Journal of Chromatography A, 1327 (2014) 90–96
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Determination of nitrofuran metabolites in shrimp by high performance liquid chromatography with fluorescence detection and liquid chromatography–tandem mass spectrometry using a new derivatization reagent Na-Na Du a , Ming-Ming Chen a,b , Liang-Quan Sheng a,b,∗ , Shui-Sheng Chen a,∗∗ , Hua-Jie Xu a , Zhao-Di Liu a , Chong-Fu Song a , Rui Qiao a a b
College of Chemistry and Chemical Engineering, Fuyang Normal College, Fuyang, Anhui 236041, China College of Chemistry and Chemical Engineering, Anhui University, Hefei, Anhui 230039, China
a r t i c l e
i n f o
Article history: Received 28 October 2013 Received in revised form 16 December 2013 Accepted 19 December 2013 Available online 28 December 2013 Keywords: Nitrofuran metabolites HPLC–FLD LC–MS/MS Shrimp
a b s t r a c t A high performance liquid chromatography with fluorescence detection (HPLC–FLD) method for the simultaneous determination of total nitrofuran metabolite residues (furazolidone, furaltadone, nitrofurantoin, and nitrofurazone) in shrimp was developed. The method involves the acid hydrolysis of protein-bound metabolites, followed by the derivatization of the freed metabolites with the new fluorescent derivatization reagent 2-hydroxy-1-naphthaldehyde (HN) and subsequent liquid–liquid extraction (LLE). Separation is achieved on a YMC-Pack Polymer C18 column under alkaline conditions, and the high fluorescence intensity of the derivatives at an emission wavelength Em = 463 nm (Ex = 395 nm) enables, for the first time, their simultaneous determination in shrimp at concentrations as low as 1 g/kg by HPLC–FLD. The method was validated using blank shrimp fortified with all four metabolites at 0.5, 1.0 and 2.0 g/kg. Recoveries were >87% with relative standard deviations of 2sigma(I)]: R1 = 0.048, wR2 = 0.142, R indices (all data): R1 = 0.0583, wR2 = 0.025, S = 1.03, Largest diff. peak and hole (eÅ−3 ): 0.31 and −0.18; ESI+ /MS (positive ion mode) m/z 356.2 [M+1]+ (calcd. for C14 H13 N2 O3 , 356.2). The mass spectrum was obtained using an ACQUITY® TQD triple-quadrupole instrument (Waters Technology, Milford, MA) equipped with an ESI interface operated in the ESI+ mode with a capillary voltage of 3.5 kV, the extractor set at 3.0 V, a source block temperature of 100 ◦ C and a desolvation temperature of 350 ◦ C. The four NF metabolite derivatives (AH–HN, SEM–HN, AOZ–HN and AMOZ–HN) are very stable and readily crystallized from solution with purities greater than 99.5% as determined by HPLC. The stock standard solutions of the derivatives were stored in a cool dark room and considered to be stable for more than 6 months. The standard solutions were analyzed every other month, and the relative deviation was less than 5%. 2.6. Sample preparation An aliquot of shrimp meat (5.0 g) was ground in the mincing machine and weighed into a 50 mL polypropylene tube. Extract solution (10 mL 0.50 M hydrochloric acid/methanol (3:7, v/v)) was immediately added, followed by selected amounts of the working standard solutions. The solution was then vigorously shaken for a couple of minutes and further dispersed by ultrasound for 40 min to enhance the rate of acid hydrolysis of the bound metabolites. Next, KCl (0.5 g) was added and the mixture was sonicated for another 20 min. After centrifugation at 15 ◦ C and 7200 rpm for 20 min, the supernatant was decanted, the derivatization agent was added (200 L of a 25 mM solution in CH3 OH) and the mixture was placed under a stream of nitrogen at 60 ◦ C until the mixed solution was concentrated to 1 mL. Prior to LLE, a 0.1 M solution of potassium dihydrogen phosphate (4 mL) was added to the solution. Ethyl acetate (2 mL × 5 mL) was used for the extraction process, and the organic phases were collected into a 10 mL test tube. The combined organic phase was evaporated to dryness under a stream of nitrogen at 40 ◦ C, and the dry residue was subsequently redissolved in acetonitrile/distilled water (1 mL, 8:2 (v/v)). The resulting solution was then filtered through a 0.22 m nylon filter into an HPLC vial. 2.7. Quantitation Experimental data were calibrated and quantified using the MassLynxTM (version 4.1) software. A calibration curve of the peak area versus concentration (g/kg) was plotted for each analyte, the least-squares regression parameters were calculated for the calibration curve and the concentrations of the test samples were interpolated from the regression parameters. Sample concentrations were determined by linear regression using the formula Y = mX + b, where Y = the peak area, and X = the concentration of
N.-N. Du et al. / J. Chromatogr. A 1327 (2014) 90–96
93
Table 3 MS/MS transitions and optimal conditions used for MS/MS analysis. Compound AMOZ–HN SEM–HN AH–HN AOZ–HN a
Transition
Cone voltage (V)
Collision energy (eV)
356.2 > 225.2a 356.2 > 127.3 230.2 > 170.2 230.2 > 143.2 270.0 > 170.0 270.0 > 101.4 257.2 > 88.3 257.2 > 170.2
30
28 14 29 20 36 20 29 20
30 46 42
Dwell (S)
Acquisition time (min)
0.1
10.80 –11.50 11.90 –12.90 13.00 –13.80 13.80 –14.50
Transitions for quantification are indicated in bold font.
