Determination of avoparcin in animal tissues and milk using LC-ESI-MS/MS and tandem-SPE

J. Sep. Sci. 2008, 31, 3871 –3878 K. Inoue et al. 3871

Koichi Inoue1

Yasuomi Mizuno2

Yukiko Yoshimi1

Mari Nunome1

Tomoaki Hino1

Kaname Tsutsumiuchi2

Hisao Oka1, 3

1Department of Physical andAnalytical Chemistry, School ofPharmacy, Kinjo GakuinUniversity, Moriyama-ku,Nagoya, Japan

2College of Bioscience andBiotechnology, ChubuUniversity, Kasugai Aichi, Japan

3Graduate School of HumanEcology, Kinjo GakuinUniversity, Moriyama-ku,Nagoya, Japan

Original Paper

Determination of avoparcin in animal tissues andmilk using LC-ESI-MS/MS and tandem-SPE

A highly sensitive and selective method using LC-ESI-MS/MS and tandem-SPE wasdeveloped to detect trace amounts of avoparcin (AV) antibiotics in animal tissuesand milk. Data acquisition using MS/MS was achieved by applying multiple reactionmonitoring of the product ions of [M + 3H]3+ and the major product ions of AV-a and-b at m/z 637 fi 86/113/130 and m/z 649 fi 86/113/130 in ESI(+) mode. The calculatedinstrumental LODs were 3 ng/mL. The sample preparation was described that theextraction using 5% TFA and the tandem-SPE with an ion-exchange (SAX) and Inert-Sep C18-A cartridge clean-up enable us to determine AV in samples. Ion suppressionwas decreased by concentration rates of each sample solution. These SPE concentra-tion levels could be used to detect quantities of 5 ppb (milk), 10 ppb (beef), and25 ppb (chicken muscle and liver). The matrix matching calibration graphs obtainedfor both AV-a (r A0.996) and -b (r A0.998) from animal tissues and milk were linearover the calibration ranges. AV recovery from samples was higher than 73.3% andthe RSD was less than 12.0% (n = 5).

Keywords: Animal tissues / Avoparcin / Liquid chromatography-tandem mass spectrometry / Milk/ Solid phase extraction /

Received: August 6, 2008; revised: September 10, 2008; accepted: September 11, 2008

DOI 10.1002/jssc.200800446

1 Introduction

Avoparcin (AV) is a macrocyclic glycopeptide antibioticcomplex produced as a fermentation product by a strainof Streptomyces candidus. AV exists in two major forms AV-a and -b (Fig. 1). AV has been used in agriculture to pro-mote the growth of poultry. The structural similarity ofAV to the human glycopeptide antibiotic vancomycinhas raised concerns regarding cross-species antibioticresistance. In Europe, the use of AV as a growth promoterhas been linked to vancomycin-resistant enterococci(VRE) in food animals and in healthy people in the com-munity with no hospital exposure [1, 2]. A recent studyindicated that banning AV from chicken farms in Taiwandecreased the frequency of VRE [3]. The finding of VRE inimported foods, such as chicken meat, however, is animportant problem with respect to human exposuredespite the banning or limitation of AV worldwide [4].

LC methods for the determination of AV have beendescribed. Kadota et al. [5, 6] published a procedure fordetecting AV in chicken muscle using SPE and LC withUV and amperometric detection. The detection limit ofAV in chicken muscle is 0.2 lg/g (0.2 ppm) using LC/amperometric detection [6]. Curren and King [7] reporteda method using pressurized hot water extraction of AVfrom kidney samples with hydrophilic interaction chro-matography (HILIC)/UV with a detection limit of 0.5 lg/g(0.5 ppm). The separation and analysis using micellarEKC (MEKC) has a standard solution detection limit of0.2 lg/mL [8]. However, this MEKC method has not beenreported to be applied for food samples. To the best ofour knowledge, there are few methods available to detectAV in food samples using chromatographic techniques.These analytical methods are not sufficient for therequired validation in accordance with the JapaneseAgricultural Chemical Residues in 2006 [9].

