Determination of the Mycotoxin Content in ... - ACS Publications

Oct 9, 2015 - Determination of the Mycotoxin Content in Distiller's Dried Grain with Solubles Using a Multianalyte UHPLC−MS/MS Method. Michalina ...
0 downloads 0 Views 555KB Size
Subscriber access provided by CMU Libraries - http://library.cmich.edu

Article

The determination of the mycotoxin content in distiller's dried grain with solubles (DDGS) using a multi-analyte UHPLC-MS/MS method Michalina Oplatowska-Stachowiak, Simon A. Haughey, Olivier P. Chevallier, Pamela Galvin-King, Katrina Campbell, Elizabeth Magowan, Gerhard Adam, Franz Berthiller, Rudolf Krska, and Christopher T. Elliott J. Agric. Food Chem., Just Accepted Manuscript • Publication Date (Web): 09 Oct 2015 Downloaded from http://pubs.acs.org on October 9, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Agricultural and Food Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 37

Journal of Agricultural and Food Chemistry

1

The Determination of the Mycotoxin Content in Distiller's Dried Grain with

2

Solubles (DDGS) Using a Multi-Analyte UHPLC-MS/MS Method

3 4

Michalina Oplatowska-Stachowiak,†* Simon A. Haughey,† Olivier Chevallier,† Pamela Galvin-King,†

5

Katrina Campbell,† Elizabeth Magowan,‡ Gerhard Adam,§ Franz Berthiller,# Rudolf Krska,# &

6

Christopher T. Elliott†

7 8



9

Malone Road, BT9 5BN Belfast, Northern Ireland, United Kingdom

Institute for Global Food Security, School of Biological Sciences, Queen’s University Belfast, 18-30

10



11

Kingdom

12

§

13

Vienna (BOKU), Konrad Lorenz Str. 20, 3430 Tulln, Austria

14

#

15

Department of Agrobiotechnology (IFA-Tulln), University of Natural Resources and Life Sciences,

16

Vienna (BOKU), Konrad Lorenz Str. 20, 3430 Tulln, Austria

Agri-Food and Biosciences Institute, Large Park, BT26 6DR Hillsborough, Northern Ireland, United

Department of Applied Genetics and Cell Biology, University of Natural Resources and Life Sciences,

Center for Analytical Chemistry and Christian Doppler Laboratory for Mycotoxin Metabolism,

17 18

*Corresponding author (Tel: +44(0)2890976531, Fax: +44(0)2890976513, E-mail:

19

[email protected])

1

ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry

20

Abstract

21

There are more than 300 potential mycotoxins that can contaminate food and feed and

22

cause adverse effects in humans and animals. The data on the co-occurrence of mycotoxins in novel

23

animal feed materials such as distiller's dried grain with solubles (DDGS) are limited. Thus a UHPLC-

24

MS/MS method for the quantitation of 77 mycotoxins and other fungal metabolites was used to

25

analyze 169 DDGS samples produced from wheat, maize and barley and 61 grain samples. All DDGS

26

samples analyzed were contaminated with 13 to 34 different mycotoxins. Fumonisins were present

27

in all 52 maize DDGS samples (81.0 to 6890 µg/kg for fumonisin B1) and deoxynivalenol in all 99

28

wheat DDGS samples (from 39.3 to 1120 µg/kg). A number of co-occurring mycotoxins were also

29

identified. Due to the high co-occurrence of mycotoxins routine screening of the animal feed

30

ingredients is highly recommended to allow the highlighted risks to be effectively managed.

31 32

Keywords

33

DDGS, fungal contamination, feed, grain, multi-mycotoxins contamination, animal health

2

ACS Paragon Plus Environment

Page 2 of 37

Page 3 of 37

34

Journal of Agricultural and Food Chemistry

Introduction

35

Mycotoxins (Figure 1) are natural toxic secondary metabolites produced by microscopic

36

filamentous fungi such as Fusarium, Aspergillus, Penicillium, Alternaria and Claviceps spp. Fungi can

37

grow on agricultural products in the field and/or during storage, transport and processing. There are

38

more than 300 (potential) mycotoxins currently known and they have many different acute and

39

chronic toxic effects in humans and animals.1, 2 Regulated mycotoxins such as aflatoxins,

40

deoxynivalenol, ochratoxin A, zearalenone, fumonisins, patulin and T-2/HT-2 toxins have well

41

described adverse effects and there are maximum, guidance or indicative limits for them in food and

42

feed in the EU.4−6 Emerging mycotoxins are a large group of diverse compounds (for example

43

enniatins, beauvericin or moniliformin) that are being frequently detected in food and feed and for

44

which only limited data on exposure and toxicity are available.7 Masked mycotoxins such as

45

deoxynivalenol-3-glucoside, zearalenone-14-glucoside or zearalenone-14-sulphate are conjugates,

46

that are formed as a result of plant detoxification processes. Their presence can lead to

