Rapid and High-Throughput Determination of Sixteen β-agonists in Livestock Meat Using One-Step Solid-Phase Extraction Coupled with UHPLC-MS/MS

β-agonists are illegally added to animal feed because they can greatly increase carcasses’ leanness, which impairs the safety of animal-derived foods and indirectly endangers human health. This study aimed to develop an ultrahigh-performance liquid chromatography-tandem mass spectrometry (UHPLC-MS/MS) method for determining sixteen β-agonists in livestock meat. The homogenized samples were subjected to enzymatic hydrolysis using β-glucuronidase/sulfatase at 40 °C for 2 h, extracted with acetonitrile containing 1% acetic acid (v/v), and purified by the one-step Qvet-AG extraction column. The residue was redissolved by 0.1% aqueous formic acid/methanol (9:1, v/v) after blow-drying by nitrogen, and then determined by UHPLC-MS/MS. The results demonstrated that the well linearity was in the range of 0.1–50 μg/L with the correlation coefficient (R2) ≥0.9928, and the limits of detection (LOD) and quantification (LOQ) were 0.01–0.11 μg/kg and 0.04–0.38 μg/kg, respectively. With intraday and interday relative standard deviations (RSDs) being less than 10%, the average recoveries of pork, beef, and lamb at various spiked levels ranged from 62.62–115.93%, 61.35–106.34%, and 62.00–111.83%, respectively. In conclusion, the established method is simple, efficient, sensitive, and suitable for the simultaneous detection of several β-agonist residues in livestock meat.


Introduction
Livestock meat has traditionally been a fundamental part of the human diet because of its tremendous nutritional value. With the improvement of living standards, meat consumption shows a trend of annual growth. In recent years, people's views on health have also changed, moving away from high-fat and high-energy diets toward low-fat and low-sugar foods. This has an impact on traditional livestock farming, such as lean pigs being more favored by consumers. β-agonists are a class of phenylethanolamine compounds with an adrenergic function; they were initially developed to treat uterine, respiratory, and cardiovascular diseases in humans and animals [1]. Later, studies found that β-agonists significantly improved lean muscle mass of carcasses and the feed conversion ratio of animals when taken at doses 5 to 10 times higher than therapeutic doses [2]. Because of their huge feeding benefits, the illicit abuse of β-agonists in animal feed occurs often, especially in livestock animals such as pigs, cattle, and sheep, which are more difficult to regulate. However, the chemical structures of β-agonists are stable and frequently difficult to be degraded in animals, making them easily residual in the human body through the

Standard Solution Preparation
A total of 100 mg/L of sixteen standards and five internal standards as standard stock solutions was stored at −20 • C for up to six months. Mixed standard solution at the concentration of 1.0 mg/L was prepared by diluting the sixteen standard stock solutions with methanol. Mixed internal standard solution at the concentration of 1.0 mg/L was prepared by diluting the five internal standard stock solutions with methanol as well. The aforementioned mixed standard solutions were stored at −20 • C for one month. According to experimental needs, the mixed standard solution and the mixed internal solution were diluted by methanol to obtain 100 µg/L of the mixed standard working solution and mixed internal standard working solution, respectively, and stored at 4 • C until use.

UHPLC-MS/MS Instrumentation and Operating Conditions
Thermo Scientific Vanquish ultrahigh-performance liquid chromatography coupled to Thermo Scientific TSQ Quantis mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) was used. Separation was performed with the Thermo Hypersil GOLD aQ column (100 mm × 2.1 mm, 1.9 µm; Thermo Fisher Scientific, Waltham, MA, USA). The column temperature was 30 • C, and the sample injection volume was 3 µL. Mobile phase A was 0.1% aqueous formic acid (v/v), mobile phase B was methanol containing 0.1% formic acid (v/v), and the flow rate was held at 0.3 mL/min throughout the analysis. Gradient conditions were 0-2.0 min (95% A), 2.1-5.0 min (70% A), 5.1-9.0 min (60% A), 9.1-12.0 min (10% A), 12.1-13.0 min (95% A), and held at this level for 3 min so that the system could re-equilibrate before the next injection.
The MS/MS was equipped with an ESI source operating in the positive ionization mode [M+H] + and the data were collected in multiple reaction monitoring (MRM) mode. The optimized electrospray ionization parameters were as follows: capillary voltage, 3.5 kV; ion source temperature, 450 • C; ion transfer tube temperature, 300 • C; vaporizer temperature, 280 • C; sheath gas, 38 Arb; auxiliary gas, 4 Arb.