There is a growing interest in the application of LC-MSto ensure food safety [10]. Many LC-MS methods for resi-due analysis of veterinary medicines in foods have beendescribed and almost all use LC-MS/MS. There are noreports of an LC-MS/MS method for detecting AV in foodsamples. The objective of this study was to develop aquantitative LC-MS/MS method that is highly sensitiveand selective for the regulatory analysis of AV residues inanimal tissues and milk.

Correspondence: Dr. Koichi Inoue, Department of Physical andAnalytical Chemistry, School of Pharmacy, Kinjo Gakuin Univer-sity, 2-1723 Omori, Moriyama-ku, Nagoya 463-8521, JapanE-mail: [email protected]: +81-52-798-0982

Abbreviations: AV, avoparcin; HILIC, hydrophilic interactionchromatography; MEV, matrix effect value; VRE, vancomycin-re-sistant enterococci

i 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com

3872 K. Inoue et al. J. Sep. Sci. 2008, 31, 3871 – 3878

2 Experimental

2.1 Chemicals and reagents

The structural formulas of the compounds studied areshown in Fig. 1. AV sulfate (953 lg/mg) was supplied byJapanese Lederle (Tokyo, Japan). AV stock solution(0.5 mg/mL) was prepared by dissolving the appropriateamount of standard in water/methanol (80:20 v/v). Purestandard solutions were prepared by diluting an aliquotof the stock solution in water/methanol (80:20 v/v). HPLC-grade water, methanol, ethanol, acetic acid (99%), formicacid (99%, LC-MS-grade), TFA (98%), sulfuric acid (H2SO4),ammonium hydroxide (25%), and sodium hydroxide(NaOH) were obtained from Wako Chemical (Osaka,Japan). Purified water was obtained from a Milli-Q purify-ing system (Millipore, Bedford, MA). Food samples suchas animal tissues (chicken muscle, liver, and beef muscle)and milk were obtained from a local store in Nagoya,Japan.

2.2 Instrumentation

LC analyses were performed using a Waters Alliance2695 system (Waters, Milford, MA). LC separation wasperformed using a TSK-GEL ODS 100 V column(2.06150 mm, 3 lm: Tosoh, Tokyo, Japan) maintained at408C. The mobile phase consisted of 0.3% formic acid inwater (Solvent A) and 0.3% formic acid in methanol (Sol-vent B). The LC linear gradient was as follows: 3% SolventB at 0 min, 50% Solvent B at 30 min, 95% Solvent B at30.1 min, 95% Solvent B at 35.1 min, and 3% Solvent B at35.2 min with a flow rate of 0.2 mL/min. The injectionvolume was 10 lL. The separated compounds weredetected with a Waters Micromass Quattro Premier triplequadrupole mass spectrometer (Waters). The mass spec-

trometer was operated with ESI(+) mode. The ESI sourceconditions were: capillary voltage of 2.8 kV, extractor of3 V, RF lens of 0 V, source temperature of 1008C, and de-solvation temperature of 4008C. The cone and desolva-tion gas flows were 50 and 950 L/h, respectively, andwere obtained from a nitrogen source (N2 Supplier Model24S, Anest Iwata, Yokohama, Japan). Argon was used asthe collision gas and was regulated at 0.35 mL/h and themultipliers were set to 650 V. The LH resolution 1, HMresolution 1, ion energy 1, LM resolution 2, HM resolu-tion 2, and ion energy 2 were 12.0, 12.0, 0.5, 12.0, 12.0,and 0.8, respectively.

2.3 Extraction of AV from animal tissues

To establish the optimal conditions, the methoddescribed below was used to detect AV in chicken musclesample (chicken breast fillet). The optimal solvent extrac-tion with 5% TFA in water was then used for other animaltissue samples.