47

underestimation of the total content of mycotoxins and potential threat due to hydrolysis of masked

48

forms to toxic parents during processing or digestion.8, 9

49

Mycotoxins can contaminate a wide range of cereals and also their products as most of them

50

are relatively stable during processing.10 Starch crops such as maize, wheat, barley, rye, triticale,

51

sorghum and oat are consumed by humans and animals worldwide but also a portion of them are

52

used to produce bioethanol and alcoholic beverages. In 2013 about 2% of total crops in the EU were

53

used in the production of fuel ethanol and 4% for alcoholic beverages.11 Distiller's dried grain with

54

solubles (DDGS) is a co-product of this ethanol production. DDGS contains valuable amounts of

55

protein, fat and fibre and has been used as a feed ingredient in diets of livestock, poultry and fish

56

since the 1990s. In Great Britain the use of distiller's co-products in animal feed has increased from

57

160.7 thousand tonnes in 1999 to 402.6 thousand tonnes in 2015.12 In Northern Ireland a steady

58

increase has also been observed and from the data available the use of 39.8 thousand tonnes was 3

ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry

59

reported in 2008 and 108.9 thousand tonnes in 2014.13 From the European perspective the

60

bioethanol industry produced 3.7 million tonnes of animal feed co-products in 2013/2014. The most

61

commonly used crops for bioethanol production in 2013 was maize (47%) and wheat (31%).11 Whilst

62

DDGS has been used as animal feed for a number of years, the data on the mycotoxins

63

contamination is limited only to a few studies have been published in recent years.14−18 One tonne of

64

grain processed for ethanol production will generate one third of tonne of feed co-products.

65

Therefore, it has been estimated that the possible mycotoxin contamination in DDGS increases 3

66

times compared to the original starting material.18 Determination of mycotoxin profiles in DDGS is

67

therefore important to identify the most commonly occurring mycotoxin cocktails and to assess the

68

safety of DDGS as an animal feed ingredient. There are only a few mycotoxins which are regulated in

69

food and feed in the EU and most analytical methods focus on these. However, the contamination

70

spectra are much broader and mycotoxin co-occurrence in food and feed is well documented.18−22

71

These co-occurrence data are provided by multi-mycotoxin LC-MS/MS methods that have been

72

developed over recent years. They allow the screening of dozens to hundreds of contaminants

73

simultaneously. Sulyok et al.23 initially developed a multi-analyte methods for the detection of 39

74

mycotoxins. The method was then further expanded to include 8724 and 10625 mycotoxins, 186

75

fungal and bacterial metabolites,26 191 mycotoxins and fungal metabolites27 and recently 295 fungal

76

and bacterial metabolites.28 Multi-mycotoxin LC-MS/MS was also developed for 288 pesticides

77

together with 38 mycotoxins29 and for 56 mycotoxins.30 Other method covering 23 mycotoxins was

78

reported by Monbaliu et al.19 Jackson et al.31 developed an isotope dilution LC-MS/MS method for 32

79

mycotoxins.

80

The aim of this study was to provide the first comprehensive data set on the co-occurrence

81

of mycotoxins in different types of DDGS samples. A UHPLC-MS/MS method that was developed by

82

Stratton et al.32 as part of the EU research project QSAFFE was transferred, optimized and validated

4

ACS Paragon Plus Environment

Page 4 of 37

Page 5 of 37

Journal of Agricultural and Food Chemistry

83

for the determination of 77 mycotoxins in animal feed ingredients. It was applied to screen for the

84

mycotoxins in DDGS and grain samples (230 in total) collected from the biofuel and feed industries.

85

Materials and methods

86

Chemicals

87

Magnesium sulphate (>99.5%), dimethyl sulfoxide (DMSO, >99.9%) formic acid for mass

88

spectrometry (98%), ammonium hydroxide solution (≥25% in water), LC-MS grade methanol (MeOH)

89

and LC-MS grade acetonitrile (MeCN) were obtained from Sigma-Aldrich (Gillingham, UK). Sodium

90

chloride (>99.5%) was obtained from Fisher Scientific (Loughborough, UK). Bondesil C18 40 µm was

91

supplied from Agilent (Wokingham, UK). A milli-Q system (Millipore, Molsheim, France) was used as

92

a source of deionized water.

