Anal. Chem. 1996, 68, 1918-1923
Detection and Confirmation of β-Agonists in Bovine Retina Using LC/APCI-MS Daniel R. Doerge,*,† Mona I. Churchwell,† C. Lee Holder,† Loyd Rowe,‡ and Steve Bajic§
National Center for Toxicological Research, Jefferson, Arkansas 72079, USDA-ARS, College Station, Texas 77845, and VG Organic, Altrincham, Cheshire WA14 5RZ, England
β-Adrenergic receptor agonists are growth-promoting drugs with the potential for illegal use in livestock, and human toxicity has resulted from consumption of contaminated meat. On-line liquid chromatography with atmospheric pressure chemical ionization mass spectrometry (LC/APCI-MS) was used for sensitive detection of several β-agonists in retina, a tissue reported to concentrate and retain such residues for extended periods. Multiresidue extraction, separation, detection, and confirmation procedures were developed for retinal tissue and applied to eyes from cattle treated with clenbuterol (69-201 ppb) and to control eyes spiked with salbutamol (100 ppb) and terbutaline (25-100 ppb). Rapid switching of the potential difference between sampling cone and skimmer in the transport region of the API source was used to optimize acquisition of the protonated molecules and characteristic fragment ions obtained by collisioninduced dissociation reactions. The respective selected ions were simultaneously acquired using a single quadrupole mass spectrometer. The accurate and precise agreement observed for diagnostic ion intensity ratios between β-agonists in retinal samples and authentic standards suggests that LC/APCI-MS can be used for confirmation of analyte structure at trace levels and does not require the use of a triple-stage quadrupole mass analyzer. The misuse of β-adrenergic receptor agonists for growthpromoting properties in livestock presents several problem for regulatory agencies. Consumption of contaminated meat products has been associated with acute human toxicity,1 but the high potency for metabolic repartitioning effects in livestock2 and subsequent rapid elimination from urine and plasma (t1/2 ≈ 0.5 day, ref 3) leaves very low residue levels for monitoring in tissues. Residues are cleared from liver more slowly (t1/2 ≈ 2 day, refs 2, 4), but pigmented retinal epithelium showed much slower residue elimination (t1/2 > 10 days, refs 2, 4), and residues have been measured up to 140 days after withdrawal.2 Screening of animal tissues for β-agonist residues can be achieved using rapid immunochemical test kits (e.g., ELISA, ref 4), but enforcement †
National Center for Toxicological Research. USDA-ARS. VG Organic. (1) Martinez-Navarro, J. F. Lancet 1990, 336, 1311. (2) Elliott, C. T.; Crooks, S. R. H.; McEvoy, J. D. G.; McCaughey, W. J.; Hewitt, S. A.; Patterson, D.; Kilpatrick, D. Vet. Res. Commun. 1993, 17, 459-468. (3) Stoffel, B.; Meyer, H. H. D. J. Anim. Sci. 1993, 71, 1875-1881. (4) Elliott, C. T.; McEvoy, J. D. G.; McCaughey, W. J.; Shortt, D. H.; Crooks, S. R. H. Analyst 1993, 118, 447-448. ‡ §
1918 Analytical Chemistry, Vol. 68, No. 11, June 1, 1996
actions require unambiguous structural determination, typically involving the use of mass spectrometry. The ability of mass spectrometry (MS) to provide unambiguous determination of analyte structure has led to its widespread use in regulatory confirmation of drug residues in food. The standards by which MS confirmations are evaluated were developed at a time when electron impact (EI) MS was the predominant ionization technique.5 The guidelines in use today vary with the regulatory agency, but typically confirmation of analyte structure in a regulatory sample requires at least three characteristic ions, and the intensity ratios of these ions must agree with those generated by an authentic standard, generally to within 5-20% depending on the method of ionization.5,6 In the case of FDA confirmatory methods, agreement to within 10% is required. Many of the compounds currently monitored by regulatory agencies are not amenable to analysis using EI, and more recently electrospray (ESI) and atmospheric pressure chemical ionization (APCI) have been used successfully.7-11 It is therefore important to determine if the criteria for confirmatory tests that were originally developed for use with EI data can be applied to data generated using atmospheric pressure ionization (API) techniques. Although EI spectra are much more amenable to library searching procedures than ESI and APCI data, it is also possible to compare API-MS responses from regulatory samples with those determined consecutively from authentic standards to achieve structural confirmation. It has been our goal to develop screening and confirmatory methods to support monitoring of β-agonist misuse in livestock using LC/APCI-MS. This technique has previously been shown to provide high sensitivity and the ability to determine β-agonist structure.7,8 GC/EI-MS methods are not applicable to underivatized β-agonists because of thermal lability, and multiresidue methods based on the formation of volatile derivatives have not been successful because of differing reactivity with derivatizing reagents and the limited spectral information available from some derivatives.12,13 (5) Sphon, J. A. J. Assoc. Off. Anal. Chem. 1978, 61, 1247-1252. (6) Heitzman, R. J., Ed. Residues in food-producing animals and their products: Reference materials and methods; Commision of the European Communities, Blackwell Scientific Publications: 1992; p 39. (7) Doerge, D. R.; Bajic, S.; Lowes, S. Rapid Commun. Mass Spectrom. 