the standard in g/kg. The correlation coefficients for each of the calibration curves were routinely >0.99. 3. Results and discussion The method described in this paper is aim to the determination of metabolite derivatives. During hydrolysis, the side chains of the metabolites, which are bound to protein residues, are released. The fluorescent reagent HN then reacts with the freed side chains of the metabolites to form the corresponding 2-hydroxy-1-naphthyl derivatives. The addition of KCl aided the removal of the proteins and increased the solubility of the analytes in the organic phase. By concentrating the samples to 1 mL under a stable stream of nitrogen at 60 ◦ C, their purity was maintained and oxygen enrichment was avoided. The simple sample treatment process involved LLE with ethyl acetate followed by dissolution of the residue in acetonitrile/water (8:2, v/v). Separation was achieved on a YMC-Pack Polymer C18 column by HPLC–FLD without any additional inconvenient procedures, such as the adjustment of the pH or SPE, which are commonly required for the analysis of NF metabolites in animal tissues [1,14,33]. 3.1. Optimization of the sample preparation conditions To determine the optimum sample preparation procedure, different hydrolysis solutions (0.25, 0.30, 0.35, 0.40, 0.45 and 0.50 M HCl/methanol (3:7, v/v)); quantities of KCl (0, 0.2, 0.4, 0.6, 0.8 and 1.0 g, which was added to remove the proteins and increase the solubility of the analytes in the organic phase); amounts of the 25 mM HN derivatizing agent solution (50, 100, 150, 200, 250 and 300 L) and pH values for extraction (2.0, 3.0, 4.0, 5.0, 6.0 and 7.0, adjusted with 1 M NaOH or 1 M HCl) were investigated, and the response of each compound under the varying conditions was evaluated. Based on the responses of all of the analytes, it was determined that the use of a 0.50 M HCl/methanol (3:7, v/v) solution with 0.5 g added KCl, 200 L of a 25 mM HN solution and a pH of 3.0 for the extraction (achieved without adjusting the pH of the extraction system) gave the best results (Fig. 1). Notably, this procedure is not only time saving but also provides thoroughly cleaned samples.
Fig. 1. Effects of the acid concentration (A), KCl quality (B), NH volume (C) and pH value in extracts (D) on the recoveries of four metabolites derivation.
3.2. Optimization of the HPLC–FLD conditions Fig. 2 shows the separation of the four target analytes by HPLC–FLD for a standard solution at 50 ng/kg (top) and shrimp fortified with a mixture of NF metabolites at 10 g/kg (middle) with a 5.8–15.0 min run time. A YMC-Pack Polymer C18 column (particle size: 6 m) was used with a 1.0 mL/min flow rate. The pH of the mobile phase was varied to determine the optimum conditions, and it was found that the control of pH of the boratebuffered solution used as the mobile phase was critical for achieving proper separation of the analytes under alkaline conditions. In particular, solutions with pH values ranging from 9 to 11 were tested, and it was determined that a pH value of 9.4 for the
Fig. 2. HPLC–FLD chromatograms of blank shrimp (bottom), shrimp fortified with a mixture of nitrofuran metabolites at 10 g/kg (middle), and a mixed nitrofuran metabolite standards at 50 ng/mL (top).
