Determination of Exposure to the Alternaria Mycotoxin Tenuazonic

Jul 25, 2016 - Acid and Its Isomer allo-Tenuazonic Acid in a German Population by ... and allo-tenuazonic acid was calculated, being 0.183 and 0.025 μ...
0 downloads 0 Views 399KB Size
Subscriber access provided by CORNELL UNIVERSITY LIBRARY

Article

Determination of the Exposure to the Alternaria Mycotoxin Tenuazonic Acid and its Isomer allo-Tenuazonic Acid in a German Population by Stable Isotope Dilution HPLC-MS 3

Yannick Hövelmann, Sebastian Hickert, Benedikt Cramer, and Hans-Ulrich Humpf J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.6b02735 • Publication Date (Web): 25 Jul 2016 Downloaded from http://pubs.acs.org on July 26, 2016

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 31

Journal of Agricultural and Food Chemistry

1

Determination of the Exposure to the Alternaria Mycotoxin Tenuazonic Acid and its Isomer

2

allo-Tenuazonic Acid in a German Population by Stable Isotope Dilution HPLC-MS3

3

Yannick Hövelmann†, Sebastian Hickert†, ‡, Benedikt Cramer†, Hans-Ulrich Humpf*†, ‡

4



5

48149 Münster, Germany,

6



7

*Corresponding author (Tel: +49 251 83 33391; Fax: +49 251 83 33396; E-mail:

8

[email protected])

Institute of Food Chemistry, Westfälische Wilhelms-Universität Münster, Corrensstraße 45,

NRW Graduate School of Chemistry, Wilhelm-Klemm-Str. 10, 48149 Münster, Germany.

1 ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry

9

Page 2 of 31

ABSTRACT

10

The content of the Alternaria toxin tenuazonic acid and its isomer allo-tenuazonic acid was

11

quantitated in urine of a German cohort (n = 48) using a newly developed and successfully

12

validated solid phase extraction based stable isotope dilution HPLC-MS3 method. Tenuazonic

13

acid was detected in all of the samples and quantifiable in 97.9% of these samples in a range

14

of 0.16 – 44.4 ng/mL (average: 6.58 ng/mL) or 0.07 - 63.8 ng/mg creatinine (average:

15

8.13 ng/mg creatinine). allo-tenuazonic acid was for the first time detected in human urine

16

(95.8% of the samples positive) and quantitated in 68.8% of the samples in a range of 0.11 –

17

5.72 ng/mL (average: 1.25 ng/mL) or 0.08 – 10.1 ng/mg creatinine (average: 1.52 ng/mg

18

creatinine), representing 3.40 – 25.0% of the sum of both isomers (average: 12.4%). Food

19

frequency questionnaires were used to document food consumption of study participants in

20

order to correlate mycotoxin exposure to nutritional habits. Though no statistically

21

significant correlation between consumption of a specific food and urinary excretion of

22

tenuazonic acid could be determined, a trend regarding elevated intake of cereal products

23

and higher excretion of tenuazonic acid was evident. Based on these results, a provisional

24

mean daily intake (PDI) for both tenuazonic acid and allo-tenuazonic acid was calculated,

25

being 0.183 µg/kg body weight and 0.025 µg/kg body weight, respectively. A combined

26

mean PDI for both isomers amounts to 0.208 µg/kg body weight with the highest individual

27

PDI for one of the participants (1.582 µg/kg body weight) slightly exceeding the threshold of

28

toxicological concern assumed for tenuazonic acid by the European Food Safety Authority of

29

1.500 µg/kg body weight. This is the first study which investigated the tenuazonic acid

30

content in human urine of a larger sample cohort enabling the calculation of PDIs for

31

tenuazonic acid and allo-tenuazonic acid.

