Article pubs.acs.org/JAFC
Cite This: J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Atmospheric Solid Analysis Probe and Modified Desorption Electrospray Ionization Mass Spectrometry for Rapid Screening and Semi-Quantification of Zilpaterol in Urine and Tissues of Sheep Shubhashis Chakrabarty,† Weilin L. Shelver, Heldur Hakk, and David J. Smith* USDA-Agricultural Research Service, Edward T. Schafer Agricultural Research Center, Biosciences Research Laboratory, 1616 Albrecht Boulevard, Fargo, North Dakota 58102, United States
Downloaded via UNIV OF SUNDERLAND on October 13, 2018 at 13:55:29 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
S Supporting Information *
ABSTRACT: Ambient ionization mass spectrometric methods including desorption electrospray ionization (DESI) and atmospheric solid analysis probe (ASAP) have great potential for applications requiring real-time screening of target molecules in complex matrixes. Such techniques can also rapidly produce repeatable semiquantitative data, with minimal sample preparation, relative to liquid chromatography−mass spectrometry (LC−MS). In this study, a commercial ASAP probe was used to conduct both ASAP-MS and modified DESI (MDESI) MS analyses. We conducted real-time qualitative and semiquantitative analysis of the leanness-enhancing agent zilpaterol in incurred sheep urine, kidney, muscle, liver, and lung samples using ASAP-MS and MDESI MS. Using ASAP, limits of detection (LOD) and quantitation (LOQ) in urine were 1.1 and 3.7 ng/mL, respectively, while for MDESI MS they were 1.3 and 4.4 ng/mL, respectively. The LODs for tissues were 0.1− 0.4 ng/g using ASAP, and 0.2−0.6 ng/g with MDESI MS. The LOQs of the tissues in ASAP were 0.4−1.2 ng/g and 0.5−2.1 ng/g in MDESI MS. Trace levels of zilpaterol were accurately analyzed in urine and tissues of sheep treated with dietary zilpaterol HCl. The correlation coefficient (R2) between semiquantitative ASAP-MS and MDESI MS results of urine samples was 0.872. The data from ASAP and MDESI MS were validated using LC−MS/MS; urinary zilpaterol concentrations ≥5.0 ng/ mL or tissue zilpaterol concentrations ≥1.5 ng/g were detected by ASAP and MDESI MS, respectively, 100% of the time. Forty samples could be analyzed in triplicate, directly from biological matrixes in under an hour. KEYWORDS: mass spectrometry, MDESI, ASAP, high-throughput screening, zilpaterol, drug residue
■
screening of drug formulations15 and for polycyclic and nitropolycyclic aromatic hydrocarbons in environmental and biological matrixes.16 Similar to ASAP, DESI has been utilized for a wide variety of applications including the analysis of pharmaceuticals,16 in the field of forensic science,17 chemical imaging from tissues,18,19 cancer research,20 metabolomics,21 and analysis of proteins.22 Recent work done by McEwen and group23 demonstrated that by switching between ASAP and DESI on a commercially available heated electrospray ionization (HESI-II) probe, the versatility of an ion source can be significantly improved to allow detection of polar biomolecules, volatiles, and small molecule drugs. The collective body of research suggested an opportunity to explore both ASAP and MDESI MS successively for high-throughput screening using a single commercially available HESI-II probe. Previously, our laboratory had investigated rapid, antibody based screening techniques for small molecules of concern in food animal production including feed additives,24,25 antibiotics,26 animal drugs,27 and environmental contaminants.28 Given their sensitivity and apparent versatility for drug identification and quantitation, we believe that ASAP and MDESI MS
