Accepted Manuscript

Title: Determination of nitrofuran metabolites in seafood byultra high performance liquid chromatography coupled totriple quadrupole tandem mass spectrometry

Authors: Noelia M. Valera-Tarifa, Patricia Plaza-Bolanos,Roberto Romero-Gonzalez, Jose L. Martınez-Vidal, AntoniaGarrido-Frenich

PII: S0889-1575(13)00024-0DOI: http://dx.doi.org/doi:10.1016/j.jfca.2013.01.010Reference: YJFCA 2287

To appear in:

Received date: 27-7-2012Revised date: 24-1-2013Accepted date: 27-1-2013

Please cite this article as: Valera-Tarifa, N. M., Plaza-Bolanos, P., Romero-Gonzalez, R.,Martınez-Vidal, J. L., & Garrido-Frenich, A., Determination of nitrofuran metabolites inseafood by ultra high performance liquid chromatography coupled to triple quadrupoletandem mass spectrometry, Journal of Food Composition and Analysis (2013),http://dx.doi.org/10.1016/j.jfca.2013.01.010

This is a PDF file of an unedited manuscript that has been accepted for publication.As a service to our customers we are providing this early version of the manuscript.The manuscript will undergo copyediting, typesetting, and review of the resulting proofbefore it is published in its final form. Please note that during the production processerrors may be discovered which could affect the content, and all legal disclaimers thatapply to the journal pertain.

Page 1 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

1

Original research article1

Determination of nitrofuran metabolites in seafood by ultra high 2performance liquid chromatography coupled to triple 3

quadrupole tandem mass spectrometry4

Noelia M. Valera-Tarifa a, Patricia Plaza-Bolaños a,b, Roberto Romero-González 1, José 5

L. Martínez-Vidal a, and Antonia Garrido-Frenich a,*6

aDepartment of Chemistry and Physics (Analytical Chemistry Area), Research Centre 7

for Agricultural and Food Biotechnology (BITAL), University of Almería, Agrifood 8

Campus of International Excellence, ceiA3, Carretera de Sacramento s/n, E-04120 9

Almería, Spain10

bDepartment of Analytical Chemistry, University of Granada, E-18071, Granada, Spain11

*Corresponding author. Tel: +34 950015985; fax: +34 95001548312

E-mail address: agarrido@ual.es (Antonia Garrido-Frenich).13

Department of Chemistry and Physics (Analytical Chemistry Area), Research Centre for 14

Agricultural and Food Biotechnology (BITAL), University of Almería, Agrifood 15

Campus of International Excellence, ceiA3, Carretera de Sacramento s/n, E-04120 16

Almería, Spain.17

Abstract18

A new analytical method has been developed for the simultaneous determination of 4 19

nitrofuran metabolites in seafood by ultra-high performance liquid chromatography 20

coupled to triple quadrupole tandem mass spectrometry (UHPLC-QqQ-MS/MS). The 21

extraction procedure was based on a simultaneous acidic hydrolysis and derivatization 22

Page 2 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

2

using 2-nitrobenzaldehyde (2-NBA), followed by a solid-phase extraction (SPE). 23

Recovery was studied by spiking blank samples at two concentration levels (1 and 10 24

μg/kg) and recoveries ranged from 73-100% and 79-103%, respectively. Precision 25

values, expressed as relative standard deviation (RSD) were ≤ 19% and ≤ 23% for intra-26

day and inter-day precision, respectively. Linearity was studied in the range 1-50 μg/kg27

and the obtained determination coefficients (R2) were ≥ 0.9900 for all compounds. 28

Limits of detection (LODs) for the derivatized nitrofuran metabolites were 0.5-0.8 29

μg/kg and limits of quantification (LOQs) were established at 1 μg/kg, whereas decision 30

limit (CCα) and detection capability (CCβ) ranged from 1.5 to 2.6 μg/kg and 1.6 to 3.1 31

μg/kg, respectively. Finally, the method was applied to real food samples, but 32

nitrofurans were not found.33

Keywords: Nitrofuran metabolites; Seafood; UHPLC-QqQ-MS/MS; Food safety; 34

Veterinary residues; Safe limit; Antibiotics in the food chain; Food analysis; Food 35

composition36

1 Introduction37

Nitrofurans are broad-spectrum antibacterial drugs, which have been widely used 38

worldwide in veterinary medicine or as feed additives in food-producing animals, 39

mainly for the treatment of gastrointestinal infections (Garrido-Frenich et al., 2009; 40

Vass et al., 2008) and they can be considered as the most used antibacterials in food 41

(Picó and Barceló, 2008). These antibiotics have been utilized in poultry, pigs, cattle, 42

cultured fish and shrimps (Vass et al., 2008). The widespread use of nitrofurans in food-43

producing animals was due to their low cost, availability and high effectiveness against 44

resistant infections. Nevertheless, nitrofurans were prohibited as veterinary drugs in the 45

European Union (EU) (Commission Regulation, 2010) and the United States (US) 46

Page 3 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

3

(FDA, 2002) due to their carcinogenic and mutagenic activity (FAO/WHO, 1993; Vass 47

et al., 2008) and the obvious hazard for human health. In fact, nitrofurans (including 48

furazolidone) has recently been listed by the Commission Regulation (EU) No 37/2010 49

(Commission Regulation, 2010) as prohibited substances for which maximum residue 50

limit (MRL) cannot be established. For this reason, the Joint FAO/WHO Expert 51

Committee on Food Additives has not recommended MRLs for nitrofurans 52

(FAO/WHO, 2010) and these compounds do not appear in the Codex MRLs for 53

veterinary drugs (Codex Alimentarius, 2011). Additionally, a minimum required 54

performance limit (MRPL) has been established for each nitrofuran metabolite at 1 55