2 ACS Paragon Plus Environment

Page 3 of 31

Journal of Agricultural and Food Chemistry

32

Keywords: mycotoxin, alternaria, mass spectrometry, stable isotope dilution assay, liquid

33

chromatography, exposure assessment, MRM3, HPLC-MS/MS

3 ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry

Page 4 of 31

34

INTRODUCTION

35

Tenuazonic acid, 1 (Figure 1) is a mycotoxin predominantly produced by fungi of the genus

36

Alternaria, whereas Pyricularia oryzae1 and Phoma sorghina2 are also known to

37

biosynthesize 1. Apart from its acute toxicity towards rodents (LD50 for mice:

38

81 - 186 mg/kg body weight; LD50 for rats: 168 - 180 mg/kg body weight)3 and chicken

39

embryos (LD50 = 0.55 mg/egg),4 1 showed adverse effects in several animal feeding trials,5-7

40

for instance causing acute vomiting, bloody diarrhea and death after 8 - 18 days after oral

41

application of 10 mg/kg body weight per day to dogs.5 1 has also been associated with

42

Onyalai, a haematologic disorder occurring in black populations of Africa2. Furthermore, 1

43

exhibits cytotoxic, antibacterial and antiviral activity8 as well as phytotoxic effects.9

44

Upon treatment with bases10-12 as well as under acidic conditions13 epimerization of 1 has

45

been observed, yielding the isomer allo-tenuazonic acid, 2 (Figure 2). As the configuration of

46

various moieties of 1 and its analogues seems to be particularly important for toxic effects in

47

different test systems,8,

48

separately for an accurate risk assessment. Mixtures of 1 and 2 have been reported to be

49

phytotoxic independent of the ratio of both isomers9, whereas 2 showed no toxic effects on

50

HT-29 cells in contrast to 1 significantly reducing cell viability14. Further toxicological data

51

concerning 2 as well as data on the chronic toxicity of both isomers in general are lacking.

52

There are many reports about the occurrence of 1 in food commodities, particularly in

53

cereals and products thereof15-17 as well as in tomatoes and respective processed products.

54

17-19

55

and infant food.22 According to the recent opinion of the European Food Safety Authority

56

(EFSA) on the risks for public health related to 1 in food, one of the most important

14

it is essential to determine the exposure towards 1 and 2

1 has further been detected in fruit juices, 15, 16, 18 beer,20 wine,21 edible oils,15 spices16

4 ACS Paragon Plus Environment

Page 5 of 31

Journal of Agricultural and Food Chemistry

57

contributors to the exposure towards 1 are grain and grain products23. Recently, 1 was for

58

the first time quantitated in human urine of six volunteers as part of a study in southern

59

Germany24. The presence of 2 in foods was only shown in tomato products so far.14

60

Quantitative analysis of 1 is usually based on high performance liquid chromatography

61

(HPLC). Due to high acidity and strong metal-chelating properties causing irreproducible

62

chromatographic behavior, chromatography of 1 can be quite demanding. Improving

63

retention by addition of modifiers to the mobile phase (e.g. Zn(II)SO4) or by application of

64

ion chromatography techniques25 often excludes mass spectrometric (MS) detection, since

65

most of the used additives are either not volatile or suppress ion intensity. One possibility to

66

achieve better chromatographic behavior on reversed phase columns and at the same time

67

allowing sensitive MS detection was presented by Siegel et al.26 in 2009 based on

68

derivatization of 1 with 2,4-dinitrophenylhydrazine (DNPH) prior to HPLC-MS/MS analysis. A

69

drawback of this approach as well as most published methods dealing with determination of

70

1 in food commodities and physiological samples is that no chromatographic separation of 1

71

and 2 is achieved. Since 2 has been shown to exhibit no cytotoxic effects in contrast to 1,14

72

determination of both isomers as a sum parameter might result in overestimating the

73

potential health risk. In a recent study a QuEChERS (quick, easy, cheap, effective, rugged and

74

safe) based stable isotope dilution method was developed for detection and quantitation of

75

1 and 2 in tomato products, without laborious derivatization and including chromatographic

76

separation of both toxins on a high carbon load column.14 The same type of high carbon load

77

column was also employed by Siegel et al.13 for the chromatographic separation of 1 and 2.