INTRODUCTION Quantitation of contaminants in food animal products by regulatory and/or enforcement officials is mostly conducted using liquid chromatography in conjunction with electrospray ionization mass spectrometry (LC−ESI-MS).1 While LC−MS analyses provide the advantages of specificity, sensitivity, and the possibility of accurate multiresidue sample analysis, it is time-consuming, expensive, and sample preparation is often labor intensive. In the past decade, several ambient ionization techniques including desorption electrospray ionization (DESI), 2 direct analysis in real time (DART),3 and atmospheric solid analysis probe (ASAP)4 have been introduced that have the potential for direct, real-time analysis of target analytes in complex biological matrixes, without prior sample treatment and without chromatographic analyte separation.5−7 The main advantage of ASAP is that it does not require a separate ion source, unlike DESI or DART. With minor modification of existing commercial ESI and APCI ion sources, i.e., mounting a flange holder for the ASAP probe, direct and rapid analysis of a wide range of volatile polar and nonpolar analytes4,8 has been possible. The technique has found application in analysis of pharmaceuticals,4,8 identification of pollen for public health purposes,9,10 analysis of crude oil and synthetic polymers,11,12 detection of rodenticides and heroin in forensic samples,13 and screening for plasticizers in gaskets used for food preparation.14 Moreover, ASAP has been used for quantitation of anabolic steroid esters for © XXXX American Chemical Society
Received: July 24, 2018 Revised: September 11, 2018 Accepted: September 25, 2018
A
DOI: 10.1021/acs.jafc.8b03925 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
Three trace-levels of zilpaterol HCl, formulated onto feed at 0.0075, 0.075, and 0.75 mg/kg, were thereafter fed to sheep for 12 days followed by collection of urine and tissue samples as detailed in the Supporting Information. Sheep were used, in lieu of cattle, as a ruminant model to determine levels of dietary zilpaterol that would cause urine and/or tissues to test positive with available rapid screening technology. Samples from a previous study in which sheep were fed dietary zilpaterol at 6 mg/kg of feed (the label instruction for cattle allow up to 7.5 mg/kg of mixed feed) were also used. The animal studies were approved by the institutional animal care and use committee (IACUC). Preparation of Stock, Working Solutions, and Standard Curves. One milliliter of a 0.5 mg/mL stock solution was prepared by dissolving 0.5 mg of zilpaterol HCl powder in 1 mL of absolute ethanol, and the stock solution was stored at −20 °C until use. From this stock solution, 0.5 mL of matrix-matched working solutions of 5 000 and 10 000 ng/mL zilpaterol HCl were prepared using control sheep urine. Working solutions were used to prepare a matrixmatched standard curve with points at 0, 10, 20, 50, 100, 250, and 500 ng/mL. For tissues, 1 mL working solutions, each containing 10, 100, 1 000, and 10 000 ng/mL zilpaterol HCl, were prepared in acetonitrile. Matrix-matched standard curves were prepared by adding the appropriate volumes of working standards to 200 mg of ground tissue (muscle, liver, kidney, and lung) followed by extraction using 200 μL of ice-cold acetonitrile as described below. Tissue variations mandated that a unique dynamic range be prepared for each matrixmatched standard curve. Collectively, tissue dynamic ranges encompassed 0.5−100 ng/mL of acetonitrile extract representing 0.5 to 100 ng/g of tissue. Standard curves were freshly prepared and run with each sample set. Sample Preparation. Analysts were blinded to sample identity by utilizing a sample coding protocol prior to ASAP and DESI analyses. No sample preparation, other than thawing, was utilized for urine samples. Prior to tissue analysis, approximately 100 mg of ground kidney, liver, lung, or muscle were transferred via plastic spatula, previously calibrated at approximately 100 mg, into 2 mL polypropylene centrifuge tubes, and 200 μL of ice-cold acetonitrile was added to each sample. Tissue aliquots, rather than exact tissue weights, were purposely used to increase the speed of the screening assay. Samples were hand shaken and immediately centrifuged (Beckman Coulter Microfuge 18 Centrifuge, Brea, CA) for 10 min at 1 887g. Supernatants were transferred to prelabeled vials using pipets, vortexed, and analyzed. The same sample preparation procedure was also followed for archived urine and tissue samples,38 except that archived kidney and liver samples were centrifuged with 200 μL of hexane for 10 min at 1 887g, and the hexane supernatant was discarded before tissue extraction with ice-cold acetonitrile. ASAP and MDESI MS Parameters. A Thermo Scientific TSQ Quantum Access MAX mass spectrometer (Thermo