μg/kg in poultry meat and aquaculture products by the EU (Commission Decision, 56

2003; Council Regulation, 2002).57

The most important nitrofurans are furazolidone, furaltadone, nitrofurazone and 58

nitrofurantoin and their metabolites, 3-amino-2-oxazolidinone (AOZ), 3-amino-5-59

morpholinomethyl-2-oxazolidinone (AMOZ), semicarbazide (SEM) and 1-60

aminohydantoin (AHD), respectively. Nitrofurans do not persist in edible tissue because 61

they are rapidly metabolized, but their toxic metabolites are strongly bound to proteins 62

and highly stable for long periods (several weeks or even months). Thus, because of the 63

rapid metabolism of nitrofurans, the analysis of these compounds is based on the 64

determination of their main metabolites (Garrido-Frenich et al., 2009).65

In 2009 and 2010, a large number of notifications on veterinary drug residues in the EU 66

reported on the occurrence of nitrofuran metabolites, mainly in crustaceans,67

predominantly shrimps. This figure has significantly increased with respect to 2008 68

(RASFF, 2009; RASFF, 2010), so the development of new analytical methods is 69

required for monitoring purposes to detect the presence of these nitrofuran metabolites 70

Page 4 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

4

in food products. Considering the extremely low MRPL for nitrofurans (1μg/kg), the 71

main challenge is the achievement of very low detection limits (LODs).72

In general, the analysis of nitrofurans requires an acidic hydrolysis of the sample (e.g. 73

diluted hydrochloric acid, HCl) to release the metabolites from the proteins, which is 74

often performed simultaneously with a derivatization step. The extraction of the free 75

metabolites is normally carried out by solid-liquid extraction (SLE), and a subsequent 76

clean-up by liquid-liquid extraction (LLE) or solid-phase extraction (SPE) (Gentili et 77

al., 2005; Samanidou and Evaggelopoulou, 2007; Stolker and Brinkman, 2005; Vass et 78

al., 2008). For SPE clean-up, polymeric and NH2 cartridges have been utilized in 79

different food matrices (Garrido-Frenich et al., 2009), such as meat (Barbosa, et al., 80

2007a; Bock et al., 2007; Finzi et al., 2005; Leitner et al., 2001; Mottier et al., 2005; 81

Verdon et al., 2007; Xia et al., 2008), shrimp (Bock et al., 2007; Chu and Lopez, 2005), 82

fish (Tsai et al.,2010) and crawfish (Ding et al., 2006). Additionally, an extra clean-up 83

of the final extract by LLE with n-hexane has been applied to remove the lipid content 84

in the samples (Tsai et al., 2010).85

Nitrofuran metabolites are determined by liquid chromatography (LC) coupled to 86

fluorescence detection (Barbosa et al., 2007b; Conneely et al., 2003) or mass 87

spectrometry (MS) (Balizs and Hewitt, 2003; Effkemann and Feldhusen, 2004). Due to 88

the high polarity of the metabolites, the retention and separation on reversed-phase 89

columns is unfavourable. Thus, a derivatization stage with 2-nitrobenzaldehyde (2-90

NBA) is strongly recommended (Garrido-Frenich et al., 2009) in order to increase the 91

compounds hydrophobicity. Moreover, the obtained nitrophenyl (NP) derivates show 92

higher molecular masses than the original compounds, improving the detection by MS 93

Page 5 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

5

and minimizing the influence of the MS background noise (Effkemann and Feldhusen, 94

2004).95

The aim of this study is the development of an analytical method for the simultaneous 96

determination of four nitrofuran metabolites, AOZ, AMOZ, AHD and SEM, in seafood, 97

by ultra-high performance liquid chromatography coupled to triple quadrupole tandem 98

MS (UHPLC-QqQ-MS/MS). For this purpose, a simplified extraction method and a 99

derivatization stage have been optimized. Besides, the use of the UHPLC technique, 100

based on the reduction of the particle size of the stationary phase (< 2 µm) or using 101

core-shell particles, is proposed to solve the problem of the time consumed during the 102

chromatographic analysis and to increase the sensitivity, with the objective to obtaining 103

quantification limits (LOQs) equal or lower than the MRPLs established by the EU.104

2 Experimental105

2.1 Materials and reagents106

The analytical standards AOZ, AMOZ, SEM-HCl and AHD-HCl (purities always 107

>99.0%) were obtained from Sigma-Aldrich (Madrid, Spain). 2-NBA was also 108

purchased from Sigma-Aldrich (Note: NBA is a possible mutagen so it is important to 109

avoid inhalation and use only in a chemical fume hood). Stock standard solutions of 110

individual compounds (with concentration between 250-374 mg/L) were prepared by 111

dissolving solid standard in 50 mL of methanol (MeOH), obtained from J.T. Baker 112

(Deventer, The Netherlands). These solutions were stored at -30ºC in the dark until use 113

and they were stable for at least 6 months (Barbosa et al., 2007a; Verdon et al., 2007).114

Acetonitrile (ACN), ethyl acetate (EtOAc) and n-hexane were obtained from Sigma-115

Aldrich (Madrid, Spain), anhydrous magnesium sulphate (97%) and di-sodium 116

Page 6 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

6

hydrogen phosphate (99%) from Panreac (Barcelona, Spain), cyclohexane from 117

Scharlau (Barcelona, Spain), ammonium formate and formic acid were purchased from 118