78

Based on these findings, the objective of this study was the development of a sensitive

79

method for the determination of 1 as well as 2 in human urine and its application to the

80

analysis of both toxins in urine samples collected as part of a recently conducted human 5 ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry

Page 6 of 31

81

study.27 Since urine is a rather complex matrix, quantitation of analytes is often hampered by

82

interfering matrix compounds. Therefore, the multiple reaction monitoring with multistage

83

fragmentation (MRM3) technique was applied in order to achieve an increase in sensitivity

84

compared to conventional MRM. The difference between both experiments is that for MRM3

85

a subsequent second fragmentation step is implemented in Q3, which is operated as a linear

86

ion trap. Thus a specific fragment ion generated due to collision of a precursor in Q2 is

87

subjected to another fragmentation and the “granddaughter” ion is finally used for detection

88

and quantitation. The second stage fragmentation increases selectivity, resulting in higher

89

sensitivity and better signal-to-noise ratios. Lastly, correlations between nutritional habits

90

and urinary excretion of 1 were investigated based on a food frequency questionnaire (FFQ).

91

MATERIALS AND METHODS

92

Chemicals and Reagents

93

All solvents used were of gradient grade and purchased from VWR (Darmstadt, Germany), if

94

not stated otherwise. ASTM type 1 water was produced with a Purelab Flex 2 system from

95

Veolia Water Technologies (Celle, Germany). Equimolar mixtures of 1 and 2 as well as of the

96

labeled standards (13C2-1/13C2-2) were synthesized in our working group19, including

97

confirmation of the respective equimolar composition by 1H-NMR19 and HPLC-MS/MS.14

98

Preparation of standard solutions

99

The equimolar mixtures of 1 and 2 as well as

13

C2-1 and

13

C2-2 were both dissolved in

100

acetonitrile and the exact concentrations determined by UV spectroscopy on a V-700 series

101

spectrometer (Jasco Labor- u. Datentechnik GmbH, Groß-Umstadt, Germany) using the

102

molar absorptivity value of 1.427 ± 0.016 x 104 L/mol/cm at 276 nm in acetonitrile.14 As the

6 ACS Paragon Plus Environment

Page 7 of 31

Journal of Agricultural and Food Chemistry

103

spectra for 1 and 2 show no significant differences,8 the sum concentration of both isomers

104

can be determined by UV spectroscopy.

105

Stock solutions of 395.5 ± 1.09 µg/mL 1/2 and 54.30 ± 2.70 µg/mL

106

obtained. The solutions were stored at -20 °C in the dark prior to usage to ensure stability

107

over the course of the experiment.13

108

13

C2-1/13C2-2 were

Calibration

109

The stock solutions of 1 and 2 as well as of the respective 13C2-labeled standards were both

110

equimolar mixtures containing the same concentration for the two isomers. Starting from

111

the stock solution, eight calibration solutions were prepared in the range between 2 –

112

200 ng/mL for unlabeled 1 and 2, respectively, by dilution with mobile phase of the starting

113

composition of the HPLC run. The calibration solutions were spiked with

114

obtain the same concentration of both labeled isomers in each calibration solution

115

(100 ng/mL).