The tissue samples (5.0 g) were cut into small pieces,homogenized, and transferred to 50 mL propylene cen-trifuge tubes. These 25 mL of water, 0.2 M H2SO4 in water/methanol (4:6 v/v) [5, 6], water/ethanol (7:3 v/v) [7], 1, 3,and 5% TFA in water, 5% formic acid or 5% acetic acidwas added to the centrifuge tube, the mixture was homo-genized for 1 min using a YELLOW LINE, D125 basic (IKAmJapan K.K., Nara, Japan) and then the mixture was centri-fuged at 6000 rpm (68426g) for 20 min without 0.2 MH2SO4 in water/methanol (4:6 v/v). When 0.2 M H2SO4 wasused, the extraction solution was adjusted to pH 4 with1 M NaOH in water and then the mixture was centri-fuged at 6000 rpm (68426g) for 20 min. The pellet wasthen homogenized and centrifuged three times with25 mL of extraction solution. The obtained supernatantswere applied to an SPE step.

2.4 Extraction of AV from milk

Milk (5 g) was transferred to a 50 mL propylene centri-fuge tube. The sample was acidified with 5 mL of 5% TFAin water, which was stored at 48C. After 10 min, the sam-ple was centrifuged (6000 rpm (68426g), 10 min). Thesupernatant was transferred to another tube and theprocess was repeated three times. This supernatant solu-tion was applied to an SPE step.

2.5 SPE

The AV-a and -b in food samples were pretreated usingfour different SPE cartridges: C18-based solid (InertSepC18-A (500 mg/6 mL); size: 500 mg/6 mL, GL Science,Tokyo, Japan), styrene divinylbenzene polymer (GL-PakPLS-2; size: 270 mg/20 mL, GL Science), and ion exchangemode (InertSep Slim-J SAX; 500 mg, GL Science) before


Figure 1. Structural formulas of AV-a (R = H) and AV-b(R = Cl).

J. Sep. Sci. 2008, 31, 3871 –3878 Liquid Chromatography 3873

LC-MS/MS analysis. In addition, combining the lower sideof ion-exchange (SAX: linked mini-column type) andupper side of InertSep C18-A (normal-cartridge type)clean-up enabled us to determine the various analytesfrom food samples. Before extracting the samples, Inert-Sep C18-A and PLS-2 SPE cartridges were conditioned byeluting 10 mL of methanol followed by 10 mL of distilledwater. The above-described 75 mL of sample solution waseluted through the SPE cartridges. The cartridges werethen washed with 10 mL of water. Methanol (5 mL) wasadded at a low flow rate to elute the compounds thatwere retained on the cartridges. The solutions wereevaporated to dryness at 308C. The sample volumes werethen adjusted to 1.0 mL of water/methanol (80:20 v/v),centrifuged (8000 rpm (121636g), 10 min), and meas-ured by LC-MS/MS. The AV pretreatment protocol usingion-exchange (SAX) and the combination of ion-exchange(SAX) and InertSep C18-A cartridge was as follows: beforeextracting the samples, the cartridge was conditioned byeluting 10 mL of methanol followed by 10 mL of a 1.0%ammonia solution in distilled water. The above-described sample solution was adjusted to pH 8 using a1.0% ammonia solution. This sample solution was thenpassed through the SPE cartridge. The cartridge waswashed with 5 mL of water, followed by 5 mL of metha-nol. Five milliliters of 3.0% formic acid in methanol wasadded at a low flow rate to elute the compounds thatwere retained on the cartridge. The solutions were evapo-rated to dryness at 308C. The samples volumes wereadjusted with water/methanol (80:20 v/v). The obtainedsample solutions were centrifuged (8000 rpm(121636g), 10 min), and measured by LC-MS/MS. Thematrix matching calibration graphs were configuredwith extracted blank samples (chicken and beef muscles,chicken liver, and milk) solutions added AV standardranging from 5, 10, or 25 ng/g to 1 lg/g.