94
N.-N. Du et al. / J. Chromatogr. A 1327 (2014) 90–96
Fig. 3. Typical product ions scan of the m/z 356.2 precursor ion of AMOZ–HN in positive ionization mode.
acetonitrile/10 mM borate-buffered solution (mobile phase B) and 10.4 for the 10 mM borate-buffered solution (mobile phase A) provided the best separation. The retention times of the analytes were 6.06, 9.00, 11.60 and 13.13 min for AH–HN, SEM–HN, AMOZ–HN and AOZ–HN, respectively, and the strongest responsive intensity for the native fluorescence of the 2-hydroxy-1-naphthyl derivatives of the NF metabolites were observed at an excitation wavelength of 395 nm and an emission wavelength of 463 nm. In addition, the possible interference of HN in the shrimp samples was investigated using the same method. An HN standard was mixed with the analytes standard solution and analyzed using the developed method. The retention time of HN was 5.44 min and thus did not overlap with any of the four target analytes.
3.3. Optimization of LC–MS/MS conditions The MS/MS parameters were optimized for 20 L injections of each analyte using the Waters MassLynxTM (version 4.1) software. Both the positive and negative electrospray ionization MRM modes were applied to evaluate the performance of this method. The best response for all of the analytes was obtained in the positive ion mode by monitoring the reaction m/z 356.2 → 225.2 for AMOZ–HN, 230.2 → 170.2 for SEM–HN, 270.0 → 170.0 for AH–HN and 257.2 → 88.3 for AOZ–HN. Table 3 presents the MRM transitions, individual cone voltages and collision energy voltages applied for the analysis. The separation was conducted on an XTerra® MS C18 column (150 mm × 4.6 mm and 3.5 m; Waters, Milford, MA) within 10–15 min. 0.1% formic acid in acetonitrile and 0.1% formic acid in water as the mobile phases. At a flow rate of 300 L/min, the retention times of the analytes were 10.88, 12.39, 13.41 and 14.06 min for AMOZ–HN, SEM–HN, AH–HN and AOZ–HN. Next, tandem MS experiments were performed on AMOZ–HN to obtain valuable information regarding the fragmentation pathway of the 2-hydroxy-1-naphthyl derivative of AMOZ. The full scan spectrum of AMOZ–HN exhibited a prominent ion at m/z 356.2 corresponding to the protonated molecule [M+H]+ . A product ion scan was then performed for selected precursor ions to identify the appropriate fragment ions. Two of the most abundant and stable product ions were chosen, and the cone voltage, collision energy and dwell time were manually optimized for each of the two MS/MS transitions. The 20 eV product ion mode spectrum of AMOZ–HN (m/z 356.2 as the precursor ion) exhibited a base peak at m/z 225.2 with two additional higher mass fragment ions at m/z 182.2 and m/z 127.3 (Fig. 3). Two MS/MS transitions, 356.2 → 225.2 and 356.2 → 127.3, were then chosen for confirmatory analysis of AMOZ–HN, with the former used as the quantifier and the latter as the qualifier. The MS/MS transitions and peak ratios for all of the
2-hydroxy-1-naphthyl derivatives of the NF metabolites are presented in Table 4. The ratio of these two transitions is in agreement with the criteria outlines in the EU Commission Decision 2002/657/EC [42]. Typical ion chromatograms of the blank shrimp and NF metabolite-fortified (1 g/kg) shrimp samples are presented in Fig. 4. 3.4. Method validation The validity of the HPLC–FLD method was also confirmed by comparing the results obtained using the new method to those obtained using an effective LC–MS/MS method. 