1993, 7, 462-464. (8) Doerge, D. R.; Bajic S.; Preece S. W.; Blankenship, L. R.; Churchwell, M. I. J. Mass Spectrom. 1995, 30, 911-916. (9) Doerge, D. R.; Bajic, S.; Lowes, S. Rapid Commun. Mass Spectrom. 1993, 7, 1126-1130. (10) Doerge, D. R.; Bajic, S.; Preece, S. W.; Howard, P. C. Rapid Commun. Mass Spectrom. 1994, 8, 603-606. (11) Doerge, D. R.; Bajic, S. Rapid Commun. Mass Spectrom. 1995, 9, 10121016. (12) Wilson, R. T.; Groneck, J. M.; Holland, K. P.; Henry, A. C. J. AOAC Int. 1994, 77, 917-924. S0003-2700(95)01174-7 CCC: $12.00
© 1996 American Chemical Society
Ionization using ESI and APCI is soft, producing predominately molecular species by proton transfer reactions, and collisioninduced dissociation (CID) methods must be used to obtain a sufficient number of diagnostic fragment ions with adequate intensity for a confirmatory method. CID using triple-quadrupole mass spectrometers is highly effective,14 but the expense of such instrumentation is a barrier to common use. It is also possible to use in-source CID to produce such fragmentation with a singlequadrupole instrument.7-9,11 Additional translational energy can be imparted to ions in the intermediate pressure region between the API source and the mass analyzer. In practice, this can be done in a number of different ways, depending on instrument design, for example, by applying a potential difference between the sampling cone and skimmer. As this voltage is increased, the energy of the molecular ions and, hence, the degree of fragmentation via CID reactions with neutral molecules are increased. In this way, the degree of fragmentation can be optimized to produce a maximum number of ions. However, it is often the case that a single voltage does not produce the desired number of ions with adequate intensity or eliminates the molecular ion. In previous studies, APCI-MS using a single-cone voltage setting was employed to reproducibly produce a sufficient number of diagnostic ions for use in confirming drug residues in tissue samples.11,15 However, this approach is limited to those analytes whose mass spectrum contains at least three ions with adequate intensity at one voltage. In the present study, we have extended this technique by using rapid switching of the sampling cone-skimmer voltage to optimize responses for target ions, including molecular species. This technique uses selected ion monitoring to analyze drug residues in tissues without sacrificing sensitivity. Furthermore, this technique is applicable to analytes for which fragmentation occurs over a wide range of cone voltages. The goal of this study was a method that uses a single-quadrupole mass spectrometer to unambiguously confirm multiple β-agonist residues present at relevant levels in bovine retina, a tissue with demonstrated propensity for concentrating drug residues.16 This was carried out using clenbuterol, a widely misused drug17 for which bovine tissues with incurred residues were available. Two related compounds with the potential for misuse in livestock production, salbutamol and terbutaline, were selected because of different chemical and metabolic properties from clenbuterol, and these β-agonists were spiked onto control retina. EXPERIMENTAL SECTION Reagents and Supplies. Standards for clenbuterol, terbutaline, and salbutamol were obtained from Sigma Chemical Co. (St. Louis, MO). Solid phase extraction (SPE) cartridges (tC18, 3 cm3) were obtained from Waters Associates (Milford, MA). HPLC grade solvents were obtained commercially, and MilliQ water was used throughout. Source of Tissues. Bovine eyeballs were obtained from a slaughterhouse for controls or from cattle dosed with clenbuterol in a pharmacokinetic study conducted by USDA-ARS Food Animal (13) van Rhijn, J. A.; Heskamp, H. H.; Essers, M. L.; van de Wetering, H. J.; Kleijnen, H. C. H.; Roos, A. H. J. Chromatog. 1995, 665, 395-398. (14) Kienhuis, P. G. M. J. Chromatogr. 1993, 647, 39-50. (15) Kiehl, D. E.; Kennington, A. S. Rapid Commun. Mass Spectrom. 1995, 9, 1297-1301. (16) Sauer, M. J.; Pickett, R. J. H.; MacKenzie, A. L. Anal. Chim. Acta 1993, 275, 195-203. (17) Food Chem. News 1995, 36, 46-47.