2.6 Method validation of matrix effect

Based on the approaches of Matuszewski et al. [11] and Vil-lagrasa et al. [12], the matrix effects were evaluated bycomparing the MS/MS responses of standard and test sol-ution. Diluted samples (0.1 lg/g; 0.1 ppm: chickenmuscle, beef muscle, chicken liver, and cow milk) wereused to validate the findings obtained with the dilutedstandard solution and to verify the absence of a matrixeffect. The first set of experiments (set 1) was conductedto evaluate the MS/MS response to the injected standard.The second set of experiments (set 2) was conductedusing samples originating from three different lots andspiked after extraction [11]. Moreover, we studied themost appropriate dilution needed for each matrix tominimize the ion suppression effect. The validation dataobtained in the above manner enabled us to determinethe matrix effect value (MEV) for the extraction proce-

dure and dilution by comparing the absolute peak areasfor the target compound obtained in sets 1 and 2. By cal-culating the peak areas obtained using a standard solu-tion in set 1 (S1), and the corresponding peak areas forthe standards spiked after (set 2: S2) extraction, MEV wascalculated as follows: MEV (%) = (S2/S1)6100 for evalua-tion of dilution levels [11]. An MEV of 100% indicates thatthe response of the standard solution and that of thesample were the same and no absolute matrix effect wasobserved. A value greater than 100% indicates ionizationenhancement and a value less than 100% indicates ion-ization suppression.

3 Results and discussion

3.1 LC-MS/MS detection of AV-a and -b

To measure AV-a and -b using the multiple reaction mon-itoring mode, full scan and product ion spectra of theanalytes were investigated under LC conditions. AV-aand -b can be detected under ESI-MS conditions followingLC separation. Curren and King [7] reported using an ana-lytical separation with the HILIC method because theseanalytes are water-soluble. The HILIC method, however,could not resolve AV-a and -b. Moreover, a triethylammo-nium phosphate mobile phase was used in this HILICmethod of AV detection, which is not suitable for ESI-MS/MS ionization. On the other hand, Kadota et al. [5, 6] useda sodium heptane sulfonic acid buffer (pH 4 with ammo-nium hydroxide solution) mobile phase and an RP col-umn based on a C18 column for separation of AV-a and -b.Recently, Ye et al. [13] reported an LC-MS/MS method fordetecting various antibiotics in drinking water using aformic acid mobile phase and an RP column. Therefore,an ESI ionization source with a formic acid-based mobilephase and RP C18 column (TSK-GEL ODS 100V) werechosen for the ionization source and separation of AV-aand -b in this experiment. Full scan spectra of the analy-tes in the ESI(+) showed that the protonated moleculeswere at [M + 2H]2+ and [M + 3H]3+ of AV-a (m/z 956 and 637)and -b (m/z 972 and 649) (Figs. 2A and C). When precursorions of [M + 3H]3+ were used, the major product ions ofAV-a and -b were observed at m/z 637 fi 86, 113, and 130,and m/z 649 fi 86, 113, and 130, respectively, wereobserved (Figs. 2B and D). In this method of LC-MS/MSanalysis of AV, formic acid was required for LC separationof AV-a and -b. The high concentration of formic acid,however, prevented the ionization of AV in ESI(+). There-fore, the concentrations (0.1, 0.3, 0.5, 0.7, and 1.0%) of for-mic acid in the mobile phase for AV-a and -b for ESI peaksresponses were investigated. The result indicated thatthe optimal concentration was 0.3% formic acid inmobile phase. ESI-MS/MS analytical factors, such as conevoltage (V) and collision energy (eV), were then investi-gated to achieve highly sensitive and selective detection



of AV (Fig. 3). Under the conditions described, the reten-tion times of AV-a and -b (1 lg/mL) were 15.16 (RSD:0.09%, n = 10) and 15.89 min (RSD: 0.13%, n = 10), respec-tively. The test for peak area repeatability indicated thatthe RSD (n = 10) of AV-a (10 ng/mL and 1 lg/mL) and -b(10 ng/mL and 1 lg/mL) ranged from 2.11 to 3.37%. Thecalibration graphs obtained for both AV-a (r A0.999) and-b (r A0.999) were linear over the calibration rangeobtained with a 10 ng/mL to 1 lg/mL standard solution.The calculated instrumental LODs were 3.0 ng/mL (30 pg)using MassLynx V. 4.0 software with an S/N of approxi-mately 3. The optimal LC-MS/MS conditions and valida-

tion are summarized in Tables 1 and 2. LC-MS/MS was asensitive and selective technique for detecting traceamounts of AV-a and -b in standard solutions. We usedthis method to detect AV residue in animal tissues andmilk.