Fluka (Buchs, Switzerland), Bondesil-C18 from Varian (Palo Alto, CA, USA), and 119

trisodium citrate (99%), sodium chloride (99.5%), anhydrous sodium sulphate (99%), 120

sodium acetate (99%), sodium hydroxide (NaOH, 97%), HCl (37%) and florisil were 121

supplied by J.T. Baker.122

For SPE, OASIS HLB (200 mg/6 cm3) and C18 Sep-Pak (500 mg/6 cm3) cartridges 123

were provided by Waters (Milford, MA, USA).124

Ultrapure water, used for the preparation of all aqueous solutions and mobile phase, was 125

obtained from a Milli-Q Gradient water system (Millipore, Bedford, MA, USA).126

50-mL polypropylene tubes and 2-mL microtubes were available for extraction 127

purposes and 0.20-µm Millex-GN nylon filters were provided by Millipore (Millipore, 128

Carrightwohill, Ireland).129

2.2 Instruments and apparatus130

Chromatographic analyses were performed in an Agilent 1290 Infinity series liquid 131

chromatograph from Agilent (Santa Clara, CA, USA). Separations were carried out 132

using a Kinetex C18 column (50 mm x 2.1 mm, 2.6 µm particle size) from Phenomenex 133

(Torrance, CA, USA). MS analyses were performed in an Agilent (Santa Clara, CA, 134

USA) 6460A triple quadrupole mass spectrometer with a focusing nitrogen Jet 135

Stream™ ion source operating in positive electrospray ionization mode (ESI+). The 136

source parameters were as follows: drying gas temperature, 325ºC; sheath gas 137

temperature, 400ºC; drying gas flow, 10 L/min; sheath gas flow, 12 L/min; nebulizer 138

pressure, 25 psi; and capillary voltage, 4000 V.139

Page 7 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

7

The detection of the analytes was performed using the selected reaction monitoring 140

(SRM) mode (three transitions for each precursor ion). Dynamic SRM acquisition was 141

used and dwell times were set automatically by the software according to the retention 142

times. The different transition reactions and other optimal MS parameters are listed in 143

Table 1.144

Control of the equipment, data acquisition and analysis were carried out using software 145

Mass Hunter workstation (Agilent).146

A Reax-2 rotary agitator from Heidolph (Schwabach, Germany) was used for sample 147

extraction and centrifugations were carried out using a high-volume centrifuge 148

Centronic II from JP Selecta (Barcelona, Spain). SPE extractions were performed with 149

an SPE manifold system supplied by Waters (Milford, MA, USA). Sample evaporation 150

and concentration was performed using a Büchi Syncore line (Flawil, Switzerland) and 151

a Stuart sample concentrator (Stone, Staffordshire, UK) equipped with a block heater.152

2.3 Extraction procedure153

2.5 g of homogenized shrimp sample was weighed into a 50-mL polypropylene tube, 154

and 10 mL of 0.2 M aqueous solution of HCl and 100 μL of 0.1 M 2-NBA of freshly 155

solution prepared in MeOH were added. Then, the mixture was shaken and 156

simultaneous hydrolysis, extraction and derivatization were carried out incubating the 157

sample overnight at 37ºC in the dark. Then, the samples were neutralized by adding 1.5 158

mL of 0.1 M di-sodium hydrogen phosphate and 0.2 mL of 2.5 M NaOH and they were 159

shaken for a few seconds. Subsequently, the tubes were centrifuged at 4000 rpm 160

(1681g) for 5 min and the supernatants were cleaned by SPE using OASIS HLB 161

cartridges (conditioning: 5 mL of ethyl acetate, 5 mL of MeOH and 5 mL of Milli-Q 162

water). After the sample loading (total volume of the supernatant), the cartridges were 163

Page 8 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

8

washed with 5 mL of water followed by 5 mL of 30% MeOH in water. The analytes 164

were eluted with 6 mL of ethyl acetate and the eluate was evaporated to dryness under a 165

gentle N2 stream. The residue was re-dissolved in 1 mL of mobile phase (10 mM 166

ammonium formate/MeOH, 50:50, v/v). Finally, the extracts were filtered through a 167

0.20-µm filter and transferred to a 2-mL vial; 5µL were injected into the 168

chromatographic system.169

2.4 Chromatographic analysis170

Chromatographic analyses were performed using gradient elution and the mobile phase 171

was composed of MeOH (eluent A) and 10 mM aqueous solution of ammonium formate 172

(eluent B). The analysis started with 20% of eluent A, which was linearly increased up173

to 100% in 3.5 min. This composition was held for 1 min. Then, a re-equilibration time 174

of 1 min was included and a total running time of 5.5 min for each sample was obtained. 175

The flow rate was 0.3 mL/min and injection volume was 5 µL. C18 column temperature 176

was maintained at 30ºC.177

2.5 Validation protocol178

Detection of compounds was based on the retention time windows (RTWs), defined as 179

the retention time averages ± three standard deviations of the retention time (RT ± 3SD) 180

obtained analyzing 10 blank samples spiked at 10 μg/kg for each compound (Table 1). 181

The selectivity of the method was tested using control blank samples. The absence of 182

any signal at the same retention time as the selected compounds, when the transitions 183

indicated in Table 1 were monitored, indicated that were no matrix interferences that 184

may give a false positive signal. Furthermore, identification was performed comparing 185

the ratio of the most intense transitions monitored for each compound with those 186

obtained using fortified samples. Table 1 shows the obtained ion ratios and 187

Page 9 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

9

identification was considered reliable if the ratio was within the criteria indicated by EU 188

(Commission Decision, 2002).189

Peak area was used as response and the linearity of the method was evaluated by 190

performing matrix-matched calibration curves using blank shrimp samples spiked 191

before extraction in the range from 1 to 50 μg/kg. Recovery studies were performed 192

using spiked shrimps samples (n = 3) at two levels (1 and 10 μg/kg) with the aim of 193

evaluating the trueness of the method. Precision of the overall method was calculated by 194

performing both intra-day (repeatability) and inter-day (reproducibility) precision 195

studies and expressed as relative standard deviation, RSD. Intra-day precision was 196

studied at the same concentration levels evaluated in the recovery study (1 and 10 197