93

Analytical standards

94

Aflatoxins B2, G1, G2 and M1, deepoxy-deoxynivalenol, fumonisins B1 and B2, verrucarol were

95

purchased from Sigma-Aldrich (Gillingham, UK). 3-acetyldeoxynivalnol, 15-acetyldeoxynivalenol,

96

aflatoxin B1, beauvericin, deoxynivalenol, deoxynivalenol-3-glucoside, 13 ergot alkaloid toxins

97

(agroclavine, ergocornine, ergocorninine, ergocristine, ergocristinine, ergocryptine, ergocryptinine,

98

ergometrine, ergometrinine, ergosine, ergosinine, ergotamine, ergotaminine), fumonisin B3,

99

fusarenon X, neosolaniol, nivalenol, ochratoxin A, roquefortine C, T-2 toxin, verruculogen,

100

zearalenone were obtained from Romer Labs (Tulln, Austria). Alternariol, alternariol monomethyl

101

ether, aurofusarin, curvularin, equisetin, ochratoxin B, stachybotrylactam were obtained from

102

Insight Biotechnology (Middlesex, UK). Citrinin, cytochalasin B, enniatins A, A1, B and B1, gliotoxin,

103

meleagrin, moniliformin, mycophenolic acid, penitrem A, α-zearalanol, β-zearalanol, zearalanone, α-

104

zearalenol, β-zearalenol were obtained from tebu-bio (Peterborough, UK). Cyclopiazonic acid was

105

obtained from Abcam (Cambridge, UK). Altenuene was purchased from Analyticon Discovery

106

(Potsdam, Germany). Diacetoxyscripenol was obtained from Discovery Fine Chemicals (Wimborne,

107

UK). Emodin and penicilic acid were purchased from Cambridge Bioscience (Cambridge, UK). Fusaric 5

ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry

108

acid and patulin were obtained from Fisher Scientific (Loughborough, UK). Apicidin, HT-2 toxin,

109

macrosporin, skyrin, sterigmatocystin and tentoxin were purchased from Enzo Life Sciences (Exeter,

110

UK). Masked mycotoxins: zearalenone-14-glucoside, zearalenone-16-glucoside, zearalenone-14-

111

sulphate, α-zearalenol-14-glucoside, β-zearalenol-14-glucoside were obtained from the Department

112

of Agrobiotechnology (IFA-Tulln), University of Natural Resources and Life Sciences, Vienna, Austria.

113

The standards that were obtained in a powder forms were prepared at the concentration 1 mg/mL

114

in appropriate amount of solvent (MeCN or MeOH) according to the manufacturer instructions. The

115

standards were stored at -20 °C or 4 °C depending on the recommended storage condition. For the

116

spiking experiments the concentrated stocks were prepared in volumetric flasks by mixing

117

appropriate amounts of single mycotoxins. The concentrated stock of all the mycotoxins included in

118

the method except for fumonisins B1, B2 and B3, 5 masked zearalenone metabolites, deoxynivalenol-

119

3-glucoside and moniliformin was defined as ALPHA. The mycotoxins were assigned to 7 different

120

calibration groups (Table 2) depending on the requirements and/or sensitivity and appropriate

121

amounts of single mycotoxins were mixed together in such a way that after 50 times dilution of the

122

ALPHA stock in solvent first calibration level was obtained for each mycotoxins . Similarly, fumonisins

123

B1, B2 and B3 were mixed together in MeCN:H2O (50:50) and this solution was defined as FUM stock.

124

The third calibration stock, defined as MM stock, contained 5 masked zearalenone metabolites,

125

deoxynivalenol-3-glucoside and monilifomin. FUM and MM concentrated stocks were prepared in

126

such a way that after 50 times dilution in solvent the first calibration level was obtained for each

127

mycotoxin. The further intermediate stocks of ALPHA and FUM were obtained by diluting the

128

concentrated stocks 10 times. These 5 solutions were used for preparing calibrants in matrix.

129

Samples

130

In total 230 samples of wheat, maize, barley and mixed DDGS, wheat grain and barley grain

131

were obtained from the biofuel and feed industries (Table 1). One biofuel plant provided a set of 52

132

wheat DDGS samples and a set of 52 wheat grain samples used to produce these DDGS samples.

6

ACS Paragon Plus Environment

Page 6 of 37

Page 7 of 37

Journal of Agricultural and Food Chemistry

133

Similarly, two sets of nine barley DDGS and barley grain samples were also received from the same

134

plant.

135

Sample preparation

136

The samples weighing between 1 and 2 kg were homogenized to obtain fine powder using a

137

laboratory blender. The sample extraction procedure was based on the QuEChERS method29,30 with a

138

few modifications. 1.00 ± 0.01 g of sample was weighted into a 50 mL polypropylene tube. 5.00 mL

139

of 2% formic acid in water (v/v) was added and the sample was left to soak for 30 min. Then 5.00 mL

140

of acetonitrile was added and the sample was vortexed for 30 min on multi-tube vortexer. Then 2.00

141

g of magnesium sulphate and 0.05 g of sodium chloride was added and the tube was immediately

142

shaken for 30 sec. After centrifugation for 5 min at 4000×g to induce separation of aqueous phase

143

from the organic phase 2.00 mL of the upper organic phase was collected and added to a 15 mL

144

centrifuge tube containing 0.10 g of C18 silica and 0.30 g of magnesium sulphate. The tube was

145

shaken immediately for 30 s and then centrifuged at 4000×g for 1 min. An aliquot of 1.00 mL of the

146

sample was collected into a HPLC glass vial and 200 µL of DMSO was added. Extract was

147

concentrated at room temperature under nitrogen on minivap blowdown evaporator for 45 min to

148

evaporate the remaining acetonitrile, then 800 µL of MeOH was added to each sample. The resulting

149

solution was filtered through 0.2 µm PTFE syringe filter (GE Healthcare, Chalfont St Giles, UK) and

150

transferred to a new vial.