Fisher Scientific, San Jose, CA) interfaced with a HESI-II probe on a Thermo IonMax source (Thermo Fisher Scientific, San Jose, CA) was used for the experiments. The IonMax source was modified by incorporating an ASAP probe (M & M Mass Spec Consulting, LLC, Hockessin, DE) onto the opening (protected by glass window) normally used for the installation of the photoionization probe.23 For sample analyses, the closed end of a melting point tube was dipped into urine samples or tissue extracts and introduced to the mass spectrometer using the ASAP probe to position the melting point tube in front of the ion transfer capillary inlet. For ASAP, 3.5 kV was applied to the corona needle to generate a discharge in the positive ion mode. Nitrogen was used as sheath and auxiliary gas (5 and 10 psi, respectively) and was supplied through the HESI-II probe. The gas was heated to 350 °C and the ion transfer capillary inlet temperature was 275 °C. The capillary and tube lens voltages were 30 and 58 V, respectively. For ASAP, solvent flow through the ESI needle was not allowed. For MDESI experiments, the position of the HESI-II probe was critical for obtaining a consistent signal. Therefore, the Z-axis was held at position B, the X axis was held at +0.5, and the Y axis position was held at 1.75 on the micrometer of the IonMax. For MDESI experiments, ionization conditions (vaporization/desolvation gas
technologies hold great potential for use in food animal screening applications. To this end, we have used zilpaterol HCl as a relevant molecule to test that premise in samples derived from animals exposed to trace levels of zilpaterol and at levels expected to produce growth effects. Zilpaterol is a β-agonist approved by the United States Food and Drug Administration (FDA) as a feed additive to increase rate of weight gain, improve feed efficiency, and increase carcass leanness in cattle during the last 20 to 40 days of feeding.29−31 A 3-day preslaughter withdrawal period is mandated in the United States, Canada, Japan, South Korea, and South Africa to allow the depletion of zilpaterol residues below maximum residue levels (MRL). Liver MRLs range between 5 and 22 ng/g while muscle MRLs range between 1 and 10 ng/g32 in countries where zilpaterol use is allowed. For most of the world, the use of zilpaterol in food animals is strictly banned. As a consequence, many trading partners of the U.S. consider the presence of any zilpaterol residue unlawful. Further, zilpaterol use is banned by sponsors of most human competitive sporting events,33 in horse and dog racing,34 and even in farm-animal competitions including county, state, and national livestock shows.35,36 Testing for residual zilpaterol, among other drugs, is routinely done by sporting event organizers, food-safety regulatory agencies, and importers/ exporters. Even with established analytical methods, rapid screening assays for zilpaterol, which provide some degree of specificity and sensitivity, are in demand. The goal of the present work was to determine if ASAP and MDESI ionization techniques could be used for rapid, semiquantitative screening of animal matrixes for animal drug residues using zilpaterol as a prototype drug. If successful, such a method might be useful to determine the magnitude of residue present in an animal and, therefore, have a broader range of application than a solely qualitative assay. Xu et al.37 have demonstrated, in principle, the concept of a quantitative rapid assay of β-agonists with internal extractive electrospray ionization mass spectrometry, using pork tissues fortified with analyte. Xu et al.37 conducted their study with an ion-source that was fabricated in-house. Here we demonstrate the practical use of ASAP and MDESI MS using a commercially available HESI-II probe but utilize urine and tissues obtained from animal feeding trials rather than fortified tissues. Further, we validated results of our ASAP and MDESI MS screening efforts with UPLC−MS/MS analyses of the samples. This study demonstrates the practicality of conducting highthroughput, semiquantitative screening of a feed additive, in real time, with little or no sample preparation, and we believe the principle can be extended to a multitude of other applications.