Sample Preparation
Pork, beef, and lamb samples (2.00 ± 0.02 g) were accurately weighed into a 50 mL polypropylene centrifuge tube after being chopped and homogenized using an HM6300 intelligent homogenizer (Lab Precision Beijing Technology Co., Ltd., Beijing, China). Samples were spiked with 50 µL of the mixed internal standard working solution (100 µg/L). A total of 10 mL of 0.2 mol/L ammonium acetate buffer solution (pH 5.2) and 40 µL βglucuronidase/sulfatase was added and mixed (Vortex Genius 3, IKA-Werke GmbH & Co., Staufen, Germany). It was incubated for 2 h at 40 • C (TW20, Julabo GmbH, Seelbach, Germany). After being cooled to room temperature, the homogenate was then centrifuged at 10,621× g for 5 min (Centrisart ® D-16C, Sartorius, Göttingen, Germany), and the supernatant was transferred to another 50 mL polypropylene centrifuge tube. Then, 10 mL of acetonitrile containing 1% acetic acid (v/v) and 2 g of NaCl was successively added into the supernatant. Next, the mixture was fully vortexed for 5 min and centrifuged at 10,621× g for 5 min again, taking 2 mL of clear supernatant passed through the Qvet-AG one-step solid-phase extraction column (6 mL-2 g, Shimadzu, Kyoto, Japan) at a rate of about 1 drop/second for further purification. The entire filtrate was then collected directly in a test tube, which then was evaporated to dryness in a 40 • C water bath under a gentle  (N-EV AP-11 nitrogen evaporator, Organomation, Berlin, MA, USA), and redissolved to 1 mL with 0.1% aqueous formic acid/methanol (9:1, v/v). After mixing thoroughly, the solution was filtered through the 0.22 µm hydrophilic filter membrane (Agilent Technologies, Santa Clara, CA, USA) for UHPLC-MS/MS analysis. The schematic diagram of the main experimental operation is shown in Figure 1. at 10,621× g for 5 min (Centrisart ® D-16C, Sartorius, Göttingen, Germany), and the supernatant was transferred to another 50 mL polypropylene centrifuge tube. Then, 10 mL of acetonitrile containing 1% acetic acid (v/v) and 2 g of NaCl was successively added into the supernatant. Next, the mixture was fully vortexed for 5 min and centrifuged at 10,621× g for 5 min again, taking 2 mL of clear supernatant passed through the Qvet-AG one-step solid-phase extraction column (6 mL-2 g, Shimadzu, Kyoto, Japan) at a rate of about 1 drop/second for further purification. The entire filtrate was then collected directly in a test tube, which then was evaporated to dryness in a 40 °C water bath under a gentle flow of nitrogen (N-EV AP-11 nitrogen evaporator, Organomation, Berlin, MA, USA), and redissolved to 1 mL with 0.1% aqueous formic acid/methanol (9:1, v/v). After mixing thoroughly, the solution was filtered through the 0.22 μm hydrophilic filter membrane (Agilent Technologies, Santa Clara, CA, USA) for UHPLC-MS/MS analysis. The schematic diagram of the main experimental operation is shown in Figure 1.

Matrix Effect Evaluation
The Matrix Effect (ME), which has an impact on the accuracy of the quantitative results, is the interference of other components in the sample with the analytical process of the target compound [21]. According to the established sample processing method, the matrix-matched standard curves (the seven concentration levels of compounds were 0.1, 0.5, 1.0, 2.0, 5.0, 10.0, and 50.0 μg/L) were obtained by the blank samples and compared with the reagent standard curves obtained by redissolved solution (0.1% aqueous formic acid/methanol, 9:1, v/v). The MEs were evaluated by calculating the relationship between the slope ratio of the matrix-matched standard curve and the reagent standard curve, and the formula is: where ka is the slope of the matrix-matched standard curve; kb is the slope of the reagent standard curve. When ME > 1, it indicates a matrix enhancement effect; when ME < 1, it indicates a matrix inhibition effect, and when ME = 1, then there is no matrix interference [22].  The Matrix Effect (ME), which has an impact on the accuracy of the quantitative results, is the interference of other components in the sample with the analytical process of the target compound [21]. According to the established sample processing method, the matrix-matched standard curves (the seven concentration levels of compounds were 0.1, 0.5, 1.0, 2.0, 5.0, 10.0, and 50.0 µg/L) were obtained by the blank samples and compared with the reagent standard curves obtained by redissolved solution (0.1% aqueous formic acid/methanol, 9:1, v/v). The MEs were evaluated by calculating the relationship between the slope ratio of the matrix-matched standard curve and the reagent standard curve, and the formula is: where k a is the slope of the matrix-matched standard curve; k b is the slope of the reagent standard curve. When ME > 1, it indicates a matrix enhancement effect; when ME < 1, it indicates a matrix inhibition effect, and when ME = 1, then there is no matrix interference [22].