151

UHPLC-MS/MS analysis

152

Chromatographic separation was carried out using an Acquity UPLC I-class system equipped

153

with a 100 mm × 2.1 mm i.d., 1.6 µm, CORTECS UPLC C18 column (Waters, Milford, MA). The column

154

temperature was maintained at 40 °C and the injection volume was 1 µL. Two mobile phases sets

155

were used: ''acidic'': (A) 0.1% formic acid in water and (B) in MeOH:MeCN (1:1, v/v); and ''neutral'':

156

(A) 1 mM ammonium formate in water (pH adjusted to 6.5) and (B) in MEOH:MeCN (1:1, v/v). The

157

gradient was the same for both sets of mobile phases: the starting composition was 1% organic

7

ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry

158

phase (B) with linear change to 99% with a flow rate of 0.4 mL/min over 10 min, followed by a

159

cleaning step consisting of maintaining the final composition of 1:99 (A:B) with an increased flow

160

rate to 0.6 mL/min over 3.5 min before returning to original conditions (99:1, A:B, flow rate

161

0.4 mL/min). 34 mycotoxins were analyzed in neutral mobile phase and 43 in acidic.

162

The Aquity UPLC system was coupled to Xevo TQ-S triple quadrupole tandem mass

163

spectrometer (Waters) with electrospray ion source (ESI) operating with polarity switching in a single

164

injection. The ion source parameters were as follows for both ES+ and ES-: capillary voltage 1.5 kV,

165

source temperature 150 °C, desolvation temperature 500 °C, desolvation gas flow 1000 L/h and cone

166

gas flow 150 L/h. Cone voltage and collision energy were optimized by infusion of each individual

167

analyte. The optimization was performed using automatic IntelliStart function.

168

Calibration

169

The quantitation was achieved by preparing a standard curve consisting of samples spiked

170

before extraction to correct for the recovery losses. The blank matrix of the same type as the set of

171

samples to be analyzed was selected. As it was impossible to find a single sample completely free

172

from all mycotoxins the matrix with the lowest amount of mycotoxins and at lowest levels was used

173

to prepare eight calibration levels by spiking with appropriate amounts of mycotoxin stock ALPHA

174

and FUM solutions. Two non-spiked samples were also prepared. The calibration curve in matrix was

175

extracted together with each batch of samples of the same matrix type. Additionally, each run

176

contained also four recovery controls: matrix spiked after extraction, two samples at level 5 and two

177

samples at level 2. Depending on the sensitivity and method requirements the mycotoxins were

178

allocated to different concentration groups (Table 2). Due to limitations in amount of standards

179

available for the masked zearalenone mycotoxins, deoxynivalenol-3-glucoside and moniliformin, it

180

was not possible to prepare calibrants spiked before extraction for these mycotoxins. MM

181

mycotoxins stock containing these 7 analytes was used to prepare additional calibration curve in

182

matrix spiked after extraction. If these toxins were present in the analyzed sample, the

8

ACS Paragon Plus Environment

Page 8 of 37

Page 9 of 37

Journal of Agricultural and Food Chemistry

183

concentrations obtained were corrected for recovery losses using the recovery data obtained during

184

validation.

185

The data were analyzed using TargetLynx processing software (Waters, Wilmslow, UK). Linear

186

1/x weighted calibration curves were calculated. When the blank matrix used for calibration was

187

contaminated with mycotoxin the area of the analyte peak in the blank was subtracted from the

188

area of each standard peak and new calibration was constructed that was used to calculate the

189

concentration of the mycotoxins in unknown samples.

190

Method characterization and validation

191

Matrix induced suppression/enhancement (SSE) was determined by comparing the response for

192

the matrix spiked with standards after extraction (at level 2) to a solvent standard at the same

193

concentration. This experiment was performed for 6 different samples of each 6 matrices: wheat

194

DDGS, maize DDGS, barley DDGS, wheat, maize and barley. SSE was calculated as the ratio of the

195

peak area of the analyte in the matrix and solvent multiplied by 100.

196

Method performance parameters: limits of detection (LOD), limits of quantitation (LOQ) and

197

recoveries were determined for 6 matrices. LOD was estimated as the lowest matrix-matched

198

calibration standard corresponding to a signal to noise ratio at least 3:1 and LOQ to at least 10:1.