3.2 Matrix effects on AV-a and -b in tissue samples

We observed a possible ion suppression of AV-a and -b insamples extracted and concentrated by SPE. Therefore, itis needed to investigate the matrix effect of the analyte


Figure 2. LC-MS/MS spectra ofthe AV standard for ESI (+).(A) MS scan mode of AV-a fromm/z 100 to 1500. (B) Daughterscan mode of AV-a [M + 3H]3+

(m/z 637) from m/z 50 to 1500.(C) MS scan mode of AV-b fromm/z 100 to 1500. (D) Daughterscan mode of AV-b [M + 3H]3+

(m/z 649) from m/z 50 to 1500.

Table 1. LC-MS/MS conditions for AV-a and -b standard solution measurements

Analyte Precursorion(m/z)

Productions(m/z)

Cone voltage(V)

Collision energy(eV)

Instrumental detec-tion limit (ng/mL)

Calibration range(ng/mL)

AV-a 637 86/113/130 20 18 3 10 –1000AV-b 649 86/113/130 20 18 3 10 –1000

Table 2. Measurement repeatability for AV-a and -b standards

Concentration Retention time Retention time Peak area Peak area(RSD:%, n = 10) (RSD:%, n = 10) (RSD:%, n = 10) (RSD:%, n = 10)

AVa AVb AVa AVb

10 ng/mL 0.46 0.49 3.03 3.371 lg/mL 0.09 0.13 2.95 2.11


from various samples for the reliability of LC-MS/MSdetection.

In general, the sample preparation is more extensiveand useful SPE than others techniques. In the SPE con-centrated solution, the concentrations of the target ana-lyte was increased as well as interfering substances forESI ionization source [14]. For the high sensitive andaccurate detection of AV from samples, we evaluated theion suppression of AV subjected to SPE and LC-ESI-MS/MS.The response obtained by AV-a and -b in the spikedextract (0.1 lg/g, 0.1 ppm in chicken muscle, cow milk,beef muscle, and chicken liver) was compared with theresponse of the same analyte prepared in water/metha-nol solution (80:20 v/v). When ion suppression occurs,

the MEV ((spiked extraction response/standardresponse)6100,%) of AV in samples is lower than 100%[11, 12]. Figure 4 shows the MEV levels obtained for vari-ous samples using the combination cleanup of an ion-exchange (SAX) and InertSep C18-A cartridge. The MEVsof AV-a in chicken muscle and liver samples concen-trated 2.5-, 4-, 5-, 8-, and 10-fold by SPE were less than 80%that of unconcentrated samples, whereas the MEV of AV-b from chicken muscle concentrated 2.5-fold was93.6 l 11.6% (Fig. 4). The MEVs of beef meat samples con-centrated four- to ten-fold ranged from 67.5 to 93.3% thatof unconcentrated samples (Fig. 4). On the other hand,the MEVs of milk samples ranged from 91.9 to 102.8%(Fig. 4). This matrix effect on AV was considered to bedifferent from suppression of the analyte response dueto coeluting matrix constituents from each sample. Shaoet al. [15] suggested that the matrix effects of other antibi-otics are different from those of spiked extracts of pig tis-sues such as muscle, kidney, and liver [15]. The present


Figure 3. Investigation of the mass spectrometric param-eters (cone voltage and collision energy) for determination ofAV standard by ESI-MS/MS. The ESI(+)-MS settings were:capillary voltage of 2.8 kV, extractor of 3 V, RF lens of 0 V,source temperature of 1008C, and desolvation temperatureof 4008C. The cone and desolvation gas flows were 50 and950 L/h, respectively. Argon was used as collision gas andwas regulated at 0.35 mL/h and the multipliers were set to650 V. (A) Cone voltages of AV-a and -b. (B) Collisionenergy of AV-a. (C) Collision energy of AV-b.