Production Laboratory, College Station, TX, in March 1993. In this study, animals were administered clenbuterol orally (20 µg kg-1 day-1) for 5 days and then withdrawn for varying times (cow 15, 50 h; cow 18, 26 h; cow 50, 6 h; cow 55, 24 h) before slaughter. Residue levels of clenbuterol in one retina from these cattle were determined in April-May 1994 by J. M. Gronek (USDA-FSIS Midwestern Laboratory) using liquid-liquid extraction and GC/ MS determination of a phosgene derivative as previously reported for analysis of liver samples.12 The second eyeball from each cow was stored at -80 °C until analyzed as described below. Sample Preparation and Extraction. Retinal tissue was removed from the back of the thawed eyeball from control and clenbuterol-dosed animals and weighed (300-700 mg/eye). The tissue was added to 0.01 M HCl (1 mL per 100 mg of minced tissue) on ice and sonicated using a Virsonic cell disruptor (Model 16-850, Gardiner, NY) at 30% of maximum power (50 W) for 10 min, followed by centrifugation at 1500 rpm for 15 min. The decanted supernatant was carefully adjusted to pH 7.0 with 0.01 M NaOH, and 1.0 mL aliquots containing 100-200 mg of tissue were diluted for SPE with an equal volume of deionized water (MilliQ, Millipore Corp., Milford, MA). The SPE cartridge was prewashed and activated sequentially with 2 mL of methanol, 2 mL of water and 1 mL of 1 mM NaOH. The retinal extract was applied to the cartridge and washed with 0.5 mL of water and 2 mL of acetonitrile. The β-agonists were eluted with 3 mL of a solution consisting of 99% methanol and 1% 100 mM ammonium acetate (pH 4.0). This solution was evaporated to dryness in a nitrogen stream at room temperature and reconstituted in 0.10.2 mL of mobile phase. For determinations from spiked tissues, the β-agonists (1 µg/mL in methanol) were applied directly onto control retinal tissue prior to homogenization. Liquid Chromatography. The LC separation used a NovaPak C18 cartridge in a radial compression module (4 µm particle size, 5 mm × 100 mm, Waters Associates) with a mobile phase consisting of 90% acetonitrile and 10% aqueous ammonium acetate (10 mM, pH 4.0) at a flow rate of 1.0 mL/min. Because of mass spectral interferences with acetonitrile, ion intensity ratio determinations for salbutamol were performed using a mobile phase consisting of 90% methanol and 10% aqueous ammonium acetate (3 mM, pH 4.0). Injection volumes of 10-20 µL were used. A switching valve (Rheodyne Model 7030, Cotati, CA) was used to divert unretained compounds to waste during the initial 4.5 min of each chromatographic run. During this time, a constant flow of the same mobile phase was delivered to the APCI probe by a Waters M6000 pump. Mass Spectrometry. A VG Platform II single-quadrupole mass spectrometer (VG Organic, Altrincham, England) equipped with an atmospheric pressure ion source and APCI interface was used. The total LC column effluent was delivered into the atmospheric pressure ion source through a heated nebulizer probe (500 °C) using nitrogen as the probe and bath gas with an ion source temperature of 120 °C. Positive ions were acquired in selected ion monitoring mode (SIM, span ) 0.02 Da, dwell ) 0.3 s). The sampling cone voltage was varied between 15 and 40 V to produce varying amounts of in-source collision-induced dissociation (see Table 1). The mass spectrometer was calibrated using a solution of poly(ethylene glycol)s (PEG 200, 25 µg/mL; 300, 50 µg/mL; 600, 75 µg/mL; 1000, 250 µg/mL; all obtained from Sigma Chemical Co.) in 50% acetonitrile in aqueous ammonium acetate (5 mM) over the mass range m/z 85-1200. Analytical Chemistry, Vol. 68, No. 11, June 1, 1996
1919
Figure 1. Cone voltage fragmentation mass spectra for β-agonists. The full-scan mass spectra were obtained from flow injection analysis of a 100 ng quantity of the respective β-agonist at a single skimmer-sampling cone voltage (20 V).
