Analytical Methods
Application of an LC–MS/MS based multi-mycotoxin method for the semi-quantitative determination of mycotoxins occurring in different types of food infected by moulds
Michael Sulyok,Rudolf Krska *,Rainer Schuhmacher
Christian Doppler Laboratory for Mycotoxin Research,Department IFA-Tulln,University of Natural Resources and Applied Life Sciences,Vienna,Konrad Lorenzstr.20,A-3430Tulln,Austria
a r t i c l e i n f o Article history:
Received 9July 2008
Received in revised form 11May 2009Accepted 22July 2009
Keywords:Food
Liquid chromatography Mass spectrometry Moulds Mycotoxins
a b s t r a c t
An existing LC–MS/MS method for multi-mycotoxin determination was extended by further 19analytes and was applied for a semi-quantitative screening of 87mouldy food samples from private households,including bread,fruits,vegetables,cheeses,nuts and jam.In the 247investigated sub-samples,49different analytes were identified,some of which were never reported before to occur in naturally con-taminated food.Enniatins and ergot alkaloids occurred in all samples of (dark)bread/pastries at low l g/kg-levels.From the remaining analytes,chanoclavine,emodin,mycophenolic acid and roquefortine C were found most frequently.Regulated mycotoxins occurred less often,but the corresponding concentra-tions exceeded the regulatory limits up to a factor of 1000in case of patulin.Moreover,considerable mycotoxin concentrations were observed in some sub-samples taken from non-mouldy spots of the investigated samples.Thus,it was concluded that it is not safe to remove the mouldy part and consume the remainder.
Ó2009Elsevier Ltd.All rights reserved.
1.Introduction
Moulds are able to infect and grow on all types of food.This is usually accompanied by changes of th
e texture,smell and taste of the infected foodstuff due to excretion of enzymes and volatile compounds by the fungus.In some cases,these changes are desired (e.g.use of non-toxigenic strains of Penicillium roqueforti for the production of blue mould cheese),but most of the time,fungal infection leads to food spoilage such as off-flavours,discoloration,rotting and disintegration of the food structure (Filtenborg,Frisvad,&Thrane,1996).
The most important aspect involved in spoilage of food is,how-ever,the formation of mycotoxins.Although approximately 400compounds are currently recognised as mycotoxins,only few of them are addressed by food legislation.Most of the existing analyt-ical methods likewise focus on these regulated icho-thecenes,aflatoxins,zearalenone,ochratoxin A,fumonisins and patulin.In contrast,most publications dealing with other mycotox-ins produced by fungi involved in food spoilage seem to derive from the field of mycology rather than from food analysis.A typical example consists of the use of the metabolic profile of crude fungal
extracts to support findings from taxonomy for the differentiation of fungal species (Nielsen &Smedsgaard,2003).
These findings are of limited relevance concerning the occur-rence of mycotoxins in naturally infecte
d food:most mycological investigations address fungal strains that have been from food and have subsequently been cultivated on synthetic cul-ture media.However,the qualitative and quantitative mycotoxin profile,which a mould produces on a food commodity depends on the ecological and processing parameters of the particular food-stuff (Filtenborg et al.,1996)and can therefore be expected to be different from synthetic media.In addition,most of the analytical methods that are used in mycological studies such as TLC (Filten-borg et al.,1996;Freire,Kozakiewics,&Paterson,2000;Overy,Seif-ert,Savard,&Frisvad,2003)or HPLC–DAD (Andersen,Smedsgaard,&Frisvad,2004;Andersen &Frisvad,2004;Larsen,Gareis,&Fris-vad,2002)are sufficiently selective for the determination of fungal extracts or for single target analysis (including a dedicated proce-dure for sample preparation)in foodstuffs,but are incapable of dealing with a large number of analytes in complicated food matri-ces.Modern methods such as HPLC coupled to (tandem-)mass spectrometry offer higher selectivity,which enables multi-analyte determination without dedicated sample clean-up in principle.However,even for those methods,multi-mycotoxin analysis in food is a real analytical challenge,as it would be advantageous to work without any clean-up and analyze raw extracts instead in or-der not to adulterate the mycotoxin pattern by sample preparation.
0308-8146/$-see front matter Ó2009Elsevier Ltd.All rights reserved.doi:10.1016/j.foodchem.2009.07.042
*Corresponding author.Tel.:+43227266280401;fax:+43227266280403.E-mail addresses:michael.sulyok@boku.ac.at (M.Schuhmacher),rudolf.krska@boku.ac.at (R.Schuhmacher).