116

13

C2-1/13C2-2 to

Sample collection

117

Urine samples (n = 48) were randomly picked out of the sample pool of a recently conducted

118

study in our working group27 with a total of 101 volunteers (44 males, 57 females),

119

predominantly university students in the age group of 20 – 30 years. The study deals with

120

the assessment of mycotoxin exposure (neither 1 nor 2 was investigated) in Germany and its

121

design was approved by the research ethical committee of the University Hospital Münster,

122

Germany (File reference: 2012-378-f-S). Participants filled out an FFQ providing information

123

on food consumption over the last 24 h, the last 30 days as well as on age, body weight and

124

height to determine possible correlations between nutritional habits and biomarker

125

excretion. To normalize biomarker concentrations to ng/mg creatinine the urinary creatinine 7 ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry

Page 8 of 31

126

levels were measured for each sample using the Jaffe method (Siemens Healthcare

127

Diagnostics, Eschborn, Germany). Measurements were carried out at the central laboratory

128

of the University Hospital, Münster, Germany. The samples were stored at -80 °C and

129

thawed directly before sample preparation.

130

Sample preparation

131

An aliquot of the urine sample (9.89 mL) was precisely transferred in a 50 mL polypropylene

132

tube. After addition of 100 µL formic acid the sample was spiked with labeled standard

133

(10 µg/mL 13C2-1/2; 10 µL = 5 ng/mL for both isomers). The sample was shaken on a vortex

134

shaker for 10 s and used for solid phase extraction.

135

Solid phase extraction

136

For solid phase extraction a 6 mL PS/DVB column (200 mg, 46.2 µm, 170 Å, Bond Elut Plexa,

137

Agilent Technologies, Waldbronn, Germany) was used. After conditioning with 3 mL iso-

138

propanol and 3 mL water at a flow rate of about 2 drops/s by applying positive pressure

139

using a syringe, the urine sample (4 mL) was passed through the column at a flow rate of

140

about 1 drop/s. Washing was carried out with 3 mL of methanol/water (20:80, v/v), followed

141

by drying the cartridge by continuously applying positive pressure for 2 min. Afterwards, the

142

analytes were eluted with 2 mL (2 x 1 mL) methanol containing 1% formic acid at a flow rate

143

of about 1 drop/s. Subsequent to evaporation to dryness at 40 °C in a nitrogen stream, the

144

remainder was reconstituted in 200 µL acetonitrile/water/formic acid (5:95:1, v/v/v), leading

145

to an almost twentyfold concentration of the analytes. After filtration through a 0.45 µm

146

PTFE Membrane (Phenomenex, Aschaffenburg, Germany) 20 µL were injected into the HPLC-

147

MS/MS system. Each sample was worked up in duplicate. Samples exceeding the calibration

8 ACS Paragon Plus Environment

Page 9 of 31

Journal of Agricultural and Food Chemistry

148

range regarding the concentration of 1 were diluted with water prior to the sample

149

preparation described above.

150

Method performance

151

Determination of the limit of detection (LOD) and limit of quantitation (LOQ) as well as of

152

the recovery rate was carried out for 2 only, since no urine sample free of 1 could be

153

obtained. In order to determine LOD, blank urine without detectable amounts of 2 was

154

spiked with 2 at four different concentrations (0.05 ng/mL; 0.30 ng/mL; 0.50 ng/mL;

155

0.75 ng/mL) in duplicate and treated as described above. Based on the resulting calibration

156

curve, the LOD was calculated as the corresponding concentration to the response of blank

157

urine (in quadruplicate) plus three times the standard deviation of the respective response.

158

The LOQ equals three times the calculated value for LOD.

159

The recovery rate was also calculated for 2 only for the above-mentioned reason. Therefore

160

blank urine was spiked with 2 at three levels within the calibration range (2.5 ng/mL;

161

5.0 ng/mL; 7.5 ng/mL) and underwent the sample preparation described above. Experiments

162

were performed in duplicate and the recovery rate was determined by averaging the single

163

recovery rates obtained for each spiked sample.

164

As the above-mentioned almost twentyfold concentration of the samples bears the risk of

165

signal suppression due to coeluting matrix compounds, the signal suppression/enhancement

166

(SSE) factor was estimated. To that end calibration solutions for 2 were prepared in the same

167

concentration range in HPLC mobile phase as well as in blank urine matrix that underwent

168

the described sample preparation, yielding two separate calibration curves. The respective

169

slopes

170

suppression/enhancement.

of

the

two

calibration

curves

were

used

to

calculate

the

signal

9 ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry

Page 10 of 31

171

Reproducibility of the method was determined by analyzing the urine sample of one of the

172

participants (2.40 ng/mL 1; 0.42 ng/mL 2) at three different days in duplicate during a course

173

of two weeks.