199

Extraction recovery was determined by analyzing the samples spiked before extraction and spiked

200

after extraction and calculating the ratio of the peak areas for each analyte. Two blank samples of

201

each of six matrices were spiked before extraction at level 2 and level 5 by adding appropriate

202

amount of solvent standards directly to the samples. Spiking concentrations for each group of

203

mycotoxins at each level are presented in Table 2. Two blank samples of each matrix were also

204

extracted at the same time and the extracts were spiked after extraction with appropriate amount of

205

solvent standards to obtain concentration level 2 and level 5 (taking into account dilution factor for

206

the sample preparation method which was 5). Two replicates at each level were prepared. The

207

extraction recovery (%) was calculated as the ratio of the peak area in the sample spiked before

9

ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry

208

extraction to the peak area in the sample spiked after extraction at the same level multiplied by 100.

209

From each experiment a set of two recovery data were obtained: one for spiking level 2 and one for

210

level 5. The extraction recovery data were collected from three different experiments for all DDGS

211

matrices and maize grain and from two experiments for wheat and barley grains. They were used to

212

calculate the mean extraction recovery ± SD.

213

Within-laboratory reproducibility study was performed for DDGS wheat, maize and barley

214

matrices for 70 mycotoxins. Seven mycotoxins (MM stock) were not included due to limited amount

215

of standards available. As it was impossible to select samples that were completely free from

216

mycotoxins, three representative samples that contained the lowest levels and smallest number of

217

mycotoxins were selected and used as blanks for validation. The samples were spiked at 1.0, 1.5 and

218

2.0 × level 5 (n=6 at each level). A calibration curve consisting of samples spiked before extraction

219

was prepared at the same time. The validation was performed as within-laboratory reproducibility

220

with three different analysts performing the experiments on three different days (separately for

221

each matrix). The concentrations of each mycotoxin in 18 spiked samples were calculated from the

222

standard curve that was prepared and extracted with each set of samples. The data were used to

223

calculate within-laboratory accuracy and precision.

224

Results and Discussion

225

Sample preparation

226

The extraction method based on QuEChERS was used in this study. First, the partition of water

227

and acetonitrile phase was achieved using magnesium sulphate to reduces polar matrix components

228

in the organic phase which was collected after this step. Then the C18 silica clean-up step was applied

229

to eliminate non-polar di- and triglycerides that normally elute in 100% of organic solvent at the end

230

of the run during column equilibration; however they can also slowly accumulate in the column

231

reducing its life.29, 30 Therefore, application of this simple step was recommended to prolong column

232

service life. Some modifications were introduced to the method. When a sample was injected 10

ACS Paragon Plus Environment

Page 10 of 37

Page 11 of 37

Journal of Agricultural and Food Chemistry

233

directly in acetonitrile the peak shapes for some analytes were not satisfactory. Dilution of the

234

sample in water 1:1 gave satisfactory peak shapes, however it was observed that, after a couple of

235

hours, samples accumulated precipitate that could have had a detrimental effect on the UPLC

236

performance. It was determined, that if the reconstitution solvent was 100% methanol, the

237

precipitation did not occur and the peak shapes were satisfactory. Therefore, the evaporation step

238

was introduced to replace acetonitrile with methanol. A small amount of DMSO was used during the

239

evaporation to act as a ‘’keeper’’ to assure analytes remained in the solution and did not bind to the

240

glass vial during evaporation.33

241

Method performance characteristics and validation

242

LOD ranged from 0.005 to 250 µg/kg. LOD and LOQ could only be determined in matrices that

243

were completely free from the analytes. No single matrix was free from beauvericin. Enniatins A, A1,

244

B and B1 were found in every samples analyzed apart from maize. There was no blank matrix

245

available for some ergot alkaloids in wheat and barley DDGS. Fumonisins were present in all maize,

246

maize DDGS and barley DDGS samples. Several other mycotoxins were also found in some of the

247

matrices limiting the possibility of determining the LOD and LOQ for each mycotoxin in each matrix.

248

In general the LOD and LOQ were higher in DDGS samples than in grain samples due to the higher

249

matrix effect.

250

Signal suppression/enhancement (SSE) for most of the analytes in grain samples (wheat, maize

251

and barley) was within the range 80−120%. Matrix-induced suppression below 80% was observed for

252

two analytes: zearalenone-16-glucoside and moniliformin in all three matrices and for a further

253

three mycotoxins: zearalenone-14-sulphate, deoxynivalenol and deoxynivalenol-3-glucoside in

254

barley only. Matrix-induced enhancement between 152−171% was observed for fumonisins in all

255

three matrices and a further 7 and 6 analytes showed more than 120% SSE in maize and barley,

256

respectively. In general, these three matrices presented similar SSE patterns with wheat causing the

257

lowest matrix effect. The matrix effect in DDGS was higher and SSE was below 80% for 17, 32 and 17

11

ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry

258

analytes in wheat, maize and barley DDGS, respectively. Matrix-induced signal enhancement was

259

observed for fumonisins, apicidin and aurofusarin in all DDGS matrices and for a further 2 and 4

260

analytes in wheat and maize DDGS, respectively. All three DDGS matrices showed a similar profile of

261

SSE. The highest variation between these three matrices was observed for aflatoxin G1, citrinin and

262

moniliformin.