Food Chemistry 119(2010)
408–416
Contents lists available at ScienceDirect
Food Chemistry
j o u r n a l ho m e p a g e :w w w.e l s e v i e r.c o m /l o c a
te/foodchem
The drawback of such a‘‘dilute and shoot”-approach is that signal suppression due to matrix effects is far more likely to occur when crude extracts are analyzed.The existing method that covers hun-dreds of fungal metabolites is used rather for qualitative screening of fungal metabolites(Nielsen&Smedsgaard,2003;Nielsen,Suma-rah,Frisvad,&Miller,2006)than for quantitative analysis,which is partially a result of the lack of availability of suitable standards.On the other hand,quantitative data on metabolites involved in food spoilage is usually restricted to a selected set of analytes that are amenable to the chosen clean-up Kokkonen,Jestoi, &Rizzo,2005).
Despite these difficulties,there is certainly a need for fast and comprehensive methods for the analysi
s of toxic metabolites pro-duced by toxigenic strains of food colonising fungi,as a simple vi-sual inspection of food is not sufficient to exclude health hazards.It was shown that toxin concentration and visible infection may not correlate in every case(Rundberget,Skaar,&Flaoyen,2004),and mycotoxins can be present in commodities without being able to detect fungi associated with the toxins and vice versa(Freire et al.,2000).In addition,the use of any mouldy material in the pro-cessing of food may contribute to the mycotoxin level in thefinal product by carry over(Filtenborg et al.,1996),e.g.use of mouldy tomatoes for the production of ketchup(Andersen&Frisvad, 2004).In such cases,the mouldy material cannot be seen in thefi-nal product.A recent report by our group(Sulyok,Berthiller,Krska, &Schuhmacher,2006)has shown that the use of mass spectrome-ters of the latest generation enables a quantitative determination of mycotoxins in crude food extracts,provided that the extraction efficiencies as well as matrix effects are sufficiently characterised for all investigated analyte/matrix combinations.Only recently, we have extended the range of analytes covered by our method and have provided preliminary data on the mycotoxin pattern in mouldy food samples(Sulyok,Krska,&Schuhmacher,2007).
seifertIn the present work,the method was further extended and ap-plied for a semi-quantitative screening of247sub-samples taken from mouldy and non-mouldy spots of87food samples from pri-vate househ
olds in Austria.It was our goal to determine the myco-toxin pattern(including toxic metabolites that have not been reported yet to occur in naturally contaminated food),that is pro-duced by moulds spontaneously infecting food,stored under typi-cal conditions at the end consumer.These results might be used to identify toxins as marker substances for food spoilage and to eval-uate the relevance of the related toxin concentrations for the end consumer.Furthermore,the distribution of the toxins between mouldy and non-mouldy parts of the same sample has been stud-ied.Some of the toxins detected were only present in the moulded part of the samples whereas others were excreted by the fungus into the surrounding food tissue.There is certainly a practical rel-evance of this issue,as it is still common practice in case of some food products to remove the mouldy parts and to consume the remainder.
2.Materials and methods
2.1.Chemicals and reagents
Methanol and acetonitrile(both LC gradient grade)were pur-chased from J.T.Baker(Deventer,The Netherlands),ammonium acetate(MS grade)and glacial acetic acid(p.a.)were obtained from Sigma–Aldrich(Vienna,Austria).Water was purified successively by reverse osmosis and a Milli-Q plus syste
m from Millipore(Mols-heim,France).Mycotoxin standards were dissolved in acetonitrile and were purchased from different sources:Brefeldin A,cytochal-asins A,B,C,D,J and H,HC-toxin,kojic acid,and3-nitropropionic acid were purchased from Sigma(Vienna,Austria),penicillic acid and roquefortine C were obtained from Iris Biotech GmbH(Mark-tredwitz,Germany),AAL TA toxin was a gift from Prof.David Gil-christ(University of California,Davis,United States),Enniatin B3 and2-amino-14,16-dimethyloctadecan-3-ol were a gift from Dr. Silvio Uhlig(National Veterinary Institute,Oslo,Norway),certified reference solutions of T2-tetraol and T2-triol were received from Biopure Referenzsubstanzen GmbH(Tulln,Austria).Alpha zearale-nol-4-glucoside and beta zearalenol-4-glucoside were synthesised in our laboratory from zearalenone using a genetically modified yeast strain expressing a glucosyl-transferase,followed by reduc-tion of the resulting zearalenone-4-glucoside with sodium borohy-dride(Krenn et al.,2007).For details concerning the other87 toxins see Sulyok et al.(2007).