RESULTS AND DISCUSSION The procedure used for extraction of three β-agonists from bovine retina tissue is a modification of the multiresidue method previously developed for human plasma.11 Recoveries of the three β-agonists from spiked control retinal tissue (65-94%, see Table 2) were similar to those previously obtained from plasma. A previous report used ethyl acetate extraction at pH 12 for the determination of clenbuterol residues in bovine retina, although only ∼25% recovery of spiked residues was initially observed.18 These workers concluded that electrostatic interactions between clenbuterol and the negatively charged mucopolysaccharides (e.g., hyaluronic acid) in the retina were a limiting factor because treatment of retina homogenates with hyaluronidase liberated quantitive amounts of spiked clenbuterol for solvent extraction. The efficient solubilization of residues using 0.01 N HCl observed in the current study is consistent with this finding because protonation of the mucopolysaccharide carboxyl groups is expected at pH 2. Our previous work showed that the mass spectra of clenbuterol, salbutamol, and terbutaline (see Figure 1 for structures) varied as a function of the skimmer-sampling cone voltage.8 At low voltage (15 V), the spectrum consisted of predominately MH+, and as the voltage was increased (20-30 V), the intensity of fragment ions increased through in-source CID reactions. Figure 1 shows the full-scan mass spectra acquired at a single, intermediate cone voltage. Some diagnostic fragmentation product ions were identified for these tert-butyl-substituted β-agonists: (MH+ (18) Tomlinson, J. A.; Flurer, R. A.; Satzger, R. D. Proceedings of the 43rd ASMS Conference on Mass Spectrometry and Allied Topics, Atlanta, GA, May 2226, 1995; p 512.
1920
Analytical Chemistry, Vol. 68, No. 11, June 1, 1996
Table 1. Cone Voltage Switching Conditions Used for Acquisition of β-Agonist-Derived Ionsa ion (m/z)
species
cone voltage (V)
240 222 166 148 226 170 152 277 259 203
salbutamol MH+ -H2O -t-Bu - H2O -t-Bu - 2H2O terbutaline MH+ -t-Bu -t-Bu - H2O clenbuterol MH+ -H2O -t-Bu - H2O
15 20 25 22 15 23 23 15 21 21
a The protonated molecules and fragment ions from the β-agonists listed were acquired using a dwell time of 0.3 s, a span of 0.02 Da, and an interchannel delay time of 0.03 s. The cone voltage was varied as a step function in concert with acquisition of the respective ion as described in the Experimental Section. The losses from the respective protonated molecule responsible for the fragment ions are listed.8
- t-Bu), (MH+ - H2O), (MH+ - H2O - t-Bu) and (MH+ - 2H2O - t-Bu). Table 1 shows the ions corresponding to these fragmentation reactions that are specific for these β-agonists. To accommodate the need for multiple diagnostic ions of adequate intensity, including fragments and the protonated molecule, in a β-agonist confirmatory method, the mass spectrometry acquisition software was used to perform rapid switching of the sampling cone voltage as a step function in concert with acquisition of the respective selected ion using a constant dwell time of 0.3 s and interchannel delay of 0.03 s. This procedure permitted acquiring data under optimal conditions for each particular ion and gives rise to increased ion intensities, and hence
Figure 2. SIM chromatograms for terbutaline, salbutamol, and clenbuterol in an extract from a spiked retina. A control bovine retina was spiked with a mixture of the three β-agonists equivalent to 100 ppb each, and a 100 mg sample was extracted and analyzed using the rapid cone voltage switching technique described in the Experimental Section.