174

Calculation of the tenuazonic acid and allo-tenuazonic acid content in urine samples

175

To evaluate the recorded data, Analyst software, ver. 1.6.2, (Sciex, Darmstadt, Germany) was

176

used. The peak areas of the quantifier MRM3 for 1 and 2 in the calibration solutions were

177

divided by the peak areas of the respective

178

ratios were plotted against the concentrations of 1 and 2 to obtain two separate calibration

179

curves. All calibration solutions were prepared in duplicate and the peak areas of both

180

measurements averaged. The slope and intercept of these curves were used to calculate the

181

concentrations of 1 and 2. The sample volume and concentration factor resulting from

182

sample preparation were taken into account; the values were not corrected by recovery

183

rates as these were found to be close to 100%.

184

13

C2-labeled substances. Resulting peak area

Correlation of specific food intake and urinary excretion of tenuazonic acid

185

In order to identify correlations between consumption of a specific food item and urinary

186

excretion of 1, FFQs provided by the participants were compared with their respective

187

urinary excretion of 1. To that end ingested amounts of the foods of interest were plotted

188

against the excretion of 1 for all the study participants. Participants were also divided into

189

percentiles representing low, medium, increased and high intake of a specific food item,

190

whereupon the averaged excretion of 1 for the respective food groups were compared.

191

HPLC-MS/MS settings

192

A QTRAP 6500 mass spectrometer (Sciex) equipped with a 1260 Infinity LC System (Agilent,

193

Waldbronn, Germany) was used for the analysis. Electrospray ionization was applied in 10 ACS Paragon Plus Environment

Page 11 of 31

Journal of Agricultural and Food Chemistry

194

positive ion mode. The source temperature was set to 500 °C, the curtain gas to 30 psi and

195

the ion source gases 1 and 2 were set to 35 and 45 psi, respectively. Ion spray voltage was

196

set to 4500 V and the collision gas parameter set to “high”. Optimization of declustering

197

potential (DP) as well as collision energy (CE) for 1 and 2 and the respective labeled

198

standards was carried out by direct infusion of these substances into the MS system using a

199

syringe pump. For MRM3 Q3 was operated in LIT mode with the scan rate set to 10000 Da/s.

200

Various ion trap fill times (30 ms, 60 ms, 90 ms, 120 ms, 150 ms, 200 ms) were investigated,

201

whereas a linear increase of signal intensity with increasing fill time was observed.

202

Therefore, the maximum fill time of 200 ms was applied in the final method for each of the

203

MRM3 transitions with the excitation time set to 25 ms. Substance specific MRM3

204

parameters were tuned with a standard solution containing the respective compounds using

205

chromatographic runs. In that process different excitation energies (AF2) were evaluated.

206

Best results regarding signal intensities were achieved with the excitation energy set at

207

0.06 V for all MRM3 transitions. MRM transitions were also recorded with a dwell time of

208

10 ms for comparative reasons.

209

Separation of 1 and 2 was achieved on a 100 mm x 2.1 mm i.d., 3 µm, Hypercarb column

210

equipped with a 10 mm x 2.1 mm guard column of the same material (Thermo Scientific,

211

Braunschweig, Germany). A binary gradient consisting of acetonitrile (A) and water (B), both

212

containing 1% formic acid, was applied with the column temperature set to 80 °C. 20 µL of

213

sample were injected under starting conditions of the HPLC run at 5% A at a flow rate of

214

250 µL/min, which were held constant for 1.5 min. After linearly increasing the percentage

215

of A to 70% and the flow rate to 325 µL/min at 14.5 min, the content of A was further

216

increased to 95% with the same flow rate at 16 min. These conditions were kept constant for

217

2.5 min, followed by 8 min of equilibrating the column under starting conditions prior to the 11 ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry

Page 12 of 31

218

next injection. A diverter valve was integrated into the method discarding the first 10 min

219

and the last 6.5 min of each HPLC run to avoid unnecessary contamination of the MS system

220

by matrix components.