3.4.1. Linearity, LOD and LOQ Table 5 summarizes the linearity, limits of detection (LOD) and limits of quantitation (LOQ) of the HPLC–FLD and LC–MS/MS methods for the four analytes. Linear calibration curves were constructed from 0.5 to 40 g/kg for each compound using the regression of the peak area versus the concentration of the standard. No effort was made to reach the upper concentration limits of the calibration curves. Both methods exhibited excellent linearity with correlation coefficients (R) from 0.9985 to 0.9998 for the HPLC–FLD method and from 0.9992 to 0.9998 for the LC–MS/MS method. The LOQ for the different analytes were calculated using a signal-to-noise ratio of 10. The LOD calculated using a signal-to-noise ratio of 3 varied from 0.19 g/kg (AOZ–HN) to 0.26 g/kg (AH–HN) with the HPLC–FLD method and 0.19 g/kg (AOZ–HN) to 0.21 g/kg (AH–HN) for the LC–MS/MS method. Both the HPLC–FLD and LC–MS/MS methods showed excellent linear responses for the analytes in the studied concentration range. Notably, the LOD and LOQ values for the four analytes obtained by the HPLC–FLD method proposed in this study and the LC–MS/MS proposed by our group and Mottier et al. [16] are almost same and fall below the MRPL of 1 g/kg established by the EU. 3.4.2. Precision and accuracy The relative standard deviation (RSD, %), accuracy of the HPLC–FLD and LC–MS/MS methods were assessed for the analysis of six replicates of blank shrimp samples and fortified shrimp samples containing 0.5, 1.0 and 2.0 g/kg of a mixture of the standards. Table 6 summarizes the accuracy and RSD values which obtained at each concentration level. The accuracy ranged from 87.4% to 107% for the HPLC–FLD method and 85.6% to 106% for the LC–MS/MS methods, while the RSD values ranged from 3.5% to 8.1% for the HPLC–FLD method and 3.0% to 8.0% for the LC–MS/MS method. These results demonstrated that the values obtained using both methods were within the acceptable range, indicated that the efficient sample preparation process did not influence the outcome,
N.-N. Du et al. / J. Chromatogr. A 1327 (2014) 90–96
95
Table 4 2-Hydroxy-1-naphthyl derivatives of nitrofuran metabolites, corresponding MS/MS transitions, and peak ratios. Analyte
Peak ratioa
MS/MS transition (m/z) (a) 356.2 [M+H] → 225.2 [M+H CO2 C4 H9 NO] (b) 356.2 [M+H]+ → 127.3 [M+H CO2 C10 H7 CH2 N2 O]+ (a) 230.2 [M+H]+ → 170.2 [M+H NH2 CONH2 ]+ (b) 230.2 [M+H]+ → 143.2 [M+H CH2 N2 CONH2 ]+ (a) 270.0 [M+H]+ → 170.0 [M+H C3 H4 N2 O2 ]+ (b) 270.0 [M+H]+ → 101.4 [M+H C10 H7 CNO]+ (a) 257.2 [M+H]+ → 88.3 [M+H C10 H7 CNO]+ (b) 257.2 [M+H]+ → 170.2 [M+H CO2 NHCHCH3 ]+ +
AMOZ–HN SEM–HN AH–HN AOZ–HN a
+
0.17 0.21 0.21 0.24
(a) = Quantifier transition; (b) = qualifier transition, ratio of qualifier to quantifier.
Fig. 4. MRM chromatograms obtained by LC–MS/MS of a blank shrimp sample (right) and a blank shrimp sample fortified at a 1 g/kg level of each derivative of nitrofuran metabolites (left).
Table 5 Regression data and the LOD and LOQ of HPLC–FLD and HPLC–ESI/MS/MS. Compound
HPLC–FLD
AMOZ–HN SEM–HN AH–HN AOZ–HN
HPLC–ESI/MS/MS
R
LOD (g/kg)
LOQ (g/kg)
R
LOD (g/kg)
LOQ (g/kg)
0.9998 0.9985 0.9993 0.9998
0.24 0.23 0.26 0.20
0.63 0.78 0.86 0.66
0.9992 0.9994 0.9996 0.9998
0.12 0.21 0.21 0.19
0.41 0.70 0.71 0.64
Table 6 Accuracy and relative standard deviation of performance data of HPLC–FLD and HPLC–ESI/MS/MS analysis of four nitrofuran metabolites in blank shrimp samples at three fortification levels. Analyte
Fortification level (g/kg)
LC–FLD/LC–MS2 found (g/kg, n = 6)
LC–FLD/LC–MS2 accuracy (%)
LC–FLD/LC–MS2 RSD (%)
AH
0.50 1.00 2.00
0.47/0.43 0.92/0.90 1.99/1.83
93.8/85.6 91.5/89.8 99.3/91.7
7.2/7.9 6.5/7.1 4.1/3.3
SEM
0.50 1.00 2.00
0.47/0.53 0.87/1.02 1.87/2.02
94.7/106 87.4/102 93.7/101
7.5/8.0 5.3/5.5 4.6/3.6
AMOZ
0.50 1.00 2.00
0.53/0.45 0.95/0.99 2.02/1.83
107/89.4 95.3/98.8 101/91.7
8.1/7.0 7.3/6.5 3.5/1.6
AOZ
0.50 1.00 2.00
0.44/0.42 0.99/1.03 1.88/1.96
87.9/85.6 98.9/103 93.8/97.9
5.5/5.3 4.9/4.7 3.7/3.0
96
N.-N. Du et al. / J. Chromatogr. A 1327 (2014) 90–96
Table 7 Comparison of analysis of nitrofuran metabolites in shrimp using HPLC–FLD versus HPLC–ESI/MS/MS (n = 6). Shrimp
SEM HPLC–FLD
1 2 3 4 5 6 7 8 9d a b c d
and development of the 12th Five-year Plan of China (DY125-15-E01); the Natural Science Foundation of Anhui Provincial University (KJ2009A127).