■
MATERIALS AND METHODS
Chemicals and Supplies. HPLC grade water and acetonitrile were purchased from EMD Millipore Co. (Darmstadt, Germany), and formic acid was bought from Sigma-Aldrich (St. Louis, MO). Capillary tubes (100 mm), glass vials (12 mm × 32 mm) and polypropylene centrifuge tubes (2 mL) were purchased from WilmadLabglass (Vineland, NJ), Thermo Fisher Scientific (Rockwood, TN), and Heathrow Scientific, (Vernon Hills, IL), respectively. Plastic spatulas used for allocating aliquots of ground tissue samples were purchased from Synaptent LLC (Chicago, IL). Zilpaterol HCl was extracted and recrystallized from Zilmax premix (4.8% zilpaterol HCl by weight), and the structure was confirmed by proton and carbon NMR spectroscopy, as described in the zilpaterol extraction procedure provided in the Supporting Information and Figure S-1. B
DOI: 10.1021/acs.jafc.8b03925 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry temperature, ion transfer capillary inlet temperature, capillary and tube lens voltages) were similar to that of ASAP, except that 3.5 kV was applied to the ESI needle instead of the corona discharge needle, and a mixture of acetonitrile and water (1:1) with 0.1% formic acid flowed through the HESI-II probe at 25 μL/min. For both ASAP and MDESI analyses, zilpaterol was analyzed by fragmenting the molecular ion m/z 262 using 24 and 30 V collision energies, respectively, and monitoring the product ion, m/z 185, in the selected reaction monitoring (SRM) mode. Quantitation. All samples, including the standards, were run in triplicate. After the introduction of a sample into the mass spectrometer, a total of 50 scans (scan time 0.25 s/scan) were collected and the average signal intensity of m/z 185 over the 50 scans was used for quantitation. Background was determined by measuring the mean signal intensity of matrix blanks (n = 5). The limit of detection for the instrument signal (iLOD) was calculated as the average of the matrix blank signals (n = 5) plus three standard deviations (SD) of the mean. Detectable zilpaterol residues in test samples occurred when the ion intensity of each sample aliquot (n = 3) exceeded the iLOD derived from matrix blanks. That is, if any one of the triplicate sample aliquots did not give an ion intensity greater than the iLOD, the sample was considered to have no detectable zilpaterol residues. The method LOD (LOD) and method limit of quantification (LOQ) of the samples were calculated as follows: LOD = (3 × SD)/slope of the standard curve, where SD represents standard deviation of the matrix blank run 5 times in the MS. LOQ = (10 × SD)/slope of the standard curve, where SD represents the standard deviation of the matrix blank run 5 times in the MS. For urine samples, the LOD and LOQ were calculated from a combination of 20 runs of blanks. When high residue concentrations (approximately >100 ng/mL) were encountered during the ASAP analysis, carryover was avoided by running three blank urine samples before analyzing the subsequent test samples. Sample carryover was not observed during urine or tissue analysis with MDESI MS analysis. Urine Preparation and Instrument Conditions for LC−MS/ MS Analysis. A Waters Acquity Ultra Performance Liquid Chromatograph interfaced with a Waters triple quadrupole mass spectrometer (Waters, Milford, MA) was used for the analysis of urine samples. Samples were prepared for LC−MS/MS analysis using the solid phase extraction (SPE) procedure and analyzed using the LC− MS instrument parameters reported by Shelver et al.39 LC−MS/MS was selected as a validating analytical method because of its wide use by regulatory agencies and because archived samples were available for analysis that had previously been measured using LC−MS/MS. A minor modification in cone voltage (40 V) was used and the collision energies for the transition of 244 (M−OH)+ > 202, 244 > 185 and 244 > 157 were 15, 20, and 30 V respectively.
were able to semiquantitatively determine zilpaterol concentrations directly from raw urine or after minimal processing (∼15 min total) of tissues collected from sheep fed supplemental rations containing zilpaterol HCl. Urine analyses were conducted rapidly with a turnaround time less than a minute per sample in triplicate; essentially no sample preparation was needed. For tissues, minimal processing involving protein precipitation and centrifugation was required, but such processing could be completed in 60 min or less for a sample set involving 18 animals. Given the requirement for manual sample introduction and individual data collection, a total of 40 samples could be successfully introduced into the mass spectrometer and analyzed in an hour. The study suggests that the aforementioned MS techniques