221

RESULTS AND DISCUSSION

222

Method development

223

In this study, a quick and effective SPE based stable isotope dilution HPLC-MS3 method was

224

developed, successfully validated and applied to the determination of both tenuazonic acid,

225

1 and allo-tenuazonic acid, 2 in human urine, without requiring laborious derivatization for

226

sample preparation. Frequently occurring difficulties associated with chromatography of 1

227

are overcome by using a high carbon load column, enabling almost complete

228

chromatographic baseline separation of both isomers (Figure 2).

229

Method performance

230

The determination of the limit of detection (LOD) and the limit of quantitation (LOQ) as well

231

as recovery rates was carried out for 2 only, since no urine sample free of 1 was available.

232

The determined parameters for 2 are applicable for 1 as well due to the same mass

233

spectrometric behavior of both isomers. A similar approach was recently applied for the

234

determination of method performance characteristics for ochratoxin A (OTA) and its isomer

235

2'R-OTA in blood samples.28 The LOD and LOQ for 2 were found to be 0.04 ng/mL and

236

0.11 ng/mL, respectively, showing an increased sensitivity compared to a previously

237

published method for the determination of 1 in urine with an LOD of 0.2 ng/mL and an LOQ

238

of 0.6 ng/mL.24 The recovery rate was determined in duplicate at three different spiking

239

levels and was calculated to be 94.1 ± 2.6% at urine concentrations for 2 of 2.5 ng/mL,

240

5.0 ng/mL and 7.5 ng/mL, respectively. Due to the lack of legal limits for 1 and 2 in the 12 ACS Paragon Plus Environment

Page 13 of 31

Journal of Agricultural and Food Chemistry

241

European Union, specific criteria for analytical methods for both toxins are missing.

242

According to the guideline CEN/TR 16059:201029 recovery rates between 50 and 120% for

243

unregulated mycotoxins are acceptable in food analysis. Applying this criterion for the

244

analysis of human urine as well, it is met by the newly developed method regarding the

245

calculated recovery rate of 94.1 ± 2.6%. Regression coefficients (R2) of >0.99 were obtained

246

for both analytes in their respective calibration range, indicating sufficient linearity. The

247

signal suppression/enhancement factor for 2 was calculated to be 88.2%, which is

248

satisfactory regarding the almost twentyfold concentration of the samples during sample

249

preparation. Reproducibility was determined by analysis of one of the urine samples at three

250

different days in duplicate and the respective relative standard deviation was found to be

251

9.7% for 1 and 13.7% for 2.

252

Application of MRM3 technology to mycotoxin analysis

253

The newly developed method is the first application of multiple reaction monitoring with

254

multistage fragmentation (MRM3) to mycotoxin analysis in physiological samples. MRM3 is a

255

technology not commonly applied to the analysis of low molecular weight molecules ([1+H]+

256

= m/z 198). Application of MRM3 usually deals with larger molecules, such as peptides in

257

complex biological matrices30,

258

MRM3 to the quantitative analysis of smaller molecules, such as retinoic acid in complex

259

matrices and metanephrines, was recently demonstrated.33, 34 Regarding mycotoxin analysis,

260

Lim et al.35 compared the application of MRM3 and MRM to the determination of

261

trichothecenes in grains and found comparable results concerning method performance

262

criteria.