a
ND ND
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Determination of nitrofuran metabolites in shrimp by high performance liquid chromatography with fluorescence detection and liquid chromatography–tandem mass spectrometry using a new derivatization reagent Na-Na Du a , Ming-Ming Chen a,b , Liang-Quan Sheng a,b,∗ , Shui-Sheng Chen a,∗∗ , Hua-Jie Xu a , Zhao-Di Liu a , Chong-Fu Song a , Rui Qiao a a b
College of Chemistry and Chemical Engineering, Fuyang Normal College, Fuyang, Anhui 236041, China College of Chemistry and Chemical Engineering, Anhui University, Hefei, Anhui 230039, China
a r t i c l e
i n f o
Article history: Received 28 October 2013 Received in revised form 16 December 2013 Accepted 19 December 2013 Available online 28 December 2013 Keywords: Nitrofuran metabolites HPLC–FLD LC–MS/MS Shrimp
a b s t r a c t A high performance liquid chromatography with fluorescence detection (HPLC–FLD) method for the simultaneous determination of total nitrofuran metabolite residues (furazolidone, furaltadone, nitrofurantoin, and nitrofurazone) in shrimp was developed. The method involves the acid hydrolysis of protein-bound metabolites, followed by the derivatization of the freed metabolites with the new fluorescent derivatization reagent 2-hydroxy-1-naphthaldehyde (HN) and subsequent liquid–liquid extraction (LLE). Separation is achieved on a YMC-Pack Polymer C18 column under alkaline conditions, and the high fluorescence intensity of the derivatives at an emission wavelength Em = 463 nm (Ex = 395 nm) enables, for the first time, their simultaneous determination in shrimp at concentrations as low as 1 g/kg by HPLC–FLD. The method was validated using blank shrimp fortified with all four metabolites at 0.5, 1.0 and 2.0 g/kg. Recoveries were >87% with relative standard deviations of 2sigma(I)]: R1 = 0.048, wR2 = 0.142, R indices (all data): R1 = 0.0583, wR2 = 0.025, S = 1.03, Largest diff. peak and hole (eÅ−3 ): 0.31 and −0.18; ESI+ /MS (positive ion mode) m/z 356.2 [M+1]+ (calcd. for C14 H13 N2 O3 , 356.2). The mass spectrum was obtained using an ACQUITY® TQD triple-quadrupole instrument (Waters Technology, Milford, MA) equipped with an ESI interface operated in the ESI+ mode with a capillary voltage of 3.5 kV, the extractor set at 3.0 V, a source block temperature of 100 ◦ C and a desolvation temperature of 350 ◦ C. The four NF metabolite derivatives (AH–HN, SEM–HN, AOZ–HN and AMOZ–HN) are very stable and readily crystallized from solution with purities greater than 99.5% as determined by HPLC. The stock standard solutions of the derivatives were stored in a cool dark room and considered to be stable for more than 6 months. The standard solutions were analyzed every other month, and the relative deviation was less than 5%. 2.6. Sample preparation An aliquot of shrimp meat (5.0 g) was ground in the mincing machine and weighed into a 50 mL polypropylene tube. Extract solution (10 mL 0.50 M hydrochloric acid/methanol (3:7, v/v)) was immediately added, followed by selected amounts of the working standard solutions. The solution was then vigorously shaken for a couple of minutes and further dispersed by ultrasound for 40 min to enhance the rate of acid hydrolysis of the bound metabolites. Next, KCl (0.5 g) was added and the mixture was sonicated for another 20 min. After centrifugation at 15 ◦ C and 7200 rpm for 20 min, the supernatant was decanted, the derivatization agent was added (200 L of a 25 mM solution in CH3 OH) and the mixture was placed under a stream of nitrogen at 60 ◦ C until the mixed solution was concentrated to 1 mL. Prior to LLE, a 0.1 M solution of potassium dihydrogen phosphate (4 mL) was added to the solution. Ethyl acetate (2 mL × 5 mL) was used for the extraction process, and the organic phases were collected into a 10 mL test tube. The combined organic phase was evaporated to dryness under a stream of nitrogen at 40 ◦ C, and the dry residue was subsequently redissolved in acetonitrile/distilled water (1 mL, 8:2 (v/v)). The resulting solution was then filtered through a 0.22 m nylon filter into an HPLC vial. 2.7. Quantitation Experimental data were calibrated and quantified using the MassLynxTM (version 4.1) software. A calibration curve of the peak area versus concentration (g/kg) was plotted for each analyte, the least-squares regression parameters were calculated for the calibration curve and the concentrations of the test samples were interpolated from the regression parameters. Sample concentrations were determined by linear regression using the formula Y = mX + b, where Y = the peak area, and X = the concentration of
N.-N. Du et al. / J. Chromatogr. A 1327 (2014) 90–96
93
Table 3 MS/MS transitions and optimal conditions used for MS/MS analysis. Compound AMOZ–HN SEM–HN AH–HN AOZ–HN a
Transition
Cone voltage (V)
Collision energy (eV)
356.2 > 225.2a 356.2 > 127.3 230.2 > 170.2 230.2 > 143.2 270.0 > 170.0 270.0 > 101.4 257.2 > 88.3 257.2 > 170.2
30
28 14 29 20 36 20 29 20
30 46 42
Dwell (S)
Acquisition time (min)
0.1
10.80 –11.50 11.90 –12.90 13.00 –13.80 13.80 –14.50
Transitions for quantification are indicated in bold font.