Determination of Linearity, Limit of Detection, and Limit of Quantification
Based on the established UHPLC-MS/MS method, the prepared matrix-matched standard working solutions were determined on the machine under optimized instrumental conditions. The standard curve was plotted by linear fitting with the ratio of compound concentration (µg/L) to the concentration of the internal standard (µg/L) as the X-axis, and the ratio of the peak area of quantitative ion chromatography for each compound to the peak area of the corresponding internal standard as the Y-axis. From this, the regression equations were obtained and the corresponding squared correlation coefficients (R 2 ) were calculated. The limits of detection (LODs) and limits of quantification (LOQs) for the target compounds were assessed by testing a series of spiked blank samples. The spiked concentration at signal-to-noise (S/N) ≥ 3 was used as LOD of the method, while LOQ of the method was the spiked concentration at S/N ≥ 10 [23].

Recovery and Precision Test
The accuracy (expressed as recovery) and the precision (as relative standard deviation, RSD) of the analytical method were evaluated using spiked blank samples at three concentration levels (0.5, 1.0, and 5.0 µg/kg), and each concentration was set in six parallels. The analysis was repeated three times (three different days), and the average recovery was calculated along with intraday and interday RSD. The results were used to assess the accuracy, reproducibility, and stability of the method. The recovery and RSD were calculated as follows: where 2.5 is the conversion multiplier, C E (µg/L) is the experimental concentration determined from the calibration curve, C S (µg/kg) is the spiked concentration, SD (µg/L) is the standard deviation, and C A (µg/L) is the average of experimental concentration determined from the calibration curve.

Analysis of Actual Sample
Thirty samples of pork, beef, and lamb were randomly chosen from three different Chinese provinces and tested for sixteen β-agonists using the method determined in this work. To further confirm the feasibility and accuracy of the method, sixteen compounds were quality controlled by blank spiking during the assay.

Data Analysis
TraceFinder 4.1 software (Thermo Fisher Scientific, Waltham, MA, USA) was used for data acquisition and processing, and OriginPro software (2022b, OriginLab Inc., Northampton, MA, USA) was used for plotting. Three parallels were performed for each experiment, and data were expressed as mean ± standard deviation.

Optimization of Chromatographic Condition
The majority of the research currently in publication uses reversed-phase columns for the separation of β-agonists, which are moderately or strongly polar compounds [24,25]. In this test, we compared the separation performances of three columns: the Acquity UPLC HSS C18 column (150 mm × 2.1 mm, 1.7 µm; Waters, Milford, MA, USA), the Accucore TM aQ column (150 mm × 2.1 mm, 2.6 µm; Thermo Fisher Scientific, Waltham, MA, USA), and the Hypersil GOLD aQ column (100 mm × 2.1 mm, 1.9 µm; Thermo Fisher Scientific, Waltham, MA, USA). The Hypersil GOLD column was chosen due to its outstanding aqueous mobile phase stability, strong retention capabilities, and excellent separation properties for polar compounds.
To better separate the sixteen β-agonists, different mobile phases such as watermethanol, water-acetonitrile, 0.1% aqueous formic acid (v/v)-methanol containing 0.1% formic acid (v/v), and 0.1% aqueous formic acid (v/v)-acetonitrile containing 0.1% formic acid (v/v) were tried. The results demonstrated that adding a certain acid to the mobile phase could enhance the ionization efficiency and make the target compounds easier to adsorb and elute from the column. Moreover, since the elution capacity of methanol is weaker than that of acetonitrile, β-agonists with similar polarity enable elution in segments. Sharp symmetrical peak patterns can be obtained with a high signal responsiveness due to the solvent's consistency. Consequently, 0.1% aqueous formic acid (v/v)-methanol containing 0.1% formic acid (v/v) was used as the mobile phase composition.
It is challenging to separate these sixteen substances using isocratic elution since the compounds have similar polarities. In order to obtain the best possible separation, the gradient elution procedure was finally determined by adjusting the elution intensity of the mobile phase. The sixteen target compounds were well separated under the determined chromatographic condition, and the analysis time was 16 min, which is faster than the method in the Chinese standard GB/T 22286-2008 for the detection of eleven β-agonists in 26 min. The ultrahigh-performance liquid chromatogram of the mixed standard solution is shown in Figure 2.
anol, water-acetonitrile, 0.1% aqueous formic acid (v/v)-methanol containing 0.1% formic acid (v/v), and 0.1% aqueous formic acid (v/v)-acetonitrile containing 0.1% formic acid (v/v) were tried. The results demonstrated that adding a certain acid to the mobile phase could enhance the ionization efficiency and make the target compounds easier to adsorb and elute from the column. Moreover, since the elution capacity of methanol is weaker than that of acetonitrile, β-agonists with similar polarity enable elution in segments. Sharp symmetrical peak patterns can be obtained with a high signal responsiveness due to the solvent's consistency. Consequently, 0.1% aqueous formic acid (v/v)-methanol containing 0.1% formic acid (v/v) was used as the mobile phase composition.
It is challenging to separate these sixteen substances using isocratic elution since the compounds have similar polarities. In order to obtain the best possible separation, the gradient elution procedure was finally determined by adjusting the elution intensity of the mobile phase. The sixteen target compounds were well separated under the determined chromatographic condition, and the analysis time was 16 min, which is faster than the method in the Chinese standard GB/T 22286-2008 for the detection of eleven β-agonists in 26 min. The ultrahigh-performance liquid chromatogram of the mixed standard solution is shown in Figure 2.