Figure 4. Comparison between signals obtained from astandard solution and signals from extracts at different con-centration levels for the evaluation of matrix effects on LC-MS/MS. Sample extracts (chicken: chicken muscle, milk:cow milk, beef: beef muscle, liver: chicken liver) were spikedwith 0.1 lg/g (0.1 ppm), and a standard solution at the sameconcentration was prepared (n = 5).


results indicate the need to investigate the relativematrix effects in the analysis of antibiotics in foodsobtained from different sources. Therefore, special atten-tion must be paid to the levels of AV detected in chickenmuscle and liver samples. Our results shown in Fig. 4indicate that the optimal concentration level using SPEwas five-fold for milk, 2.5-fold for beef meat, and 1.0-foldfor chicken muscle and liver using the response of themajor product ions of AV-a and -b at m/z 637 fi 113 andm/z 649 fi 113 (RSD a10%, n = 5). These SPE concentra-tion levels could be used to detect quantities of 5 ppb(milk), 10 ppb (beef), and 25 ppb (chicken muscle andliver). It is the capable combined LC-MS/MS and SPE tech-nique to process decreased ion suppression and allowed

the detection of trace levels of AV in the tissue and milksamples.

3.3 Investigation of sample preparation of AV

Curren and King [7] reported that the extraction of AVfrom kidney sample was developed by accelerated sol-vent system with 30% ethanol at 758C and 50 atm. Forthis extraction to be achieved, these procedures requirespecific instruments for pressurized hot system and alarge amount of solvents [7]. Moreover, this limit of quan-titation was 0.5 ppm, and only kidney sample wasapplied for validation [7]. Solvent extraction wasrequired to obtain complete recovery from spiked ani-mal tissues. A water-based extraction solvent for AV isbetter than an organic solvent because of the water-solu-bility of AV [7]. In this study, we investigated severalextraction solvents (the levels of spiked AV: 0.1 lg/gchicken muscle sample), such as water, 0.2 M H2SO4 inwater/methanol (40:60 v/v) [5, 6], water/ethanol (70:30v/v) [7], 1 and 5% TFA in water, 5% acetic acid, and 5% for-mic acid. Table 3 shows that 5% TFA resulted in higherAV recovery from chicken tissues than other solventextractions. Therefore, this result indicated that 5% TFAextraction of AV from animal tissues is useful for detect-ing trace levels of AV using SPE and LC-MS/MS.

In this study, commonly SPE was performed using anextraction cartridge, because the SPE method is very easyand uses only a small amount of organic solvent. AVrecovery using SPE with the RP mode and the mixed-phase mode cartridges (RP and anion-exchange) wasexamined. The AV was recovered using chicken tissue(5 g) samples spiked with 0.1 lg/g (0.1 ppm) AV. Extrac-tions using the SPE cartridges were performed accordingto the above-described methods. The AV recovery did notexceed 30% when the InertSep C18-A and PLS-2 SPE car-tridges were used. Based on these results, the RP mode isnot sufficient to remove the high concentrations of con-taminating involatile materials present in biologic sam-ples. Therefore, an SPE pre-concentration procedure withRP cartridges may improve AV recovery, but it may alsoincrease the concentrations of interfering high-concen-tration involatile substances in the samples. King et al.[16] reported that involatile materials must be removedfrom the sample to avoid the ionization suppression typi-cally observed with sample extraction. On the otherhand, the anion-exchange mode is highly selective foracidic compounds, thereby providing clean extractsfrom the sample matrix. The ion-exchange (SAX) phasemode cartridge, however, had lower recoveries for AV-spiked samples than did the RP mode cartridge. It may bethat the capacity of the ion-exchange (SAX) phase modecartridge for trapping acidic analytes from samples wastoo small and became saturated by the high matrix con-tamination. As a result, SPE preparation of AV was