the standard in g/kg. The correlation coefficients for each of the calibration curves were routinely >0.99. 3. Results and discussion The method described in this paper is aim to the determination of metabolite derivatives. During hydrolysis, the side chains of the metabolites, which are bound to protein residues, are released. The fluorescent reagent HN then reacts with the freed side chains of the metabolites to form the corresponding 2-hydroxy-1-naphthyl derivatives. The addition of KCl aided the removal of the proteins and increased the solubility of the analytes in the organic phase. By concentrating the samples to 1 mL under a stable stream of nitrogen at 60 ◦ C, their purity was maintained and oxygen enrichment was avoided. The simple sample treatment process involved LLE with ethyl acetate followed by dissolution of the residue in acetonitrile/water (8:2, v/v). Separation was achieved on a YMC-Pack Polymer C18 column by HPLC–FLD without any additional inconvenient procedures, such as the adjustment of the pH or SPE, which are commonly required for the analysis of NF metabolites in animal tissues [1,14,33]. 3.1. Optimization of the sample preparation conditions To determine the optimum sample preparation procedure, different hydrolysis solutions (0.25, 0.30, 0.35, 0.40, 0.45 and 0.50 M HCl/methanol (3:7, v/v)); quantities of KCl (0, 0.2, 0.4, 0.6, 0.8 and 1.0 g, which was added to remove the proteins and increase the solubility of the analytes in the organic phase); amounts of the 25 mM HN derivatizing agent solution (50, 100, 150, 200, 250 and 300 L) and pH values for extraction (2.0, 3.0, 4.0, 5.0, 6.0 and 7.0, adjusted with 1 M NaOH or 1 M HCl) were investigated, and the response of each compound under the varying conditions was evaluated. Based on the responses of all of the analytes, it was determined that the use of a 0.50 M HCl/methanol (3:7, v/v) solution with 0.5 g added KCl, 200 L of a 25 mM HN solution and a pH of 3.0 for the extraction (achieved without adjusting the pH of the extraction system) gave the best results (Fig. 1). Notably, this procedure is not only time saving but also provides thoroughly cleaned samples.
Fig. 1. Effects of the acid concentration (A), KCl quality (B), NH volume (C) and pH value in extracts (D) on the recoveries of four metabolites derivation.
3.2. Optimization of the HPLC–FLD conditions Fig. 2 shows the separation of the four target analytes by HPLC–FLD for a standard solution at 50 ng/kg (top) and shrimp fortified with a mixture of NF metabolites at 10 g/kg (middle) with a 5.8–15.0 min run time. A YMC-Pack Polymer C18 column (particle size: 6 m) was used with a 1.0 mL/min flow rate. The pH of the mobile phase was varied to determine the optimum conditions, and it was found that the control of pH of the boratebuffered solution used as the mobile phase was critical for achieving proper separation of the analytes under alkaline conditions. In particular, solutions with pH values ranging from 9 to 11 were tested, and it was determined that a pH value of 9.4 for the
Fig. 2. HPLC–FLD chromatograms of blank shrimp (bottom), shrimp fortified with a mixture of nitrofuran metabolites at 10 g/kg (middle), and a mixed nitrofuran metabolite standards at 50 ng/mL (top).