enhanced sensitivity, relative to that obtained by acquiring the same ions using a single cone voltage setting. Ion intensity ratios were determined from the area under the respective selected ion chromatogram and were computed relative to the area for the protonated molecule. For example, the signal for protonated clenbuterol (m/z 277) was acquired using a low voltage (15 V) while m/z 259 and 203 fragment ions were acquired at a higher voltage (21-25 V). Voltages for acquisition of the m/z 203 ion were optimized on a daily basis in the range of 21-25 V to produce an intensity approximately equal to that of the protonated molecule and m/z 259 fragment ion. Although this change was made to optimize fragmentation on different days, identical conditions were used for ion intensity ratio determinations from samples and standards on any given day. This procedure regularly gave ratios approximately equal to unity for the ions from clenbuterol and salbutamol, although it was not possible to increase the m/z 170 ion from terbutaline beyond ∼25% of the base peak (see Tables 3, 4). The goal of this optimization, i.e., ratios tending to a value of 1, was to minimize the variability observed in the individual ion intensity ratios since the relative effect of random noise is greater on low-intensity ions as opposed to more intense ones. Because the chromatographic separation (see Figure 2) between clenbuterol (7.9 min) and terbutaline and salbutamol (6.0 and 6.8 min, respectively) was adequate, the seven ions for the first two eluting compounds were acquired in one time function (seven ions total; total duty cycle, including interchannel delay, of 2.31 s) and clenbuterol in another (three ions; total duty cycle of 0.99 s). The LC separation readily accommodated the length of the acquisition cycle needed since the chromatographic peak widths afforded at least 10 acquisition points and minimal effects
on peak shape were observed (see Figure 2). Control retina showed no peaks for any of the β-agonists (see Figure 3 for clenbuterol data; salbutamol and terbutaline data not shown). Approximate method detection limits (S/N ) 3) using the cone voltage switching technique were extrapolated from the signals and noise present in the 100 ppb spiked control retina data (see Figure 2) to be approximately 100 pg for terbutaline, 120 pg for salbutamol, and 60 pg on-column for clenbuterol. Using the described extraction and LC/MS procedure, this corresponds to 10, 12, and 6 ppb in retina, respectively. If necessary, these limits could be reduced ∼10-fold by increasing the injection volume and increasing the amount of tissue processed. Method detection limits were also computed from single ion monitoring of the respective (M + H+) mass chromatogram. Calibration plots generated from replicate extractions of retina spiked with the three β-agonists at various levels (see Table 2) gave highly linear responses for each analyte (r2 g 0.999) up to 100 ppb with slope values for clenbuterol, 29.2; salbutamol, 34.4; terbutaline, 39.2 (data not shown) area counts/pg injected. Using the method of Foley and Dorsey for calculating chromatographic detection limits (DL ) 3sn/s, where sn is the standard deviation of the noise and s is the slope of the response), values of 3, 5, and 13 ppb in retina were determined for clenbuterol, salbutamol, and terbutaline, respectively.19 This comparison confirms our observation that sensitivity is minimally affected when acquiring multiple ions under the rapid cone voltage switching conditions as opposed to one selected ion. (19) Foley, J. P.; Dorsey, J. G. Chromatographia 1984, 18, 503.