31

or complex food matrices.32 However, the application of

13 ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry

Page 14 of 31

263

As shown in Figure 3, the second fragmentation step leads to significantly reduced

264

background signals compared to conventional MRM technique. Further improvement of the

265

signal-to-noise levels was achieved by extension of the linear ion trap fill time to 200 ms. The

266

introduction of a further fragmentation step apparently does not influence the method

267

performance criteria. The long fill time does not lead to a saturation of the ion trap, as the

268

calibration curves show excellent linearity in the respective calibration range. Since the Q3 is

269

operated in a full scan mode, which is possible due to the fast scan rate of 10,000 Da/s, all

270

possible secondary fragments are recorded simultaneously. This allows to use any of the

271

secondary fragments for quantitation retrospectively in the unlikely case of interfering

272

signals for the chosen quantifier or qualifier MRM3 transitions. Lastly, the complete

273

secondary product ion spectrum adds additional information to assure the correct

274

identification of analytes.

275

Determination of tenuazonic acid and allo-tenuazonic acid in urine samples

276

The developed stable isotope dilution HPLC-MS3 method was applied to the analysis of 1 and

277

2 in urine samples of 48 volunteers. The urine samples were randomly picked out of a larger

278

sample pool (n = 101) without any regard to information given in the FFQs or previous

279

analytical results. As shown in Figure 4, all samples tested positive for 1 with 97.9% of the

280

samples being above the LOQ and in the range between 0.16 – 44.4 ng/mL or 0.07 –

281

63.8 ng/mg creatinine with average concentrations of 6.58 ng/mL or 8.13 ng/mg creatinine.

282

The results regarding the content of 1 in urine are roughly in the range of these obtained in a

283

recently published study24 in which 1 was quantitated in the 24 h-urine of six volunteers in

284

the range of 1.3 – 17.3 ng/mL or 2.3 – 10.3 ng/mg creatinine. Furthermore, in the present

285

study 2 was for the first time detected in human urine with 95.8% of the samples being

286

above the LOD and 68.8% above the LOQ in the range of 0.11 - 5.72 ng/mL or 0.08 – 14 ACS Paragon Plus Environment

Page 15 of 31

Journal of Agricultural and Food Chemistry

287

10.1 ng/mg creatinine. Average values of 2 are 1.25 ng/mL or 1.52 ng/mg creatinine. The

288

high levels of 2 found in the analyzed urine samples, representing 3.40 – 25.0% of the sum of

289

both isomers (average: 12.4%), demonstrate the relevance of 2 as a mycotoxin. Until now, 2

290

has only been detected in tomato products.14 Since 1 has been shown to epimerize under

291

acidic conditions in aqueous solutions yielding the isomer 2,13 epimerization of 1 during

292

gastric digestion might also be a source for the detected amount of 2 in the analyzed urine

293

samples. Furthermore, 2 is not produced by Alternaria species under alkaline or neutral

294

conditions but is detectable in acidic cultures14 supporting the hypothesis of an acidic-

295

catalyzed epimerization of 1 to 2. In conclusion, it remains unclear whether the amount of 2

296

in human urine is solely attributed to the consumption of food samples contaminated with 2

297

or to epimerization of 1 during gastric digestion. In that regard a human or animal trial is

298

necessary to further investigate the origin of 2 in human urine and should consequently be

299

addressed in the future.

300

Correlation of specific food intake and urinary excretion of tenuazonic acid

301

To correlate consumption of a specific food group and changes in urinary excretion of 1,

302

FFQs provided by the participants regarding certain food intake were compared with their

303

respective urinary excretion of 1. Depending on the answers given in the food frequency

304

questionnaires, study participants were divided into percentiles representing low, medium,

305

increased and high intake of the respective food group. Special emphasis was put on

306

consumption of cereal products and 24-h dietary recalls, since 1 has been shown to

307

commonly occur in cereals and products thereof23 as well to be almost completely excreted

308

after 24 h.24 In case of plotting the single data points for cereal consumption against urinary

309

excretion of 1, no correlation between the two was evident (R2