263

The extraction recovery for most of the analytes was in the range 70−110% for all the matrices

264

tested. Incomplete extraction was observed for deoxynivalenol-3-glucoside (32−48%). The extraction

265

recovery for moniliformin was only 13−21%, therefore the method was not suitable for quantitative

266

analysis of this mycotoxin. Additionally, the recovery was determined to be slightly lower than 70%

267

for citrinin, aurofusarin, fusaric acid, nivalenol and penitrem A in some matrices. Ergocornine,

268

ergocristine, ergocryptine, ergosine and ergotamine but not ergometrine gave recoveries higher

269

than 110% and this effect was more significant in grain than in DDGS. Analysis of the grain samples

270

spiked before extraction with ergotaminine only confirmed that some of the analyte was converted

271

to ergotamine during extraction procedure but such conversion did not happen in the matrix

272

samples spiked after extraction.

273

The method accuracy was tested in three types of DDGS samples for 70 mycotoxins. Seven

274

toxins were not included in this study due to limited amount of standards available. The accuracy

275

was typically within the range 80−120% with results for citrinin, ergocryptine, ergotamine,

276

ergocornine and fusaric acid in wheat DDGS and for citrinin and beauvericin in maize DDGS outside

277

this range. The precision was below 20% for most of the analytes, with only a few analytes such as

278

beauvericin, citrinin, some ergot alkaloids, fusaric acid, penicillic acid, patulin giving higher values.

279

Sample analysis

280

The developed UHPLC-MS/MS method was applied to screen 169 DDGS samples and 61 grain

281

samples for 77 mycotoxins.

282

Wheat DDGS (n=47), maize DDGS (n=52) and mixed DDGS (n=9)

12

ACS Paragon Plus Environment

Page 12 of 37

Page 13 of 37

Journal of Agricultural and Food Chemistry

283

The number of mycotoxins detected in wheat DDGS samples ranged from 14 to 27 (Table 3). All

284

of the samples contained between 39.3 to 1120 µg/kg of the regulated Fusarium toxin

285

deoxynivalenol. Other emerging Fusarium toxins: enniatins A, A1, B, B1 and beauvericin were also

286

detected in every sample and enniatin B level had the highest concentration ranging from 174 to

287

1490 µg/kg. Almost half (47%) of the samples contained zearalenone. Other mycotoxins produced by

288

Fusarium including fumonisins, apicidin, aurofusarin, equisetin and fusaric acid were found in a small

289

number of samples. All the samples contained between 6 to 12 different ergot alkaloids that are

290

produced mainly by Claviceps purpurea fungi. The range of contamination was very broad with total

291

ergot alkaloids content ranging from 4.7 to 1230 µg/kg with a median value of 7.7 µg/kg. Low levels

292

(up to 3.5 µg/kg) of regulated Aspergillus/Penicillium toxin ochratoxin A were found in 81% of the

293

samples. Penicillium toxins, namely mycophenolic acid and meleagrin were found in 94% and 21% of

294

the samples, respectively.

295

Maize DDGS samples contained between 13 to 28 mycotoxins. All samples were contaminated

296

with regulated Fusarium toxins fumonisins, with fumonisin B1 having the highest concentration

297

which ranged from 81.0 to 6890 µg/kg. All samples also contained other Fusarium toxins namely

298

enniatins A, A1, B, B1, beauvericin and fusaric acid. Beauvericin concentration was higher compared

299

to enniatins and it ranged from 46.1 to 561 µg/kg. Equisetin was found in 98% of the samples. Other

300

Fusarium toxins found were T-2 toxin (81% of the samples), deoxynivalenol (62%), zearalenone

301

(42%) and β-zearalenol (10%), aurofusarin (42%) and neosolaniol (15%). Masked mycotoxin

302

deoxynivalenol-3-glucoside was found in 10% of the samples that were also highly contaminated

303

with deoxynivalenol. With regard to Aspergillus toxins aflatoxin B1 was present in 75% of the

304

samples, aflatoxin B2 in 63%, aflatoxin G1 in 37% and aflatoxin G2 in 2%. Traces of aflatoxin M1 were

305

also found in 50% of the samples. Aflatoxin M1 is a metabolite of aflatoxin B1 that is excreted into

306

milk, however it has also been reported to be present at low levels in such commodities as corn34

307

and peanuts.35 Aspergillus/Penicillium toxins ochratoxin A, ochratoxin B and cyclopiazonic acid were

13

ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry

308

detected in 67%, 23% and 73% of the samples, respectively. Other Penicillium metabolites:

309

mycophenolic acid, meleagrin, roquefortine C were also found. Alternaria toxins: alternariol,

310

alternariol monomethyl ether and tentoxin were found in 33, 21 and 27% of the samples,

311

respectively.