2.2.Samples
Eighty-seven spontaneously moulded foodstuffs(including19 breads/pastries,20fruits,14vegetables,6cheeses,5jams,6nuts and17others),which had been provided by staff members of our institute,were sampled for mycotoxins in this study.Few of the samples were
completely covered by mould,whereas most samples exhibited one or several–sometimes differently coloured –mouldy spots.The latter samples can be considered to be realistic in private household practice.
After visual inspection of each sample,several sub-samples per individual sample were prepared by cutting mouldy and(if avail-able)non-mouldy spots of the individual sample using a scalpel. (Note that the term‘‘non-mouldy”is used throughout the manu-script although we are aware that fungal mycelium may have also been present in those parts of the samples,which did not exhibit visible fungal infection.There was nofixed distance between the sampled mouldy and non-mouldy spots;one sub-sample was usu-ally taken from the maximum distance from the fungal infection.). The surface area of the sub-samples was approximately1cm2and their thickness ranged between0.5and1cm in order to make sure that the main part of the sampled volume consisted of food matrix.
Thefinal set of247sub-samples included68originating from bread/rolls(49with mouldy spots/19spots without visible infec-tion),49from fruits(30/19),34from vegetables(22/12),14from cheeses(10/4),13from nuts(10/3),12from jams(6/6)and57 from other foodstuffs(34/23).
2.3.Sample preparation and estimation of matrix effects
Extraction was carried out using a mixture of acetonitrile/ water/acetic acid79+20+1(v+v+v),with ratios between3 and16mL solvent/g sample depending on the texture of the sam-ple.After extraction,the samples were centrifuged,diluted1+1 and injected as described in detail by Sulyok et al.(2007).For the estimation of matrix effects,raw extracts of sample spots without visible fungal infections were fortified using a multi-analyte stan-dard on one concentration level,diluted and analyzed and the cor-responding peak areas were compared to a standard prepared and diluted in neat solvent.
2.4.LC–MS/MS parameters
Detection and quantification was performed with a QTrap4000 LC–MS/MS System(Applied Biosystems,Foster City,CA)equipped with a TurboIonSpray electrospray ionization(ESI)source and an 1100Series HPLC System(Agilent,Waldbronn,Germany).Chro-matographic and mass spectrometric parameters of87of the inves-tigated analytes are described by Sulyok et al.(2007).MS and MS/ MS parameters of the additional19analytes were optimised by infusion of standard solutions into the mass spectrometer.For a de-tailed list of these compounds and the related parameters see Table
M.Sulyok et al./Food Chemistry119(2010)408–416409
1.Quantification was performed in the Selected Reaction Monitor-ing(SRM)mode.Limits of detection were calculated from the signal to noise ratios(LOD=3*S/N)of the respective SRM chromato-grams as derived from the analysis of liquid standards.Enhanced Product Ion scans for identity confirmation were acquired using the third quadrupole as linear ion trap applying a dynamicfill time, whereas other parameters such as the declustering potential and the collision energy were set to the optimised values that had been determined for the respective analytes.
3.Results and discussion
3.1.Evaluation of the extended HPLC–MS/MS method
As our institute is continuously expanding the multi-mycotoxin method,we applied the latest version to this experiment,as to get as much information on contamination as possible.As our ap-proach is based on the analysis of diluted crude extracts without further sample pretreatment,the inclusion of additional analytes does not pose any problems as long as they are compatible with the chromatographic conditions and as long as they are extractable (at least to a partial,reproducible extent)using the acidified aceto-nitrile/water mixture that is applied in our method.However,it must be kept in mind that these conditions are a compromise resulting from the chemically diverse set of anal
ytes and may be far from optimal for some compounds.For example,the extension of the existing method by the highly polar substances3-nitroprop-ionic acid and kojic acid resulted in small retention factors of0.97 and1.16,respectively,of these two compounds on the applied C18-based column.Although a full method validation for the additional analytes has not been performed yet,preliminary results indicate that the method performance parameters will probably be not sig-nificantly different compared to the set of87substances:Limits of detection ranged from0.04for enniatin B3to160l g/kg for kojic acid(see Table1).The linearity of the detector signals of each of the19analytes has been confirmed over a concentration range of at least2orders of magnitude.Analyte losses due to incomplete extraction and/or matrix effects have been investigated in maize and were comparable to our initial method(Sulyok et al.,2006; data not shown).