can be utilized for rapid screening and/or semiquantitative zilpaterol analysis from live animals or carcasses, in real time without the need for rigorous sample processing as mandated for LC−MS/MS analyses. Optimization of MS Parameters for ASAP and MDESI MS. Zilpaterol ionized very well in positive ion ASAP and MDESI MS (Figure S-2). The fragmentation pattern of zilpaterol is shown in Figure 1A with MS/MS spectra derived from ASAP and MDESI MS shown in Figure 1B,C, respectively. Using optimized collision energies of 24 V for ASAP and 30 V for MDESI MS, zilpaterol (m/z 262) yielded major product ions of m/z 244, 202, 185, and 157 in both methods. MDESI MS produced additional fragments of m/z 116 and 120. The fragment at m/z 185 was the base peak for both methods and was therefore chosen for the study instead of fragments m/z 242 and 202. In studies by Sung et al.40 and Suo et al.,41 fragments at m/z 244 and 185 were used with LC−MS for detecting zilpaterol due to the higher relative sensitivity of these fragments. Shelver et al.38 used m/z 244, 202, and 185 for quantification of zilpaterol in tissues and urine using LC−MS. For ASAP there is a general consensus that complete analyte vaporization to produce high signal intensity without thermal degradation should be the main objective.4,15,42 In our study ASAP desolvation gas temperature was accordingly optimized. A vaporization/desolvation gas temperature of 350 °C produced the highest signal intensity for fragment m/z 185 without analyte degradation apart from the loss of a hydroxyl group (to produce m/z 244). The loss of the benzylic hydroxyl group was also observed in ESI where a low desolvation gas temperature of 50 °C was used, along with a capillary ion transfer tube temperature of 275 °C (Figure S-3). A vaporization/desolvation temperature of 350 °C was also used for MDESI experiments since high signal intensities of m/ z 185 were produced at that temperature. A solvent flow rate of 25 μL/min of acetonitrile and water (1:1) with 0.1% formic acid produced better signal intensity (Figure S-4) relative to solvent flow rates of 2−5 μL/min that are typically employed in conventional DESI experiments.6 Our DESI conditions also differed from the McEwen group23 who did not use a high desolvation gas temperature. We speculate that high temperature and voltage in the HESI-II probe produces a stream of gaseous ions that causes analyte desorption and ionization from the sample matrix located on the melting point tube. Such a phenomena is presumably similar to desorption atmospheric pressure chemical ionization (DAPCI), a variant of DESI where the ionized solvent vapor helps in desorption6,43,44 but which utilizes a corona discharge to
■
RESULTS AND DISCUSSION In this study we conducted high-throughput screening of zilpaterol HCl using ASAP and MDESI MS from a commercial HESI-II probe without changing the ion source. The instrument parameters were first optimized, the day-to-day repeatability and linearity of standard curves were determined, followed by semiquantitative analyses of the drug in urine and tissues after determining its LOD and LOQ in the respective biological matrixes. We note that the assay is semiquantitative because matrix differences among animals were not minimized using typical analyte extraction and/or solid phase extraction techniques and that sample volumes introduced into the mass spectrometer were not standardized. After analysis, the data obtained from urine and tissue samples using ASAP and MDESI MS were correlated with LC−MS/MS data obtained from the same samples. In addition, archived samples that were previously evaluated using HPLC and LC−MS/MS were analyzed using ASAP and MDESI MS. The MS techniques C
DOI: 10.1021/acs.jafc.8b03925 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
samples consecutively in a single run. The data clearly demonstrated differences in signal intensity of control and incurred samples containing zilpaterol. For quantification of residues, samples were manually analyzed individually in triplicate with each subaliquot analysis and ion intensity considered independent of other subsamples. Interday Repeatability and Linearity of Standard Curves. During method optimization, an across-day variation in signal intensity of samples was observed. With urine, relative standard deviation (RSD) of signal intensities of 10, 20, 50, 100, 250, and 500 ng/mL urine standards across 4 days was 17.7% in ASAP and 12.6% in MDESI MS. The RSDs of zilpaterol standards fortified into kidney, muscle, liver, and lungs (4−6 concentrations in the range of 0.5−100 ng/mL) across 4 days were 27.7, 21.3, 19.8 and 19.6%, respectively, in ASAP MS. Similarly, in MDESI MS the RSDs were 29.5, 23.7, 26.1, and 