312

Mixed DDGS samples contained mixtures of mycotoxins characteristics noted above for both

313

wheat and maize DDGS. The number of mycotoxins in each sample was from 16 to 31. 100% of

314

samples contained Fusarium toxins: deoxynivalenol, fumonisins, zearalenone and enniatins. Other

315

Fusarium metabolites found were fusaric acid, equisetin, T-2 toxin, apicidin and aurofusarin. From 3

316

to 12 different ergot alkaloids were found in each sample. Penicillium toxins: mycophenolic acid and

317

roquefortine C were detected in 78% and 22% of the samples. Ochratoxin A was found in 22% of the

318

samples. Alternaria toxins: alternariol, alternariol monomethyl and tentoxin were also present.

319

Mycotoxin contamination is a result of infection with toxin producing fungi belonging to

320

Fusarium, Aspergillus, Penicillium, Alternaria and Claviceps genera during the growth of the cereals

321

in the field and/or post-harvest during grain storage and processing. EU Commission Directive

322

2003/10/EC3 sets the maximum limits for aflatoxin B1 (5−20 µg/kg) and Commission

323

Recommendation 2006/576/EC5 guidance values for deoxynivalenol (900−12000 µg/kg), zearalenone

324

(100−3000 µg/kg), ochratoxin A (50−250 µg/kg) and fumonsins B1+B2 (5000−60000 µg/kg) in animal

325

feed. The exact value depends on animal species and feed type. From a regulatory perspective the

326

most prevalent mycotoxins identified in this study were Fusarium toxins: deoxynivalenol in wheat

327

DDGS and fumonisins in maize DDGS. These toxins appear in DDGS at high levels and as guidance

328

limits are established for feed the monitoring of DDGS quality in terms of mycotoxin contamination

329

should focus on them. Two analyzed maize DDGS samples exceeded the guidance limit for

330

deoxynivalenol in cereal by-products to be used as feed ingredients (12000 µg/kg). The rest of the

331

samples were contaminated with mycotoxins below the guidance or indicative limits. However, the

14

ACS Paragon Plus Environment

Page 14 of 37

Page 15 of 37

Journal of Agricultural and Food Chemistry

332

co-occurrence of mycotoxins in every sample analyzed raises the question about possible synergistic

333

and/or additive effects of the cocktails of these contaminants on animal health and performance.

334

Data on the contamination of DDGS with mycotoxins are limited to a few publications focusing

335

mainly on regulated mycotoxins. The analysis of 67 maize DDGS samples for aflatoxins,

336

deoxynivalenol, fumonisins, T-2 toxin and zearalenone collected from US bioethanol industry

337

between 2009 and 2011 revealed the presence of deoxynivalenol at level higher than 2 mg/kg in

338

12% of the samples, fumonisins at the level higher than 5 mg/kg in 6% of the samples and

339

zearalenone in most of the samples at the concentration between 0.1 to 0.3 mg/kg.36 Rodrigues and

340

Chin15 analyzed 409 maize DDGS samples for aflatoxins, zearalenone, deoxynivalenol, fumonisins

341

and ochratoxin A and found 92% of the samples contaminated with more than two mycotoxins.

342

Khatibi et al.17 analyzed 141 corn DDGS samples collected from 78 bioethanol plants in the USA in

343

2011 for the presence of deoxynivalenol, 3- and 15-acetyldeoxynivalenol, nivalenol and zearalenone.

344

The authors found > 1 mg/kg of deoxynivalenol, 15-acetyldeoxynivalenol and zearalenone in 30, 15

345

and 3% of the analyzed samples, respectively. The analysis of 59 maize DDGS samples from Thailand

346

for fumonisins B1 and B2, deoxynivalenol, zearalenone and beauvericin showed co-occurrence of

347

these five mycotoxins in 50.8% of the samples.16 Li et al.37 analyzed 17 DDGS samples from China and

348

found all of them contaminated with deoxynivalenol, zearalenone and ochratoxin A and 16 of them

349

with aflatoxin B1. A more comprehensive study on DDGS was published by Zachariasova et al.18

350

Among other samples 71 maize DDGS and 16 wheat DDGS were analyzed for 56 mycotoxins. The

351

authors found that DDGS samples contained a broad spectrum of mycotoxins including

352

deoxynivalenol and its metabolites, T-2/HT-2 toxins, zearalenone and its metabolites, fumonisins,

353

enniatins, beauvericin, Alternaria toxins, ergot alkaloids, sterigmatocystin, ochratoxin A and

354

mycophenolic acid. Similar mycotoxin profiles in DDGS samples were identified in this study.