94
N.-N. Du et al. / J. Chromatogr. A 1327 (2014) 90–96
Fig. 3. Typical product ions scan of the m/z 356.2 precursor ion of AMOZ–HN in positive ionization mode.
acetonitrile/10 mM borate-buffered solution (mobile phase B) and 10.4 for the 10 mM borate-buffered solution (mobile phase A) provided the best separation. The retention times of the analytes were 6.06, 9.00, 11.60 and 13.13 min for AH–HN, SEM–HN, AMOZ–HN and AOZ–HN, respectively, and the strongest responsive intensity for the native fluorescence of the 2-hydroxy-1-naphthyl derivatives of the NF metabolites were observed at an excitation wavelength of 395 nm and an emission wavelength of 463 nm. In addition, the possible interference of HN in the shrimp samples was investigated using the same method. An HN standard was mixed with the analytes standard solution and analyzed using the developed method. The retention time of HN was 5.44 min and thus did not overlap with any of the four target analytes.
3.3. Optimization of LC–MS/MS conditions The MS/MS parameters were optimized for 20 L injections of each analyte using the Waters MassLynxTM (version 4.1) software. Both the positive and negative electrospray ionization MRM modes were applied to evaluate the performance of this method. The best response for all of the analytes was obtained in the positive ion mode by monitoring the reaction m/z 356.2 → 225.2 for AMOZ–HN, 230.2 → 170.2 for SEM–HN, 270.0 → 170.0 for AH–HN and 257.2 → 88.3 for AOZ–HN. Table 3 presents the MRM transitions, individual cone voltages and collision energy voltages applied for the analysis. The separation was conducted on an XTerra® MS C18 column (150 mm × 4.6 mm and 3.5 m; Waters, Milford, MA) within 10–15 min. 0.1% formic acid in acetonitrile and 0.1% formic acid in water as the mobile phases. At a flow rate of 300 L/min, the retention times of the analytes were 10.88, 12.39, 13.41 and 14.06 min for AMOZ–HN, SEM–HN, AH–HN and AOZ–HN. Next, tandem MS experiments were performed on AMOZ–HN to obtain valuable information regarding the fragmentation pathway of the 2-hydroxy-1-naphthyl derivative of AMOZ. The full scan spectrum of AMOZ–HN exhibited a prominent ion at m/z 356.2 corresponding to the protonated molecule [M+H]+ . A product ion scan was then performed for selected precursor ions to identify the appropriate fragment ions. Two of the most abundant and stable product ions were chosen, and the cone voltage, collision energy and dwell time were manually optimized for each of the two MS/MS transitions. The 20 eV product ion mode spectrum of AMOZ–HN (m/z 356.2 as the precursor ion) exhibited a base peak at m/z 225.2 with two additional higher mass fragment ions at m/z 182.2 and m/z 127.3 (Fig. 3). Two MS/MS transitions, 356.2 → 225.2 and 356.2 → 127.3, were then chosen for confirmatory analysis of AMOZ–HN, with the former used as the quantifier and the latter as the qualifier. The MS/MS transitions and peak ratios for all of the
2-hydroxy-1-naphthyl derivatives of the NF metabolites are presented in Table 4. The ratio of these two transitions is in agreement with the criteria outlines in the EU Commission Decision 2002/657/EC [42]. Typical ion chromatograms of the blank shrimp and NF metabolite-fortified (1 g/kg) shrimp samples are presented in Fig. 4. 3.4. Method validation The validity of the HPLC–FLD method was also confirmed by comparing the results obtained using the new method to those obtained using an effective LC–MS/MS method. 