13.1% for kidney, muscle, liver, and lungs, respectively. Inexact delivery of a consistent sample volume into the mass spectrometer on a day-to-day and even sample-to-sample basis very likely contributed to the relatively high variance relative to typical LC−MS analyses, but exact causes of the variation have not been explored. Regardless of the sample-tosample variation, the assay was able to return very rapid results with little ambiguity with respect to the order of magnitude of the residue. Such data can provide substantial information about the status of exposure to a drug. For example, animals on active feeding regimens with β-agonists will typically excrete hundreds to thousands of nanograms of zilpaterol per mL of urine,38,39,46 whereas animals exposed to trace doses or in active withdrawal from the drug will excrete substantially lesser amounts, in the ng/mL range.25,47 For tissues, LODs were substantially lower than typical tolerances, indicating that atmospheric ionization techniques have sufficient sensitivity for regulatory or trade screening applications. Regardless of the day-to-day signal intensity variation, the linearity of standard curves in ASAP and MDESI MS were excellent (R2 > 0.98), especially for what would have to be considered a semiquantitative method. The mean R2 of standard curves prepared in urine across 4 days was 0.993 for ASAP and 0.995 for MDESI MS, respectively. The mean slopes of standard curves of urine samples were 14.55 and 17.26 in ASAP and MDESI MS, respectively. For tissue samples, mean regression coefficients ranged from 0.985 to 0.994 using ASAP and 0.981 to 0.990 with MDESI MS while mean slopes ranged from 99.27 to 183.88 in ASAP and from 104.38 to 143.98 in MDESI MS (Table S-1). Regardless of the absolute across-day signal variation at a given zilpaterol concentration, regression analysis of urine standard curve data combined across analysis day produced correlations of 0.994 and 0.996 in ASAP and MDESI MS, respectively (Figure S-6). Similarly, the regression analyses of standard curves for tissue samples yielded excellent linearity (R2 > 0.99) in both MS techniques (Figure S-7). Within-day standard curves used to quantify the urine and tissue samples in ASAP and MDESI MS are shown in Figures S-8 and S-9. Limit of Detection and Quantification. The ASAP LOD and LOQ for urine was 1.1 and 3.7 ng/mL, respectively, while for MDESI MS the urine LOD and LOQ was 1.3 and 4.4 ng/ mL, respectively. Although more sensitive urine assays are available for zilpaterol (0.1 ng/mL),48 the advantage of ASAP or DESI techniques is speed at which an order of magnitude response may be obtained without cleanup steps such as solid phase extraction, providing a very rapid technique to screen
Figure 1. Fragmentation pattern of zilpaterol in positive ion mode. (A) Structural representation of fragmentation pattern. The molecular ion [M + H]+, m/z 262, produced major product ions of m/z 244 after the loss of hydroxyl group, m/z 202 after the loss of hydroxyl group and the isopropyl group, and m/z 185 after losing the hydroxyl along with isopropylamine. (B) The MS/MS spectra of 10 μg/mL zilpaterol spiked into urine in SRM mode in ASAP and (C) in MDESI MS. Collision energy 24 and 30 V were used in ASAP and MDESI MS, respectively.
ionize solvent vapor and the gaseous ions help in desorption ionization of samples.6,45 Quantification of Zilpaterol from Urine and Tissue Samples. Signal Comparisons. Figure 2 demonstrates that signal intensity was proportional to zilpaterol concentration and that blank sample matrix could be easily distinguished from urine containing both low or high zilpaterol concentrations (∼5−500 ng/mL), regardless of technique. Similarly, various concentrations of zilpaterol (∼2−100 ng/g) were detected in tissue samples as presented in Figure S-5. Such sensitivity is more than adequate to detect zilpaterol in urine of animals being actively fed zilpaterol38,39,46 or for animals fed label doses and provided up 20 days of withdrawal, depending upon dose and species.38,39 With background signal variance defined, ion intensities of control tissue samples could be clearly differentiated from samples with low concentrations (2 ng/g) of zilpaterol. Data shown in Figure 2 and Figure S-5 were collected sequentially by manually analyzing several D
DOI: 10.1021/acs.jafc.8b03925 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
Figure 2. Ion chromatograms of zilpaterol in consecutively run sheep urine samples. Blank urine and urine samples containing high concentrations of zilpaterol are shown (A) using ASAP-MS, (B) using MDESI MS, while urine samples containing low and high concentrations of zilpaterol are shown (C) using ASAP and (D) using MDESI MS.