μg/kg), analyzing three replicates for each level, whereas inter-day precision was 198

evaluated spiking three samples at 1 and 10 μg/kg (n = 3) in different days.199

Furthermore, LODs and LOQs, defined as the lowest concentration of the analyte for 200

which signal-to-noise ratios were 3 and 10, respectively and to complete the validation 201

procedure, decision limit (CCα) and detection capability (CCβ) were calculated using the 202

calibration curve procedure applying the MRPL value (1 g/kg) as MRL (Verdon et al., 203

2006).204

2.6 Samples205

Fresh shrimps and prawns were purchased from local markets located in Almeria 206

province (Spain). Then, samples were minced with a blender and the homogenized 207

material was stored in a freezer at -30ºC until analysis. Shrimp samples showing the 208

absence of the target analytes were used as blank samples in the preparation of 209

calibration standards and recovery studies for the optimization and validation procedure.210

Page 10 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

10

3 Results and discussion211

3.1 UHPLC-MS/MS analysis212

The optimization of the MS parameters (fragmentor voltage and collision energy) was 213

carried out by performing several injections (without LC separation) of an intermediate 214

solution in MeOH of the derivatized nitrofuran metabolites (20 mg/L), using ESI+ as 215

ionization source. The optimization was assisted by Optimizer™ software, which 216

automatically determined the precursor ion and the optimal fragmentor voltage by 217

performing injections of the solution described above in the off-column mode. Then, 218

setting the optimized fragmentor voltage, the precursor ion was fragmented changing 219

the collision energy and the software proposed a number of product ions. Those ions 220

showing higher abundance and m/z ratio were selected (up to a total of three). Besides, 221

the most intense transition was used for quantification purposes and the other two 222

transitions were employed for analyte identification (Table 1).223

In order to optimize the chromatographic conditions, different mobile phases were 224

evaluated (data not shown), utilizing MeOH as organic solvent and several aqueous 225

solutions of formic acid, ammonium acetate and formate at different concentrations. 226

The results showed that the best sensitivity was achieved using MeOH and an aqueous 227

solution of ammonium formate (10 mM), and they were selected as mobile phase 228

components for further experiments. Other parameters such as column temperature (30-229

40ºC), flow rate (0.2-0.4 mL/min), injection volume (5 and 10 µL) and gradient profile 230

were evaluated, setting as optimal conditions those indicated in Section 2.4. Finally, 231

Figure 1 shows an example of UHPLC-MS/MS chromatograms of nitrofuran 232

metabolites at 1μg/kg.233

Page 11 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

11

On the other hand, a study of the stability of the derivatized nitrofuran metabolites was 234

carried out by injecting matrix-matched standards of the derivatized compounds at 50 235

μg kg-1, which were subjected to three different storage conditions: room temperature, 236

5ºC (refrigerator) and -30ºC (freezer). In general, it was observed that the compounds 237

were stable at least 1 week in all storage conditions tested.238

3.2 Optimization of the extraction procedure239

Methods reported in literature are time-consuming, especially due to the necessary 240

derivatization stage. First, three extraction procedures were evaluated in order to 241

optimize the main stages of the procedure: extraction, hydrolisis and derivatization. 242

Experiments were performed using 2.5 g of blank shrimp samples spiked at 100 μg/kg. 243

First, a simultaneous hydrolysis and extraction step was carried out using ACN (1% 244

formic acid), applying a final derivatization stage overnight (T ≈ 37ºC) with 2-NBA 245

(Method 1). On the other hand, a different method was tested performing a 246

simultaneous overnight hydrolysis and derivatization step in the first place (water 1% 247

formic acid and 2-NBA), and after that, extraction was carried out (ACN 1% formic 248

acid) (Method 2). And finally, in Method 3, hydrolysis (water 1% formic acid), 249

extraction (ACN 1% formic acid) and derivatization (2-NBA) were carried out 250

separately in the order listed. The obtained results are shown in Table 2 and in general, 251

the best results were observed carrying out the three main steps separately (Method 3), 252

although recoveries for NP-SEM metabolite were too low in all the cases evaluated 253

(<47%).254

In order to improve the LOQs of the method, a concentration stage (concentration factor 255

= 2) was included in the procedure in a sample concentrator using vacuum and heating 256

(40ºC) in the Syncore line after hydrolysis and extraction (performing both separately), 257

Page 12 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

12

and applying a final derivatization step (Method 4). In general, good peak shapes and 258

high peak intensity were observed, although NP-SEM detection was again problematic 259

in most of cases and NP-AHD recoveries were >120% (Table 2).260

In consequence, a SPE stage was evaluated after the extraction and before the 261

derivatization (spiked samples at 25 μg/kg), using C18 and OASIS HLB cartridges 262

(Method SPE 1), with the aim of applying a simultaneous sample preconcentration and 263

clean-up step. The best results were obtained using OASIS HLB cartridges (Figure 2). It 264

is important to notice that the metabolites were better retained by the sorbent in their 265

derivatized form, and thus the derivatization reaction was applied after the SLE and 266

before the SPE stage (Method SPE 2). However, additional modifications were 267

necessary to improve analyte retention and recovery values (Figure 2). Thus, hydrolysis, 268

extraction and derivatization were carried out simultaneously (adding HCl and 2-NBA) 269

and then, sample pH was adjusted to 7 with di-sodium hydrogen phosphate and NaOH 270

to minimize acid pH influence after hydrolysis in the cartridge retention (Method SPE 271