Analytical Chemistry, Vol. 68, No. 11, June 1, 1996
1921
Figure 3. SIM chromatograms for clenbuterol-derived ions in a standard, a retinal extract from a dosed cow, and a control retinal extract. Authentic standard clenbuterol (1.0 ng on-column, panel B) and extracts from control (panel A) and clenbuterol-dosed (panel C) bovine retinal tissue were injected into the LC/APCI-MS system, and selected ions were acquired at different cone voltages as described in the Experimental Section. The integrated areas are listed above the respective peak. Table 2. Recoveries of β-Agonists from Spiked Retinaa spike level (ppb)
clenbuterol
salbutamol
terbutaline
100 50 25
87 ( 3.8 76 ( 1.9 65 ( 6.2
94 ( 4.9 80 ( 2.7 77 ( 6.6
88 ( 3.4 70 ( 2.6 69 ( 4.0
a Control retinal tissue was spiked with a mixture of the three β-agonists at the indicated levels and extracted as described in the Experimental Section. Average recoveries and standard deviations were determined from control retina extracts (n ) 4).
Although accurate quantification of clenbuterol residues in retina was not a primary goal of this study, it was possible to determine the residue levels present in retinas from clenbuteroldosed cows using external standard calibration of interspaced injections (see Table 3). This was necessary to minimize the minor (∼5% of peak area) fluctuations in absolute signal intensity often seen between consecutive injections. The values determined using LC/MS in 1995 using LC/APCI-MS (69-201 ppb, uncorrected for recovery) were similar (∼68-100%) to those previously determined in 1994 (J. Gronek, unpublished) using a GC/MS method.12 This suggests that only small amounts of clenbuterol decomposed during a storage period of ∼2 years.15 Future studies that emphasize accurate quantification will require synthesis of stable isotope-labeled β-agonists in order to perform isotope dilution experiments. 1922 Analytical Chemistry, Vol. 68, No. 11, June 1, 1996
Table 3. Quantification of Clenbuterol Residues in Dosed-Cow Retina Using LC/APCI-MSa cow no.
Clenbuterol (LC/APCI-MS)
50 55 15 18
94 ( 9 69 ( 5 136 ( 21 201 ( 11
a LC/MS quantification was obtained using external standard calibration as described in the Experimental Section.
Figure 3 shows the comparison between selected ion chromatograms for standard clenbuterol, incurred residues in bovine retina, and a control retina. No interference was observed in control retinal extracts at any of the three masses, although a possible metabolite peak with retention time slightly longer than that of clenbuterol was noted. The correspondence of retention times for standards and incurred residue samples and the lack of such peaks in the control sample are consistent with the presence of clenbuterol. It was observed that the ion intensity ratios for the three characteristic ions in both samples and standards were reproducible (relative standard deviation e3% for any given ion, data not shown). Moreover, all ratios showed concordance between incurred residue samples from four different cows dosed with clenbuterol and the accompanying clenbuterol standard. It
Table 4. Ion Intensity Ratios for Incurred Clenbuterol Residues in Bovine Retinaa ratio sample
m/z 259/277
m/z 203/277
cow 50 (n ) 3) % difference cow 55 (n ) 2) % difference cow 15 (n ) 4) % difference cow 18 (n ) 6) % difference
0.79 ( 0.02 3.7 0.77 6.1 0.70 ( 0.02 4.5 0.78 ( 0.02 1.3
0.89 ( 0.02 6.3 0.88 7.4 0.91 ( 0.01 0 1.09 ( 0.04 1.8
Table 5. Ion Intensity Ratios for Residues of Terbutaline (A) and Salbutamol (B) Spiked into Bovine Retina and Comparison with Values from Authentic Standardsa (A) Terbutaline 100 ppb average ratio for m/z 170/226 % difference average ratio for m/z 152/226 % difference
(B) Salbutanol (100 ppb) average ratio for m/z 222/240 % difference average ratio for m/z 166/240 % difference average ratio for m/z 148/240 % difference
a
Ion intensities were determined from the peak areas under the respective selected ion chromatogram, and the ratios were computed relative to the area for m/z 277. The number of replicate determinations (n) and the standard deviation of the average were determined as follows: calf 50, 2 retinal subsamples, 2 injections from one, 1 injection from the other; calf 55, 2 retinal subsamples, 1 injection each; calf 15, 2 retinal subsamples, 2 injections each; calf 18, 2 retinal subsamples, 3 injections each. Average intensity ratio values, determined from clenbuterol residues in retinal extracts and clenbuterol standards injected just prior to and just after the incurred samples, were compared as the relative difference between incurred residueand standard-derived ratios: % difference ) [(standard ratio average - retina extract ratio average)/standard ratio average] × 100. Retinas from cows 50 and 55 were analyzed in May 1995, cow 15 in June 1995, and cow 18 in August 1995.