Figure 5. LC-MS/MS chromatograms of AV-a (m/z637 fi 113) and -b (m/z 637 fi 113) in chicken and beefmuscle samples for recovery of lower levels. MS condition:ESI(+). Multiple reaction monitoring (MRM) of m/z637 fi 113 for AV-a and m/z 637 fi 113 for AV-b. (A) Stand-ard solution of AV-a (25 ng/mL). (B) Standard solution of AV-b (25 ng/mL). (C) Chicken liver sample spiked at 0.025 lg/g(25 ppb) for recovery of AV-b. (D) Chicken liver samplespiked at 0.025 lg/g (25 ppb) for recovery of AV-b. (E)Chicken muscle sample spiked at 0.025 lg/g (25 ppb) forrecovery of AV-b. (F) Chicken muscle sample spiked at0.025 lg/g (25 ppb) for recovery of AV-b. (G) Blank ofchicken muscle sample for monitoring AV-a (ND: notdetected).


improved using the tandem-combination of an ion-exchange (the lower side) and InertSep C18-A (the upperside) cartridge clean-up. AV recoveries from biologicalsamples were higher than 75% and RSD was less than10% (n = 5). Using the tandem-combination of an ion-exchange and InertSep C18-A, the LC-ESI-MS/MS was asensitive, selective, and accurate method for the determi-nation of AV in animal tissues or milk.

3.4 Validation of the LC-MS/MS and SPE methodfor detecting AV in animal tissues and milk

The matrix matching calibration graphs were configuredusing extracted blank solutions by SPE. The extractedblank samples and calibration ranges were chickenmuscle (25 ng/g–1 lg/g), beef muscle (10 ng/g–1 lg/g),chicken liver (25 ng/g–1 lg/g), and milk (5 ng/g–1 lg/g),respectively. These calibration curves in blank samplesshowed adequate linearity with correlation coefficients(chicken muscle: r2 = 0.997 (AV-a) and 0.998 (AV-b), beefmuscle: r2 = 0.996 (AV-a) and 0.999 (AV-b), chicken liver:r2 = 0.999 (AV-a) and 0.999 (AV-b), milk: r2 = 0.999 (AV-a)and 0.999 (AV-b)). Inter-day variation of the calibrationcurve slopes, measured as the RSD, was less than 10%(n = 5). For quantification we used this matrix matchingcalibration curve obtained from each sample.

For recovery test, the results of the sample extractionof AV using smaller amounts of 5 ppb (milk sample),10 ppb (beef muscle sample), and 25 ppb (chicken muscle

and liver samples) are summarized in Table 4. Thesespike levels were selected for the evaluation of the loweramount of quantitation without ion suppression. Inaddition, chromatograms showing the recovery of AVfrom chicken liver and muscle samples are shown in Fig.5. The overall recoveries ranged from 73.3% (RSD: 6.3%,n = 5) to 103.5% (RSD: 6.5%) for AV from animal tissueand milk samples (Table 4). The recoveries and quantifi-cation levels obtained with our method are much higherfor AV from animal tissues and milk than other chroma-tographic methods. This developed method was appliedto monitoring AV in the commercially available animaltissue and milk (chicken muscle n = 3; beef muscle n = 3;chicken liver n = 1; cow milk n = 3) in Japan. The AV in allsamples was not detected by this analytical method withLC-MS/MS and tandem-SPE.

4 Concluding remarks

A simple, sensitive, and specific LC-MS/MS method wasdeveloped for the quantitative determination of AV-aand -b in animal tissues and milk at levels from 5 to25 ppb. This reliability of the LC-MS/MS and tandem-SPEprocedure was more useful than other chromatographicmethods, and fulfilled the Japanese validation criteriafor the performance of analytical methods and interpre-tation of results.

The authors declared no conflict of interest.


Table 3. Effect of solvents on AV recovery from chicken tissues (n = 5)

Extraction solution Recovery (%) [RSD (%)]

AV-a AV-b

Water 42.9 [9.4] 39.4 [9.5]0.2 M H2SO4 in water/methanol (40:60 v/v) 39.1 [13.0] 33.7 [12.9]Water/ethanol (70:30 v/v) 43.8 [6.3] 38.6 [2.7]1% TFA in water 63.6 [7.1] 60.5 [7.5]5% TFA in water 77.9 [3.0] 79.7 [5.8]5% Acetic acid in water 48.3 [5.8] 43.9 [2.6]5% Formic acid in water 36.1 [7.8] 35.5 [7.3]

AV was recovered from chicken muscle samples (5 g) spiked with a standard solution at 0.1 mg/g. The combination of an ion-exchange (the lower side) and an InertSep C18A (the upper side) cartridge was used for SPE clean-up.