3). These modifications improved the results because all the derivatized metabolites 272

were effectively retained by the sorbent (Figure 2). Finally, the selected method 273

submitted to validation was Method SPE 3, and the experimental procedure has been 274

indicated in Section 2.3. It should be mentioned that the three different developed 275

methodologies applying SPE as clean-up are summarized in Figure 3.276

It has to be pointed out that relatively similar extraction methods (with hydrolysis, 277

extraction and derivatization steps) have been applied in other reported studies (Xia et 278

al., 2008) but some differences were observed. For instance, a smaller sample size (2.5 279

g) was used in comparison to others studies, such as Xia et al. (2008) in which 5.0 g of 280

meat sample were required. Other studies used even smaller quantities of sample such 281

Page 13 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

13

as Ding et al. (2006), which employed 2 g of sample or Douny et al. (2013), which 282

required 1.0 g of shrimp. A lower amount of sample will reduce the amount of co-283

extracted material in the final extract but it can also affect negatively precision of the 284

method. In this study, 2.5 g was used as a compromise solution with adequate results. In 285

relation to sample extraction, Finzi et al. (2005) and, more recently, Barbosa et al. 286

(2012), Douny et al. (2013) and Radovnikovic et al. (2011), applied a LLE step with 287

ethyl acetate (x 2) after hydrolysis (HCl), derivatization (2-NBA) and pH adjustment, 288

obtaining recoveries between 54-130% and RSD values lower than 20% for 289

repeatability. In the present study any LLE step was needed, reducing sample-handling, 290

solvent consumption and increasing sample throughput.291

With respect to the chromatographic system applied in this study, one of its major 292

advantages is the short analytical run of 5.5 min, which allows an adequate resolution 293

and peak shape in order to separate the analytes from matrix interferences. In this sense, 294

other methods already described for nitrofuran metabolite analysis have used high-295

performance liquid chromatography (HPLC) instead of UHPLC as instrumental 296

technique and they generally describes run times ranging from 13 (Barbosa et al., 297

2007b) to more than 25 min (Bock et al., 2007; Douny et al., 2013; Xia et al., 2008). 298

Despite Radovnikov et al. (2011) also used a UHPLC system, they needed longer 299

running time, 9 min, for complete separation of nitrofuran metabolites. In addition, they 300

also required larger injection volume (20 μL) and higher flow rate values (0.5 mL/min) 301

than this study (flow rate of 0.3 mL/min and injection volume of 5 μL )302

3.3 Method validation303

The optimized method was validated studying linearity, selectivity, trueness (expressed 304

as recovery), intra-day precision, inter-day precision, LODs, LOQs, CCα and CCβ.305

Page 14 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

14

Adequate linearity was obtained for all compounds, with determination coefficients (R2) 306

higher than 0.9900 in all the cases within the assayed range (Table 3). Recovery studies 307

were performed as detailed in section 2.5. Table 3 shows the obtained recovery values 308

ranging from 73% to 100% at 1 μg/kg, and from 79% to 103% at 10 μg/kg. It can be 309

observed (Table 3) that intra-day precision was lower than 20% for the two levels 310

assayed, whereas inter-day precision values were lower than 20%, except for NP-311

AMOZ at 1 μg/kg (23%).312

If the obtained results were compared with previous methods, it can be noted that 313

recoveries values obtained in our study (73-103%) are in good agreement with other 314

values described in literature (Bock et al., 2007, Ding et al., 2006, Tsai et al., 2010) in 315

seafood samples, which ranged from 79 to 110 %. Likewise, repeatability values 316

ranging from 0.1 (Tsai et al., 2010) to 22% (Ding et al., 2006) have been observed in 317

those research works.318

LODs ranged from 0.5 to 0.8 μg/kg and LOQs were established at 1 μg/kg in order to 319

facilitate subsequent quality controls (Table 3). It can be observed that the estimated 320

LODs are lower than the MRPLs for nitrofurans metabolites fixed by the EU in 321

aquaculture animals. Finally, Table 3 also shows that estimated CCα values were within 322

the range of 1.5 (NP-AOZ) to 2.6 μg/kg (NP-SEM), whereas CCβ values ranged from 323

1.6 (NP-AOZ) to 3.1 μg/kg (NP-SEM) (Table 3). With regards to these results, it should 324

be noticed that CCα and CCβ values are higher than the MRPL because, as previously 325

cited, this limit has been used replacing the MRPL in the calculation of these 326

parameters. In this context, it should be mentioned that other studies, which have 327

carried out conventional validations have estimated LOQs and LODs values for 328

nitrofuran metabolites while they have not additionally calculated CCα and CCβ (Finzi et 329

Page 15 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

15

al., 2005; Leitner et al., 2001). Likewise, in other different studies, only CCα and CCβ330

parameters are often estimated for these compounds, for instance in the case of Barbosa 331

et al. (2012),Bock et al. (2007) and Radovnikov et al. (2011) research works. By the 332

contrast, in our study we have estimated all the 4 parameters (LOQ, LOD, CCα and 333

CCβ) in order to validate our method. In this sense, it should be specified that in 334

comparison with the CCα and CCβ values estimated in the most recent works regarding 335

nitrofuran metabolites determination (Barbosa et al., 2012; Douny et al., 2013; 336

Radovnikov et al., 2011), CCα and CCβ obtained in this study are slightly higher, and 337

this can be due to the aforementioned strategy followed to calculate them. However, it 338

should be mentioned that some of the most recent works (Barbosa et al., 2012; 339