Since the goal of this work was to provide information on the occurrence of toxic fungal metabolites in mouldy food and a rough estimate of the related concentrations,a detailed investigation on the extraction efficiencies and matrix effects in the different matri-ces was considered to go beyond the scope of this work.Judging from the results of the detailed investigation on extraction efficien-cies and matrix effects of multi-mycotoxin and multi-residue anal-ysis in various matrices(Spanjer,Rensen,&Scholten,2008;Mol et al.,2008),analyte losses due to incomplete extractio
n exceed a factor of2only for a few analyte/matrix combinations if an(acid-ified)acetonitrile/water mixture is used as extraction solvent.At the same time,in these two reports matrix effects exceeded a fac-tor of three only in few cases,but since this is an instrument-dependent parameter,we verified thisfinding by spiking various raw extracts of non-mouldy spots.The suppression of the analyti-cal signal exceeded a factor of2in the spiked raw extracts only in 12%of all investigated analyte/matrix combinations(ergopeptides and the Alternaria toxins generally being the most critical analytes in that aspect)and exceeded a factor of10in only5out of the2412 investigated analyte/matrix combinations(avenacein Y,tentoxin and HC-toxin in apricots,HC-toxin in treenuts and tomatoes).Both, incomplete extraction and matrix effects lead to an underestima-tion of the actual concentration in the mouldy samples.However, this combined error caused by matrix effects and incomplete extraction is in the worst case a factor of2–3for most analytes and(with the few exceptions stated above)a factor of approxi-mately6for the ergopeptides and citrinin(that exhibited a rather low extraction efficiency of30%in the model matrix that was investigated by Sulyok et al.,2007),which we consider acceptable for meeting our goals.
3.2.Overview of frequency and concentrations of mycotoxins in the investigated samples
Forty-nine different analytes were identified in the set of247 sub-samples of the87investigated food s
amples.Their frequencies
Table1
LC–MS/MS parameters of the19additional analytes.
Analyte t R a(min)m/z precursor ion DP b(V)m/z product ions c Rel.int.d CE e(V)c CXP f(V)c P g Dwell time(ms)LOD(l g/kg)
Kojic acid  3.22143.1[M+H]+56113.2/69.2  3.9531/2310/10+1100/25160
T2-tetraol  5.79316.2[MNH4]+31215.3/281.40.6713/1316/8+1100/2570 Penicillic acid9.30171.2[M+H]+46125.2/97.10.3817/238/16+360/2020
AAL TA-toxin11.96522.3[M+H]+71328.5/292.4  1.1335/4120/16+545/1512
T2-triol12.51400.2[M+NH4]+41215.2/281.30.3417/1312/16+545/1540 Roquefortine C12.54390.2[M+H]+61193.2/322.30.4339/2910/18+545/154 Cytochalasin J13.21452.2[M+H]+31434.5/416.50.9413/2112/12+6100/255 Cytochalasin D13.48525.2[M+NH4]+31430.5/490.50.3623/1712/14+6100/254
Brefeldin A13.49281.0[M+H]+36245.3/263.30.7211/914/14+6100/2560 Cytochalasin B13.75480.2[M+H]+51462.5/444.50.3323/2310/12+750/2010 Cytochalasin H13.76494.2[M+H]+26434.5/416.5  1.3511/1912/12+750/2030 Cytochalasin C13.98525.2[M+NH4]+31430.5/490.50.3623/1712/6+750/202 Cytochalasin A14.97478.2[M+H]+71460.5/120.20.3823/3912/8+860/2530
Enniatin B315.39629.4[M+NH4]+46196.3/214.30.6041/4110/12+960/200.04
2-Amino-14,16-
di-methyloctadecan-3-ol
16.00314.3[M+H]+41296.5/125.20.0225/2518/6+950/2020
3-Nitropropionic acid  2.93118.0[MÀH]ÀÀ3546.0–À16À3À110025
HC-toxin10.79435.2[MÀH]ÀÀ95184.0/113.10.21À36/À52À13/À3À2100/10012
b-ZOL-glucosid11.95541.3[M+Ac]ÀÀ28319.1/481.10.56À32/À14À15/À11À3100/1001
a-ZOL-glucosid12.92541.3[M+Ac]ÀÀ28319.1/481.10.48À32/À14À15/À11À3100/1000.8
a Retention time.
b Declustering potential.
c Values are given in the order quantifier ion/qualifier ion.
d Intensity of th
e qualifier transition/intensity o
f the quantifier transition.
e Collision energy.
f Cell exit potential.
g Retention time period.