targeting animal health drugs, Xu et al.37 quantified the βagonists ractopamine, tulobuterol, salbutamol in spiked pork tissue samples using internal extractive electrospray ionization and reported part per trillion level LODs ranging from 0.002 to 0.006 μg/kg from a linear trap MS. At this point, it must be noted that ASAP-MS has not been used extensively for analysis of compounds in tissue extracts while DESI-MS is mostly used for chemical imaging of tissue samples. Comparison of ASAP, MDESI MS, and LC−MS/MS Data. In developing ASAP and MDESI as potential screening tools for zilpaterol, it was incumbent to compare the rapid screening results with a standard technique with wide use and acceptance. To this end, LC−MS/MS was employed as the validating standard because of its acceptance by regulatory agencies such as the USDA FSIS.52 Figure 3 shows a comparison of ASAP and MDESI MS analyses of urine from four sheep treated with 1/10th (0.75 mg/kg feed) the label dose (7.5 mg/kg feed) of zilpaterol HCl for 6 consecutive days and from the same sheep after a withdrawal period of 1 day. Measured zilpaterol concentrations in urine collected after 6 days of treatment were very similar in both ASAP and MDESI MS analysis, ranging from 163 to 445 ng/mL using ASAP and from 221 to 437 ng/mL using MDESI MS. At lower zilpaterol concentrations, i.e., after a 1-day withdrawal period, both techniques returned comparable zilpaterol concentrations (21.5−36.5 ng/mL in ASAP and 15.2−42.3 ng/mL in MDESI MS). A mean difference of 19.0 and 4.8 ng/mL in zilpaterol concentrations was observed between ASAP and MDESI MS techniques, respectively, for these treatment and withdrawal days. Since the ASAP and MDESI MS methods we employed were inherently semiquantitative, absolute values returned by either technique were perhaps not as important as the magnitude of residue present. In this case, either method was capable of repeatedly distinguishing the order of
samples. Wang et al.49 used ASAP to measure zilpaterol in pork urine with a reported LOD of 0.1 ng/mL, but a solid phase extraction step was required. Without sample cleanup, LODs were 0.5−1.0 ng/mL, similar to our results. The LODs of kidney, muscle, liver, and lung were 0.1, 0.2, 0.3, and 0.4 ng/g using ASAP and 0.5, 0.2, 0.3, and 0.6 ng/g, respectively, with MDESI MS. The LOQs of kidney, muscle, liver, and lung in ASAP were 0.4, 0.6, 1.1, and 1.2 ng/g, and 1.6, 0.5, 0.9, and 2.1 ng/g in MDESI MS respectively. Compared to urine samples, the LODs and LOQs were lower for tissue samples in both ASAP and MDESI MS, which can be attributed to the nominal sample processing performed for tissues before their analysis. Of note, these detection limits are of sufficient sensitivity, well below the MRLs of most counties that have MRLs32 to rapidly detect food-animal exposures to zilpaterol but are well above the remarkable LODs reported by Xu et al.37 of 0.002−0.006 μg/kg for β-agonists in pork tissue using internal extractive electrospray ionization mass spectrometry. Carrizo et al.7 detected nicotine biomarkers such as 9-hydroxy-phenanthrene and 1-hydroxy-pyrene in urine using ASAP without any pretreatment of the samples and obtained LODs and LOQs of the compounds in the range of 0.1−0.5 and 1−3 μg/mL, respectively. Crevilin et al.42 reported LODs of amphetamine and other central nervous system stimulants including methamphetamine, ephedrine, sibutramine, and fenfluramine in urine using ASAP in the range of 0.002−0.4 ng/mL. Repeatability was 10−14% (RSD; calculated by running 1 ng/ L of each compound in triplicate). Kennedy et al.50 coupled DESI-MS with solid-phase microextraction to quantitate cocaine and its metabolites from unprocessed urine and reported LOD/LOQ in the range of 10−160 ng/mL. Similarly, Rosting et al.51 conducted high-throughput screening of methadone, amitriptyline, nortriptyline, and pethidine from urine and obtained LODs between 4 and 17 ng/mL. In a study E
DOI: 10.1021/acs.jafc.8b03925 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