Table 4. Recovery tests of AV in animal tissues and milk samples (n = 5)

Sample Spiked levels(ppb)

AV-a AV-b

Recovery (%) RSD (%) Recovery (%) RSD (%)

Chicken muscle 25 95.5 7.8 82.6 7.1Chicken liver 25 79.2 11.9 73.3 6.3Beef muscle 10 103.5 6.5 99.1 5.2Milk 5 97.6 2.6 93.8 9.5


5 References

[1] Bogaard, A. E., van den Stobberingh, E. E., Int. J. Antimicrob. Agents2000, 14, 327 – 329.

[2] Klare, I., Konstabel, C., Badstubner, D. O., Werner, G., Witte, W.,Int. J. Food Microbiol. 2003, 88, 269 – 290.

[3] Lauderdale, T. L., Shiau, Y. R., Wang, H. Y., Lai, J. F., Huang, I. W.,Chen, P. C., Chen, H. Y., Lai, S. S., Liu, Y. F., Ho, M., Environ. Micro-biol. 2007, 9, 819 – 823.

[4] Ike, Y., Tanimoto, K., Ozawa, Y., Nomura, T., Fujimoto, S.,Tomita, H., Lancet 1999, 353, 1854.

[5] Kadota, M., Imanaka, M., Ogawa, N., Kumashiro, K., Mori, T.,Oka, H., Ikai, Y., Horie, M., Suzuki, S., Nakazawa, H., J. Food Hyg.Soc. Jpn. 1994, 35, 23 – 28.

[6] Kadota, M., Imanaka, M., Hino, S., Nanba, J., Take, S., Nakazawa,H., J. Food Hyg. Soc. Jpn. 2003, 44, 69 – 72.

[7] Curren, M. S., King, J. W., J. Chromatogr. A 2002, 954, 41 – 49.

[8] Lucas, C., Foley, J. P., Ahuja, E. S., Biomed. Chromatogr. 2003, 17,172 – 181.

[9] Introduction of the Positive List System for Agricultural Chemi-cal Residues in Foods, Ministry of Health, Labour and Welfarein Japan (http://www.mhlw.go.jp/english/topics/foodsafety/positivelist060228/index.html).

[10] Pic�, Y., Font, G., Ruiz, M. J., Fern�ndez, M., Mass Spectrom. Rev.2006, 25, 917 – 960.

[11] Matuszewski, B. K., Constanzer, M. L., Chavez-Eng, C. M., Anal.Chem. 2003, 75, 3019 – 3030.

[12] Villagrasa, M., Guillam�n, M., Eljarrat, E., Barcel�, D., J. Chroma-togr. A 2007, 1157, 108 – 114.

[13] Ye, Z., Weinberg, H. S., Meyer, M. T., Anal. Chem. 2007, 79, 1135 –1144.

[14] Dams, R., Huestis, M. A., Lambert, W. E., Murphy, C. M., J. Am. Soc.Mass Spectrom. 2003, 14, 1290 – 1294.

[15] Shao, B., Jia, X., Wu, Y., Hu, J., Tu, X., Zhang, J., J. Rapid Commun.Mass Spectrom. 2007, 21, 3487 – 3496.

[16] King, R., Bonfiglio, R., Fernandez-Metzler, C., Miller-Stein, C.,Olah, T., J. Am. Soc. Mass Spectrom. 2000, 11, 942 – 950.


Determination of avoparcin in animal tissues and milk using LC-ESI-MS/MS and tandem-SPE

Documents

Transcript of Determination of avoparcin in animal tissues and milk using LC-ESI-MS/MS and tandem-SPE