410M.Sulyok et al./Food Chemistry119(2010)408–416
and the corresponding median and maximum concentrations as well as the type of contaminated foo
d commodities are given in Ta-ble2;a distribution of the measured concentrations is given in Figs.1and2.In three fruits(orange,plum,apricot)and three veg-etables(two cucumbers,one tomato),none of the investigated tox-ins was found despite the presence of moulds.
The reported frequency of the different toxins is of limited va-lue for the evaluation of their relevance,as the frequency ranking is biased by the differences in the LODs of the analytes and by the different number of sub-samples that were analyzed for each type of matrix.This is particularly true for the enniatins that were de-tected in all bread and pastry samples at concentrations in the low l g/kg-range,which were partially below the LOD of some of the other analytes.Enniatin contamination of grains in the l g/kg range was already reported by other by Jestoi et al.(2004),and our results confirm recentfindings by Noser, Schmutz,Schmid,and Schneider(2007)regarding contamination offlour and bread.Generally,the enniatin concentrations did not vary significantly between mouldy and non-mouldy parts of bread and pastry samples(one example is depicted in Fig.3), which indicates a more or less homogeneous distribution in the sample deriving from contamination of the raw product grain. We also found enniatins in a cranberry and a date sample at con-centrations of several hundred l g/kg.Both samples were heavily infected by a grey mould and the cranberry sample also contained the Fusarium avenaceum metabolite avenacein
Y.While those two samples would have probably not been consumed due to their visible infection,bread and pastries seem to be a potential source for enniatin uptake.However,to our knowledge it has not been investigated so far whether an uptake of such low concentrations of enniatins has consequences for the human health.A recent study concluded that due to their cytotoxic effects observed in hu-man cell lines at the micromolar level and their lipophilicity that could cause bioaccumulation in animal and human tissue,ennia-tins might represent a potential hazard(Dornetshuber et al., 2007).
Table2
Overview of the number and type of contaminated samples and of the corresponding mycotoxin concentrations.
Analyte n a N b Median conc.(l g/(l g/kg)Type and number of contaminated sub-samples
Enniatin B10041  2.2950Bread(68),nut(6),fruit(5),vegetables(5),jam(5),other(11)
Enniatin B19638233600Bread(67),fruit(6),vegetables(5),jam(4),nut(3),other(11)
Emodin88357.52900Bread(44),fruit(13),vegetables(7),jam(3),nut(2),other(19)
Enniatin A18635  1.51300Bread(61),fruit(5),nut(5),vegetables(3),jam(1),other(11)
Enniatin A77280.231000Bread(60),fruit(5),vegetables(2),nut(1),other(9)
Mycophenolic acid773482078,000Bread(34),vegetables(7),cheese(5),fruit(4),nut(2),jam(5),other(20) Roquefortine C763312084,000Bread(30),fruit(14),vegetables(12),nut(5),cheese(2),jam(1),other(12) Chanoclavine6733311500Bread(40),vegetables(7),cheese(6),fruit(5),nut(4),jam(1),other(4) Ergometrinine57187.6160Bread(52),cheese(4),vegetables(1),
Ergosinine5616  1.2  6.9Bread(56)
Ergosine5616  1.411Bread(56)
Ergotaminine5115  1.810Bread(51)
Ergocryptine5015  3.522Bread(50)
Ergotamine5016  3.523Bread(50)
Ergometrine5015  3.075Bread(52),cheese(3),vegetables(1)
Ergocristinine4915  2.314Bread(48),vegetables(1)
Ergocristine4914  3.518Bread(50)
Ergocryptinine4915  1.8  6.6Bread(50),vegetables(1)
Ergocorninine4815  2.112Bread(50),vegetables(1)
Ergocornine4713  2.716Bread(50)
Deoxynivalenol431462350Bread(42),nut(1)
Tentoxin4115  1.89.3Bread(35),fruit(1),vegetables(1),other(4)
Citrinin41203203,400,000Bread(14),fruit(13),vegetables(5),nut(1),other(8)