3.4.1. Linearity, LOD and LOQ Table 5 summarizes the linearity, limits of detection (LOD) and limits of quantitation (LOQ) of the HPLC–FLD and LC–MS/MS methods for the four analytes. Linear calibration curves were constructed from 0.5 to 40 g/kg for each compound using the regression of the peak area versus the concentration of the standard. No effort was made to reach the upper concentration limits of the calibration curves. Both methods exhibited excellent linearity with correlation coefficients (R) from 0.9985 to 0.9998 for the HPLC–FLD method and from 0.9992 to 0.9998 for the LC–MS/MS method. The LOQ for the different analytes were calculated using a signal-to-noise ratio of 10. The LOD calculated using a signal-to-noise ratio of 3 varied from 0.19 g/kg (AOZ–HN) to 0.26 g/kg (AH–HN) with the HPLC–FLD method and 0.19 g/kg (AOZ–HN) to 0.21 g/kg (AH–HN) for the LC–MS/MS method. Both the HPLC–FLD and LC–MS/MS methods showed excellent linear responses for the analytes in the studied concentration range. Notably, the LOD and LOQ values for the four analytes obtained by the HPLC–FLD method proposed in this study and the LC–MS/MS proposed by our group and Mottier et al. [16] are almost same and fall below the MRPL of 1 g/kg established by the EU. 3.4.2. Precision and accuracy The relative standard deviation (RSD, %), accuracy of the HPLC–FLD and LC–MS/MS methods were assessed for the analysis of six replicates of blank shrimp samples and fortified shrimp samples containing 0.5, 1.0 and 2.0 g/kg of a mixture of the standards. Table 6 summarizes the accuracy and RSD values which obtained at each concentration level. The accuracy ranged from 87.4% to 107% for the HPLC–FLD method and 85.6% to 106% for the LC–MS/MS methods, while the RSD values ranged from 3.5% to 8.1% for the HPLC–FLD method and 3.0% to 8.0% for the LC–MS/MS method. These results demonstrated that the values obtained using both methods were within the acceptable range, indicated that the efficient sample preparation process did not influence the outcome,
N.-N. Du et al. / J. Chromatogr. A 1327 (2014) 90–96
95
Table 4 2-Hydroxy-1-naphthyl derivatives of nitrofuran metabolites, corresponding MS/MS transitions, and peak ratios. Analyte
Peak ratioa
MS/MS transition (m/z) (a) 356.2 [M+H] → 225.2 [M+H CO2 C4 H9 NO] (b) 356.2 [M+H]+ → 127.3 [M+H CO2 C10 H7 CH2 N2 O]+ (a) 230.2 [M+H]+ → 170.2 [M+H NH2 CONH2 ]+ (b) 230.2 [M+H]+ → 143.2 [M+H CH2 N2 CONH2 ]+ (a) 270.0 [M+H]+ → 170.0 [M+H C3 H4 N2 O2 ]+ (b) 270.0 [M+H]+ → 101.4 [M+H C10 H7 CNO]+ (a) 257.2 [M+H]+ → 88.3 [M+H C10 H7 CNO]+ (b) 257.2 [M+H]+ → 170.2 [M+H CO2 NHCHCH3 ]+ +
AMOZ–HN SEM–HN AH–HN AOZ–HN a
+
0.17 0.21 0.21 0.24
(a) = Quantifier transition; (b) = qualifier transition, ratio of qualifier to quantifier.
Fig. 4. MRM chromatograms obtained by LC–MS/MS of a blank shrimp sample (right) and a blank shrimp sample fortified at a 1 g/kg level of each derivative of nitrofuran metabolites (left).
Table 5 Regression data and the LOD and LOQ of HPLC–FLD and HPLC–ESI/MS/MS. Compound
HPLC–FLD
AMOZ–HN SEM–HN AH–HN AOZ–HN
HPLC–ESI/MS/MS
R
LOD (g/kg)
LOQ (g/kg)
R
LOD (g/kg)
LOQ (g/kg)
0.9998 0.9985 0.9993 0.9998
0.24 0.23 0.26 0.20
0.63 0.78 0.86 0.66
0.9992 0.9994 0.9996 0.9998
0.12 0.21 0.21 0.19
0.41 0.70 0.71 0.64
Table 6 Accuracy and relative standard deviation of performance data of HPLC–FLD and HPLC–ESI/MS/MS analysis of four nitrofuran metabolites in blank shrimp samples at three fortification levels. Analyte
Fortification level (g/kg)
LC–FLD/LC–MS2 found (g/kg, n = 6)
LC–FLD/LC–MS2 accuracy (%)
LC–FLD/LC–MS2 RSD (%)
AH
0.50 1.00 2.00
0.47/0.43 0.92/0.90 1.99/1.83
93.8/85.6 91.5/89.8 99.3/91.7
7.2/7.9 6.5/7.1 4.1/3.3
SEM
0.50 1.00 2.00
0.47/0.53 0.87/1.02 1.87/2.02
94.7/106 87.4/102 93.7/101
7.5/8.0 5.3/5.5 4.6/3.6
AMOZ
0.50 1.00 2.00
0.53/0.45 0.95/0.99 2.02/1.83
107/89.4 95.3/98.8 101/91.7
8.1/7.0 7.3/6.5 3.5/1.6
AOZ
0.50 1.00 2.00
0.44/0.42 0.99/1.03 1.88/1.96
87.9/85.6 98.9/103 93.8/97.9
5.5/5.3 4.9/4.7 3.7/3.0
96
N.-N. Du et al. / J. Chromatogr. A 1327 (2014) 90–96
Table 7 Comparison of analysis of nitrofuran metabolites in shrimp using HPLC–FLD versus HPLC–ESI/MS/MS (n = 6). Shrimp
SEM HPLC–FLD
1 2 3 4 5 6 7 8 9d a b c d
and development of the 12th Five-year Plan of China (DY125-15-E01); the Natural Science Foundation of Anhui Provincial University (KJ2009A127).
a
ND ND