agonists and veterinary drugs. The semiquantitative results obtained from ASAP and MDESI MS analyses of urine were also compared to results obtained after analyses using the most highly regarded method for zilpaterol quantification, internal standard quantification using LC−MS/MS. Correlations between LC−MS/MS and ASAP (R2 = 0.744) and MDESI MS (R2 = 0.791) obtained (Figure 4B) from the urine data suggested that either ASAP or MDESI MS would serve as an acceptable screening method. For a semiquantitative method, these values are quite encouraging, especially given that the samples used for the LC−MS/MS analysis underwent a timeconsuming SPE process compared to the raw urine samples used for the ASAP and MDESI MS analyses and the volume of sample introduced into the mass spectrometer was tightly controlled at the LC-autosampler. Similar correlation of 0.70 between DESI and LC−MS was obtained by Abbassi-Ghadi et al.53 for the analysis of glycerophospholipid species in human esophageal cancer tissues. Figure 5 summarizes data obtained from ASAP, MDESI MS, and LC−MS/MS analyses of kidney, muscle, liver, and lung of sheep dosed with feed containing 0.75 mg/kg of zilpaterol and slaughtered with a 0-d withdrawal period (i.e., presumably sheep with the greatest tissue concentrations of zilpaterol). Positive signals for zilpaterol in kidney, liver, and lung were obtained using ASAP, MDESI MS, and LC−MS/MS analysis; however, in some cases, the values were less than LOQ. For muscle, all four animals were positive using LC−MS/MS, in contrast to MDESI MS and ASAP in which only two and three, respectively, of four samples were positive. Concentrations of zilpaterol in kidney ranged from 0.6−4, 2.0−6.2, and 4.0−4.7 ng/g by ASAP, MDESI MS, and LC−MS/MS, respectively. In muscle, the concentrations were higher in the range of 6.4−9.8 ng/g for ASAP and 4.8−5.4 ng/g for MDESI MS. In comparison, zilpaterol concentration in muscles analyzed using LC−MS/MS, ranged from 1.0−1.5 ng/g. In liver and lung, the concentrations of zilpaterol were 2.2−7.0 and 1.2− 2.6 ng/g, respectively, using ASAP. However, two out of four lung samples were less than LOQ in ASAP. In MDESI MS, three out of four liver samples and all lung samples were less than the LOQ. The zilpaterol concentrations in liver and lung analyzed using LC−MS/MS were 3.1−4.3 and 1.8−2.1 ng/g, respectively. Analysis of Archived Samples. Because we were interested in determining the utility of ASAP and MDESI MS over a range of potential zilpaterol concentrations, we used each method to analyze archived urine and tissue samples in which zilpaterol had been previously quantified using HPLC and LC−MS.38 The archived samples were from sheep that had been dosed with dietary zilpaterol HCl at 6 mg/kg feed (the level required for growth effects) and which contained higher zilpaterol concentrations than those measured in sheep from the current study. For urine, the results obtained from ASAP and MDESI analyses of the archived samples (Table 1) were in general agreement with the published data, except that MDESI MS tended to return urine concentrations lower than ASAP when urinary zilpaterol concentrations were high. Nevertheless, the ASAP and MDESI data followed a similar pattern and concentration range as the LC−MS data with differences only in the absolute numbers.38 Zilpaterol concentrations in muscle (withdrawal day 0, W0) were similar when analyzed using all three MS methods (ASAP, MDESI MS, and LC−MS/ MS). However, ASAP and MDESI data obtained from kidney and liver samples from the W0 group were mostly higher than
Figure 3. Urinary zilpaterol concentrations as measured by ASAP (black) or MDESI MS (orange) from sheep fed zilpaterol HCl for 6 days (circles; zilpaterol concentrations >100 ng/mL) and from the same sheep after a 1-day zilpaterol HCl withdrawal period (diamonds; zilpaterol concentrations ∼20 ng/mL or less). 38 The data demonstrate the similarity of results obtained using ASAP and MDESI MS and each method’s utility for rapidly detecting order of magnitude differences in concentrations.
magnitude difference in zilpaterol concentration present in urine collected during the feeding and withdrawal periods. More broadly, analysis of 252 urine samples collected from sheep given different zilpaterol doses (0.0075, 0.075, and 0.75 mg/kg feed) and analyzed using both ASAP and MDESI MS resulted in a fairly good correlation (R2 = 0.872) between the methods (Figure 4A). As such, we believe that either ASAP or DESI could be used for applications involving rapid screening of zilpaterol in urine and we are currently testing whether either method could be used to detect a broader set of β-
Figure 4. Correlations of urinary zilpaterol results obtained (A) by ASAP and MDESI MS and (B) by LC−MS/MS analysis and ASAPMS (blue) or MDESI MS (orange) analysis. Insets for panels A and B show the correlations at zilpaterol concentrations from 0 to 100 ng/ mL. F
DOI: 10.1021/acs.jafc.8b03925 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
Figure 5. Tissue zilpaterol concentrations as measured by ASAP (blue), MDESI MS (orange), and LC−MS/MS (gray) in tissues of four sheep fed zilpaterol HCl (0.75 mg/kg) for 12 days and sacrificed on withdrawal day 0: (A) kidney, (B) muscle, (C) liver, and (D) lung. An asterisk (*) indicates that the sample concentration was