Festuclavine31175215,000Bread(21),cheese(3),vegetables(2),nut(2),fruit(1),jam(1),other(1) Meleagrin22121401,000,000Vegetables(6),bread(5),fruit(2),nut(1),other(8)
Beauvericin2111  2.930,000Vegetables(5),jam(2),bread(1),fruit(1),other(12)
Agroclavine179415400Bread(12),cheese(3),vegetables(1),other(1)
Patulin1673000080,000Fruit(11),nut(1),other(4)
Alternariol1610501800Fruit(9),vegetables(3),bread(1),jam(1),other(2)
Alternariolmethyl-ether1499.4650Fruit(6),bread(2),nut(2),jam(1),other(3)
Chatoglobosin A1472100210,000Fruit(9),nut(2),other(3)
Ochratoxin A12726017,000Bread(2),vegetables(2),cheese(1),other(7)
Penitrem A121019007700Bread(5),vegetables(3),fruit(2),other(2)
Deoxynivalenol-3-glucoside962499Bread(9)
3-Nitropropionic acid94420460,000Vegetables(8),other(1)
Paxillin9869700Bread(3),fruit(2),vegetables(1),nut(1),other(2)
Sulochrin931701000Fruit(3),other(6)
Sterigmatocystin95173000Bread(4),cheese(2),vegetables(1),other(2)
Cytochalasin B93886000Bread(5),other(4)
Ergine83  3.418Bread(8)
Fumonisin B26212029,000Vegetable(4),bread(2)
Moniliformin41180000230,000Vegetable(4)
Fumonisin B341310450Vegetable(4)
Fumonisin B14123008500Vegetable(4)
Avenacein Y4316028,000Fruit(2),bread(1),vegetables(1),
Elymoclavine313159Bread(3)
Kojic acid3236005100Fruit(2),vegetable(1)
Zearalenone22640Bread(1),other(1)
Zearalenone-4-glucoside1162,000Other(1)
a Number of positive sub-samples.
b Number of positive samples;in total,87mouldy food samples were investigated and were divided into247mouldy and non-mouldy sub-samples.
M.Sulyok et al./Food Chemistry119(2010)408–416411
In a similar manner,dark and wholemeal bread and pastries were contaminated with each of the five ergopeptides investigated and with ergometrine (and with the corresponding epimers)at the low l g/kg-level,whereas the white bread samples and related products did not contain these toxins.Again,a homogenous distri-bution of the toxins between mouldy and non-mouldy spots was observed (see Fig.3),confirming that the contamination originated from contaminated raw grain products.In contrast to the ennia-tins,the symptoms of ergot alkaloid intoxications are well known,yet there are no limits for their maximum allowed concentration in food.The highest concentration determined in this study was 200l g/kg (sum of all ergot alkaloids)in a sample of dark bread.This is below the proposed guideline limit of 400–500l g/kg for er-got alkaloids in cereals for human consumption,which has re-cently been discussed by Lampen and Klaffke (2006).However,as matrix
effects were most pronounced for this substance class (see previous section),this limit was possibly exceeded in some samples.A potential health hazard deriving from continuous up-take of low levels of ergot alkaloids can therefore not be com-pletely ruled out and as a consequence,further studies on their occurrence in grain products are required.
Beside enniatins and the ergot alkaloids,the compounds found most frequently were chanoclavine,emodin,mycophenolic acid and roquefortine C and their occurrence was less focused on bread and pastry samples.The occurrence of high concentrations of the latter two compounds in mouldy food and cheese is well described (Kokkonen et al.,2005;Rundberget et al.,2004).These two sub-stances seem to be suitable markers for the infection of food by Penicillium species,as we observed concentrations of >100l g/kg of other toxins without the co-occurrence of roquefortine C
or
Fig.1.Ranges of the observed mycotoxin concentrations in the investigated sub-samples (n =number of positive
sub-samples).
Fig.2.Ranges of the observed mycotoxin concentrations in the investigated sub-samples (continued).
412M.Sulyok et al./Food Chemistry 119(2010)408–416

版权声明：本站内容均来自互联网，仅供演示用，请勿用于商业和其他非法用途。如果侵犯了您的权益请与我们联系QQ:729038198，我们将在24小时内删除。