should be noted that efficient chromatographic separation of clenbuterol from the possible metabolite (see Figure 3C) was essential to this agreement since coelution of these peaks would have adversely affected the observed ion intensity ratios. To analyze the accuracy of incurred residue ratio determinations with respect to authentic clenbuterol standard ratios, the data sets shown in Table 4 were compared as the % difference ) [(standard ratio average - retina extract ratio average)/standard ratio average] × 100. The data show the % difference between ion intensity ratios determined on three different days (over a 3 month period) for clenbuterol standards and clenbuterol incurred in retinal tissue. Data for each retinal sample were obtained by injecting 2-3 clenbuterol standards followed by 3-6 incurred sample injections and an additional 2-3 standard injections. The average ratios for clenbuterol in the retinal extracts were then compared with the composite average for the determinations of standard injections. As shown in Table 4, the values of the ion intensity ratios determined for clenbuterol in tissue extracts and a corresponding amount (1 ng) of standard agreed to within 7% difference for all four dosed cow eyes tested. This degree of match between sample and standard ion intensity ratios for the five ions monitored is sufficient for regulatory confirmation using the commonly held criteria of at least three characteristic ions with concordance between sample and standard ion intensity ratios.5,6 However, the data in Table 4 also show that some variation in the ratios was observed on different days of analysis. Irrespective of this minor variation, it was observed that in all cases the ratios derived from standards accurately and precisely matched those derived from incurred residues on any given day. Similar analysis of the ion intensity ratios from terbutaline- and salbutamol-spiked retina was also performed (see Table 5).
0.26 (2) 4.4 0.98 (2) 5.1
50 ppb
25 ppb
0.25 (3) 5.1 0.96 (3) 0.1
0.25 (3) 3.2 1.03 (3) 0.5
1.02 (4) 0.3 0.98 (4) 0.3 1.07 (4) 1.6
a Control retina were spiked with salbutamol (100 ppb) or terbutaline (25-100 ppb) and the samples extracted and analyzed by LC/APCIMS as described in the Experimental Section. Ion intensities ratios for salbutamol and terbutaline, as defined in Table 1, were computed relative to the area of the MH+ ion, m/z 240 and 226, respectively. Ion intensity ratios for spiked residues were compared to those generated by interspaced injections of an equivalent amount of authentic standards as described in Table 4. The average ratios for the spiked samples and the number of determinations (n) are listed. The relative standard deviations for all average ratios listed were e4%.
Concordance between standard and spiked retinal extract ratios were obtained for terbutaline as low as 25 ppb and for salbutamol at 100 ppb, although lower levels were not tested. The results with these three analytes suggest that this procedure may be applicable for multiresidue screening of target β-agonists in retina. It will be important to verify these results with tissues from salbutamol- and terbutaline-dosed animals when possible. These results show that the described method for confirmation of β-agonist residues using a single quadrupole mass spectrometer meets the generally accepted criteria for a regulatory method.5,6 Rapid switching of the cone voltage enabled accurate and precise measurement of ion intensity ratios that are needed to confirm the presence of β-agonists in animal tissues. As the instrumentation described in this study becomes more widely available in regulatory laboratories, it will be important to determine the reproducibility of these procedures through interlaboratory comparisons of instrumentation from different sources. ACKNOWLEDGMENT We gratefully acknowledge the assistance provided by our USDA-FSIS colleagues Dr. C. L. Deyrup, A. C. Henry, and J. M. Gronek.
Received for review December 4, 1995. Accepted March 26, 1996.X AC951174F X
Abstract published in Advance ACS Abstracts, May 1, 1996.
Analytical Chemistry, Vol. 68, No. 11, June 1, 1996
1923