Radovnikov et al., 2011), analyzed eggs and plasma samples respectively, instead of 340

seafood, so obtained results are not fully comparable.341

3.4 Analysis of real samples342

17 crustacean samples, including red shrimp and prawn samples were analyzed to 343

evaluate the applicability of the validated method. The analyzed samples include 344

different origin (native and aquaculture), geographic origin (EU and non-EU countries) 345

and format (fresh and frozen samples). Internal quality control was carried out with the 346

aim of checking the quality of the results: a matrix-matched calibration, a reagent blank, 347

a matrix blank and several spiked blank samples were evaluated. The reagent blank was 348

obtained by performing the whole process without sample, with the objective of 349

eliminating possible false positives as a result of contamination in the instrument or 350

solvent used. Spiked samples at two levels of fortification (1 and 10 μg/kg) were used in 351

order to control the extraction efficiency.352

Page 16 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

16

It was found that the analyzed samples were tested negative for nitrofuran metabolites 353

according to the prohibition of these compounds established by the EU.354

4 Conclusions355

An analytical method was developed, optimized and validated for the simultaneous 356

determination of four nitrofuran metabolites in seafood by UHPLC-QqQ-MS/MS 357

according the European legislation (MRPL = 1 μg/kg). Despite the efforts to develop a 358

short method, the complexity of the matrix, the need for derivatization and the low 359

sensitivity showed by the analytes hindered this purpose, and finally only two stages 360

were necessary. The optimized extraction method required a simple and simultaneous 361

hydrolysis, extraction and derivatization combined with a final SPE clean-up stage. In 362

addition, the use of UHPLC-QqQ-MS/MS allows the optimization of analytical 363

methods with shorter analysis times (5.5 min) and it may improve sensitivity and 364

resolution. In relation to the validation parameters, good linearity, recovery, precision, 365

LODs, LOQs, CCα and CCβ values were obtained, indicating that the proposed method 366

could be used in routine analysis or in monitoring programs. In addition, it can be 367

observed that the estimated LODs are lower than the MRPLs for nitrofuran metabolites 368

set by the EU in aquaculture products. Finally, the method has been applied to the 369

analysis of real samples and the presence of nitrofuran metabolites was not detected at 370

levels above the LOQ.371

Acknowledgments372

The authors gratefully acknowledge the Spanish Ministry of Economy and 373

Competitiveness (MINECO-FEDER) for financial support (Project Ref. AGL2010-374

21370). NMVT acknowledges her personal funding from the Research Group 375

Page 17 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

17

“Analytical Chemistry of Contaminants”, University of Almeria. PPB is grateful for 376

personal funding through Juan de la Cierva Program (MINECO-European Social Fund, 377

SMSI-ESF). RRG is also grateful for personal funding through the Ramón y Cajal 378

Program (MINECO-ESF).379

References380

Balizs, G., Hewitt, A. (2003). Determination of veterinary drug residues by liquid 381

chromatography and tandem mass spectrometry. Analytica Chimica Acta, 492, 382

105–131.383

Barbosa, J., Ferreira, M.L., Ramos, F., Noronha da Silveira, M.I. (2007a). 384

Determination of the furaltadone metabolite 5-methylmorpholino-3-amino-2-385

oxazolidinone (AMOZ) using liquid chromatography coupled to electrospray 386

tandem mass spectrometry during the nitrofuran crisis in Portugal. Accreditation 387

and Quality Assurance, 12, 543–551.388

Barbosa, J., Moura, S., Barbosa, R., Ramos, F., Noronha da Silveira, M.I. (2007b). 389

Determination of nitrofurans in animal feeds by liquid chromatography-UV 390

photodiode array detection and liquid chromatography ionspray tandem mass 391

spectrometry. Analytica Chimica Acta, 586, 359–365.392

Barbosa, J., Freitas, A., Mourão, J.L., Noronha da Silveira, M.I., Ramos, F. (2012). 393

Determination of furaltadone and nifursol residues in poultry eggs by liquid 394

chromatography-electrospray ionization tandem mass spectrometry. Journal of 395

Agricultural and Food Chemistry, 60, 4227 – 4234.396

Bock, C., Gowik, P.,Stachel, C. (2007). Matrix-comprehensive in-house validation and 397

robustness check of a confirmatory method for the determination of four 398

Page 18 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

18

nitrofuran metabolites in poultry muscle and shrimp by LC-MS/MS. Journal of 399

Chromatography B, 856, 178–189.400

Chu, P.S., Lopez, M.I. (2005). Liquid chromatography–tandem mass spectrometry for 401

the determination of protein-bound residues in shrimp dosed with nitrofurans. 402

Journal of Agricultural and Food Chemistry, 53, 8934–8939.403

Codex Alimentarius.MRLs for the Compedium of methods of analysis as suitable for 404

support to Codex MRLs. Available on 405

http://www.codexalimentarius.net/vetdrugs/data/index.html?lang=en. (Accessed 406

July 2012).407

Commission Decision 2003/181/EC, 13 March 2003, amending Decision 2002/657/EC 408

as regards the setting of minimum performance limits (MRPLs) for certain 409

residues in food animal origin (2003). Official Journal of the European 410

Communities, L71, 17–18.411

Commission Regulation (EU) No 37/2010 of 22 December 2009 on pharmacologically 412

active substances and their classification regarding maximum residue limits in 413

foodstuffs of animal origin (2010). Official Journal of the European Union, L15, 414

1–72.415

Conneely, A., Nugent, A., O´Keeffe, M., Mulder, P.P.J., van Rhijn, J.A., Kovacsics, L., 416

Fodor, A., McCracken, R.J., Kennedy, D.G. (2003). Isolation of bound residues 417

of nitrofuran drugs from tissue by solid-phase extraction with determination by 418

liquid chromatography with UV and tandem mass spectrometry detection. 419

Analytica Chimica Acta, 483, 91–98.420

Page 19 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

19

Corcia, D., Nazzari, M. (2002). Liquid chromatographic-mass spectrometry methods for 421

analyzingantibiotic and antibacterial agents in animal food products. Journal of 422