355

Regulated mycotoxins such as deoxynivalenol and fumonisins have well known adverse effects in

356

humans and animals. This study identified common co-occurrence of other mycotoxins for which

15

ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry

357

toxicity data are limited. Enniatins and beauvericin were found in all samples analyzed. They have

358

ionophoric properties and can promote transport of ions through biological membranes leading to

359

changes in normal physiological concentrations of these ions.7 Enniatins and beauvericin may

360

bioaccumulate due to their lipophilic nature. Their influence on the toxicity of other mycotoxins is

361

unknown. Another important toxin, fusaric acid, acts synergistically with deoxynivalenol when given

362

to immature pigs causing depression in weight gain and feed intake.38 Fusaric acid acts also

363

synergistically with fumonisins as enhanced toxicity was observed when fusaric acid and fumonisin

364

B1 were administered together to developing chicken embryo.39 Mycophenolic acid has

365

immunosuppressive properties; it inhibits antibody formation and the production of cytotoxic T

366

cells.40 100% of wheat DDGS was contaminated with ergot alkaloids. Ergot alkaloids can act on a

367

number of neurotransmitter receptors and interfere with reproductive processes.41 The maximum

368

content of rye ergot sclerotia in feedingstuff is 1000 mg/kg in the EU.42 While there are currently no

369

other regulatory limits for ergot alkaloids in feed in the EU, European Food Safety Authority (EFSA)

370

established NOAEL (no-observed adverse effect level) for ergot alkaloids at 0.15 mg/kg of feed for

371

pigs and 1.4 mg/kg of feed for poultry. Six wheat DDGS samples analyzed in this study exceeded or

372

were close to the NOAEL value for pig feed. The inclusion rate of DDGS in animal diet is generally in

373

the range of 10–20% therefore if these wheat DDGS batches were used as pig feed ingredient, the

374

ergot alkaloid content would be diluted below NOAEL level although it is illegal in the EU to

375

knowingly dilute contaminated materials.

376

Comparison of wheat grain and wheat DDGS samples

377

Fifty two samples of wheat grain and wheat DDGS were obtained from a bioethanol plant in

378

the EU between September and November 2014. Due to the manufacturing practice one wheat

379

grain batch could not be matched directly with another DDGS batch as different batches can be

380

mixed together during production process. As a result the data were analyzed as two sets: grain

381

wheat and DDGS separately (Table 4). A total of between 6 to 17 mycotoxins were found in grain 16

ACS Paragon Plus Environment

Page 16 of 37

Page 17 of 37

Journal of Agricultural and Food Chemistry

382

wheat samples and 17 to 24 in wheat DDGS samples. Deoxynivalenol and ergot alkaloids were not

383

detected in every grain sample, but they were found in all DDGS samples. This could be a result of

384

their presence below the LOD values in grain. After concentration during ethanol production they

385

were detectable in DDGS samples. The variation in the level of contamination for these two groups

386

of mycotoxins in wheat grain was very high ranging from traces to heavy contamination, while DDGS

387

samples were more uniformly contaminated with these mycotoxins. Deoxynivalenol was found in all

388

wheat DDGS samples and the concentration ranged from 177 to 394 µg/kg. The concentration factor

389

for deoxynivalenol calculated on the basis of the median values for DDGS and grain was 2.1. This is in

390

agreement with Hanschemann and Krieg,43 who found deoxynivalenol and zearalenone to increase

391

2−4 times during production of bioethanol from triticale. Also Schaafsma et al.,14 reported

392

deoxynivalenol concentration 3 times higher in DDGS than in maize used for its production.

393

Traces of ergot alkaloids were found only in 10 grain samples and high levels of contamination in

394

3 samples (total ergot alkaloids content: 78, 228 and 688 µg/kg). All DDGS samples were

395

contaminated with at least 10 ergot alkaloids and the sum of ergot alkaloids content ranged from

396

17.6 to 189 µg/kg.

397

The median concentrations of enniatins A, A1, B, B1 and beauvericin were higher in DDGS

398

samples by the factors of 6.5, 7, 6.5, 5 and 2.9, respectively when comparing with grain. This

399

indicates the presence of the concentration effect on these mycotoxins during bioethanol

400

production. Low levels of Fusarium toxins fumonisins, fusaric acid, zearalenone, Penicillium toxin

401

mycophenolic acid and Aspergillus/Penicillium toxin ochratoxin A were found in a higher number of

402

wheat DDGS samples than wheat grain samples, again indicating concentration of these toxins

403

during bioethanol production. All grain samples were contaminated with aurofusarin from 1.4 to 400

404

µg/kg. Aurofusarin was found in only 9 DDGS samples indicating that this toxin does not accumulate

405

in DDGS or is transformed during bioethanol production to other metabolites. DDGS was less

17

ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry

406

contaminated with apicidin and equisetin comparing to grain samples suggesting that these toxins

407

do not accumulate in DDGS or are converted to other metabolites.

408

Comparison of barley grain and barley DDGS samples

409

Nine samples of barley grain and 9 samples of barley DDGS were obtained from the same

410

bioethanol plant in the EU. A total between 31 to 34 mycotoxins were found in each DDGS sample

411

and 7 to 16 in barley grain (Table 5). All barley DDGS samples were contaminated with regulated

412

Fusarium toxins deoxynivalenol, fumonisins and zearalenone. Masked mycotoxin deoxynivalenol-3-

413

glucoside was found at a level