Chromatography A, 974, 53–89.423

Council Regulation 2002/1756/EC, 23 September 2002, amending Directive 424

70/524/EEC concerning additives in feedingstuffs as regards withdrawal of the 425

authorisation of an additive and amending Commission Regulation (EC) No 426

2430/1999 (2002). Official Journal of the European Communities, L265, 1–2.427

Ding, T., Xu, J., Shen, C., Wang, K. (2006). Determination of trace level nitrofuran 428

metabolites in crawfish meat by electrospray LC-MS/MS on the Finnigan TSQ 429

Quantum Discovery MAX. LC-GC North America, 24, 49.430

Douny, C., Widart, J., De Pauw, E., Silvestre, F., Kestemont, P., Tu, H.T, Phuong, N.T., 431

Maghuin-Rogister, G., Scippo, M.-L. (2013). Development of an analytical 432

method to detect metabolites of nitrofurans. Application to the study of 433

furazolidone elimination in Vietnamese black tiger shrimp (Penaeus monodon). 434

Aquaculture, 376 – 379, 54 – 58.435

Effkemann, S., Feldhusen, F. (2004). Triple-quadrupole LC-MS-MS for quantitative 436

determination of nitrofuran metabolites in complex food matrixes. Analytical and 437

Bioanalytical Chemistry, 378, 842–844.438

FAO/WHO: Fortieth report of the Joint FAO/WHO Expert Committee on Food. 439

Evaluation of certain veterinary drug residues in food [Internet] (1993).World 440

Health Organization. Geneva. Retrieved July, 2012 from: 441

http://www.who.int/ipcs/publications/jecfa/reports/en/index.html.442

Page 20 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

20

FAO/WHO: Report of the nineteenth session of the codex committee on residues of 443

veterinary drugs in foods [Internet] (2010). Joint FAO/WHO Food standards 444

programme, Codex Alimentarius commission. Retrieved July, 2012 from: 445

http://www.codexalimentarius.net/web/archives.jsp?lang=en.446

FDA. Department of Health and Human Services. Food and Drug Administration, 447

Topical Nitrofurans; Extralabel Animal Drug Use; Order of Prohibition (2002). 448

Federal Register, 67, 5470–5471.449

Finzi, J.K., Donato, J.L., De Nucci, M.S.G. (2005). Determination of nitrofuran 450

metabolites in poultry muscle and eggs by liquid chromatography-tandem mass 451

spectrometry. Journal of Chromatography B, 824, 30–35.452

Garrido-Frenich, A., Plaza-Bolaños, P., Aguilera-Luiz, M.M., Martínez-Vidal, J.L. 453

(2009). Chapter 1, Recent advances in the analysis of veterinary drugs and 454

growth-promoting agents by chromatographic techniques. In T. J. Quintin (Ed.), 455

Chromatography Types, Techniques and Methods (p.p.1-102). New York, NY, 456

USA: Nova Science Publishers, Inc.457

Gentili, A., Perret, D., Marchese, S. (2005). Liquid chromatography-tandem mass 458

spectrometry for performing confirmatory analysis of veterinary drugs in animal-459

food products. Trends in Analytical Chemistry, 24, 704–733.460

Leitner, A., Zöllner, P., Lindner, W. (2001). Determination of the metabolites of 461

nitrofuran antibiotics in animal tissue by high-performance liquid 462

chromatography-tandem mass spectrometry. Journal of Chromatography A, 939,463

49–58.464

Page 21 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

21

Mottier, P., Khong, S.P., Gremaud, E., Richoz, J., Delatour, T., Goldmann, T., Guy, 465

P.A. (2005).Quantitative determination of four nitrofuran metabolites in meat by 466

isotope dilution liquid chromatography-electrospray ionisation-tandem mass 467

spectrometry. Journal of Chromatography A, 1067, 85–91.468

Picó, Y., Barceló, D. (2008). The expanding role of LC-MS in analyzing metabolites 469

and degradation products of food contaminants. Trends in Analytical Chemistry, 470

27, 821–835.471

Radovnikovic, A., Moloney, M., Byrne, P., Danaher, M. (2011). Detection of banned 472

nitrofuran metabolites in animal plasma samples using UHPLC-MS/MS. Journal 473

of Chromatography B, 879, 159 – 166.474

RASFF (The Rapid Alert System for Food and Feed): Annual Report 2009. Directorate 475

General for Health & Consumers, European Commission. Retrieved July, 2012 476

from: http://ec.europa.eu/food/food/rapidalert/docs/report2009_en.pdf.477

RASFF (The Rapid Alert System for Food and Feed): Annual Report 2010. Directorate 478

General for Health & Consumers, European Commission. Retrieved July, 2012 479

from: 480

http://ec.europa.eu/food/food/rapidalert/docs/rasff_annual_report_2010_en.pdf.481

Samanidou, V.F., Evaggelopoulou, E.N. (2007). Analytical strategies to determine 482

antibiotic residues in fish. Journal of Separation Science, 30, 2549–2569.483

Stolker, A.A.M., Brinkman, U.A.Th. (2005). Analytical strategies for residue analysis 484

of veterinary drugs and growth-promoting agents in food-producing animals–a 485

review. Journal of Chromatography A, 1067, 15–53.486

Page 22 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

22

Tsai, C.W., Tang, C.H., Wang, W.H. (2010). Quantitative determination of four 487

nitrofurans and corresponding metabolites in the fish muscle by liquid 488

chromatography-electrospray ionization-tandem mass spectrometry. Journal of 489

Food and Drug Analysis, 18, 98–106.490

Vass, M., Hruska, K., Franek, M. (2008). Nitrofuran antibiotics: a review on the 491

application prohibition and residual analysis. Veterinarni Medicina, 53, 469–500.492

Verdon, E., Hurtaud-Pessel, D., Sanders, P. (2006). Evaluation of the limit of 493

performance of an analytical method based on a statistical calculation of its 494

critical concentrations according to ISO standard 11843: Application to routine 495

control of banned veterinary drug residues in food according to European 496

Decision 657/2002/EC. Accreditation and Quality Assurance, 11, 58–62.497

Verdon, E., Couedor, P., Sanders, P. (2007). Multi-residue monitoring for the 498

simultaneous determination of five nitrofurans (furazolidone, furaltadone, 499

nitrofurazone, nitrofurantoine, nifursol) in poultry muscle tissue through the 500

detection of their five major metabolites (AOZ, AMOZ, SEM, AHD, DNSAH) by 501

liquid chromatography coupled to electrospray tandem mass spectrometry–In-502

house validation in line with Commission Decision 657/2002/EC. Analytica 503

Chimica Acta, 586, 336–347.504

Xia, X., Li, X., Zhang, S., Ding, S., Jiang, H., Li, J., Shen, J. (2008). Simultaneous 505

determination of 5-nitroimidazoles and nitrofurans in pork by high-performance 506

liquid chromatography-tandem mass spectrometry. Journal of Chromatography 507

A, 1208, 101–108.508

509

Page 23 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

23

Figure captions510511

Fig. 1. UHPLC-MS/MS chromatograms of derivatized nitrofuran metabolites in blank shrimp 512samples spiked at 1 µg/kg.513

514Fig. 2. Recovery values (%) obtained when applying different SPE clean-up (shrimp spiked samples 515at 25 µg/kg). Error bars indicated the standard deviation (n =3).516

517Fig. 3. Scheme of the different SPE methodologies used in method optimization: (a) hydrolisis, 518extraction, clean-up and derivatization (C18) in separated stages; (b) separated stages but with 519derivatization prior to clean-up (OASIS HLB); and (c) simultaneous hydrolysis, extraction and 520derivatization and separated clean-up (OASIS HLB).521

522

523

Page 24 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

1

Table 1

UHPLC-MS/MS conditions and retention time windows (RTWs) for nitrophenylderivates of nitrofuran metabolites

Original nitrofuran Metabolite Derivatized metabolite RTW

(min)

Fragmentorv

oltage (V)

Quantification

transition

(collision

energy, eV)

Confirmation

transitions

(collision

energy, eV)

Ion ratio (%)

FurazolidoneMWª= 225.2

AOZ MW = 102.1

NP-AOZ MW = 235.1

1.87-1.97 110 236 >134 (4) 236>104 (16)

236>51(60)

58.7

15.4

Furaltadone MW=324.3

AMOZ MW = 201.2

NP-AMOZ MW = 334.2

2.19-2.31 110 335>291 (4) 335>100 (28)

335>262(10)

34.4

25.0

Nitrofurazone MW = 198.1

SEM MW = 75.1

NP-SEM MW = 208.1

1.90-2.05 50 209>166 (4) 209>192 (4)

209>91(20)

81.1

29.9

Nitrofurantoin MW = 238.2

AHD MW = 115.1

NP-AHD MW = 248.1

1.81-1.92 110 249>104(20) 249>76 (44)

249>134(4)

49.7

189.6

aMW: molecular weight

Table(s)1-3

Page 25 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

2

Table 2

Recovery values (%) obtained at 100 μg/kg for some of the different extraction

procedures evaluated.

ACN extraction

Method 1 Method 2 Method 3 Method 4

Hydrolysis +

Extractionª/

Derivatization

Hydrolysis +

Derivatizationb/

Extraction

Hydrolysis/

Extraction/

Derivatizationc

Hydrolysis/Extraction/

Concentration/Derivatization

Compound % Recovery (% RSDd)

NP-AOZ 106 (29) 36 (3) 70 (3) 62 (12)

NP-AMOZ 102 (16) 29 (1) 104 (2) 103 (4)

NP-SEM N.E.e 13 (21) 47 (43) 24 (267)

NP-AHD 118 (69) 165 (1) 99 (8) 136 (14)

aHydrolysis and extraction steps were carried out simultaneously. bHydrolysis and derivatization stages were applied simultaneously. cHydrolysis, extraction and derivatization steps were carried out separately. dRSD: relative standard deviation (n = 3). eN.E.: not extracted (recovery < 10%).

Page 26 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

3

Table 3

Summary of main validation parameters of the developed UHPLC-MS/MS method

Compound Linearity range

(μg/kg) R

2

Recovery

(RSD intra-day, %)ª RSD inter-day (%)

b

LOD

(μg/kg)

LOQ

(μg/kg)

CCα (μg/kg)

CCβ

(μg/kg)

1 μg/kg 10 μg/kg 1 μg/kg 10 μg/kg

NP-AOZ 1-50 0.9955 82(4) 90(8) 13 15 0.5 1.0 1.5 1.6

NP-AMOZ 1-50 0.9995 73(4) 79(6) 23 5 0.6 1.0 2.0 2.3

NP-SEM 1-50 0.9991 100 (13) 103(15) 3 7 0.6 1.0 2.6 3.1

NP-AHD 1-50 0.9995 80 (10) 99(19) 11 13 0.8 1.0 2.0 2.2 an = 3; RSD values are shown in brackets. bn = 3.

Page 27 of 30

Accep

ted

Man

uscr

ipt

Figure2

Page 28 of 30

Accep

ted

Man

uscr

ipt

Figure3

Page 29 of 30

Accep

ted

Man

uscr

ipt

Figure1

Page 30 of 30

Accep

ted

Man

uscr

ipt

JFCA-D-12-00630 Valera-Tarifa et al.

Highlights

A method for the analysis of nitrofuran metabolites by UHPLC-MS/MS was

developed

Sample treatment includes hydrolysis, derivatization and extraction procedures

The method has been validated and good performance characteristics were

obtained

The method was applied to seafood and nitrofuran metabolites were not found

*Highlights (for review)