Optimization of Mass Spectrometry Condition
Sixteen single standard solutions and five individual internal standard solutions of 100 μg/L were injected into the mass spectrometer at a flow rate of 3 μL/min by using a needle pump. The Q1 full scan was turned on in positive ion mode to determine the

Optimization of Mass Spectrometry Condition
Sixteen single standard solutions and five individual internal standard solutions of 100 µg/L were injected into the mass spectrometer at a flow rate of 3 µL/min by using a needle pump. The Q1 full scan was turned on in positive ion mode to determine the molecular ion of each standard. Then, the instrument scanned in SIM Q3; for each target compound, the two pairs of characteristic ion pairs with the highest response were selected as qualitative and quantitative ion pairs. Finally, various mass spectral parameters were optimized in the MRM mode (Table 1).

Optimization of Enzymatic Condition
Enzymolysis and acid hydrolysis are typically utilized to achieve the accurate quantification of β-agonists in organisms [26]. Glucuronidase is also mostly used for enzymolysis; however, the enzymatic effects are greatly influenced by the temperature and time of the enzymolysis. According to the reported literature [1,27,28], the blank spiking recovery was employed as an indicator to compare the effect of different temperature and time on the enzymatic effects. As shown in Table 2, the enzymatic effect at 40 • C for 2 h was comparable to the effect at 37 • C for 12 h. The recovery of seven compounds, including ISS, TUL, RAC, etc., was poor when enzymolysis was performed at 55 • C, suggesting that the temperature affected the enzyme activity and consequently the recovery of the target compounds. As a result, the present method identified 2 h of enzymolysis at 40 • C. Compared with testing methods in the Chinese Standards (GB/T 21313-2007 [19], and SN/T 1924-2011 [20]) using enzymolysis at 37 • C for 12 h, this method effectively reduced the pretreatment time and improved the detection efficiency.

Optimization of Extraction Solvent
β-agonists have moderate or strong polarity, and methanol and acetonitrile are typically used as extraction solvents in the residue analysis of such substances, which can reduce the interference of high-fat and high-protein matrices [29]. Based on the spiked recovery of the target compounds, the effects of no-extraction, methanol extraction, and acetonitrile containing 1% acetic acid (v/v) extraction were compared. Table 3 shows the poor recovery of multiple compounds in the absence of no extraction and methanol extraction. Compared with methanol, acetonitrile brought in lesser polar interfering substances during extraction and provided a better subsequent purification effect. Additionally, the addition of a tiny quantity of acid to acetonitrile was more favorable for the extraction of weakly basic β-agonists, and these results are consistent with the published literature [30,31], so the experiment chose acetonitrile containing 1% acetic acid (v/v) as the extraction solvent.

Optimization of Solid-Phase Extraction Column
There are still various impurities in the extract to interfere with the targets' response on the instrument even after extraction with acetonitrile containing 1% acetic acid (v/v). Further purification is necessary, and solid-phase extraction (SPE) has been widely used as a mature sample pretreatment technique, which can effectively separate the target and impurities. Its advantages over conventional liquid-liquid extraction include lesser solvent consumption, convenient operation, and superior stability [24]. By reviewing the literature, the Oasis MCX column [32,33] (60 mg/3 mL, Waters, Milford, MA, USA), the Oasis HLB column [34,35] (200 mg/3 mL, Waters, Milford, MA, USA), and the Bond Elut C18 column [36,37] (200 mg/3 mL, Agilent Technologies, Santa Clara, CA, USA) were more commonly used in previous β-agonists assays. Qvet-AG, as a new one-step veterinary drug pretreatment column, is less utilized at present. The purification effects of the four SPE columns mentioned above for β-agonists in this experiment were compared based on the spiked recoveries of the sixteen compounds. As shown in Table 4, the sample had the best purification effect on the Qvet-AG column, and the spiking recoveries of all the compounds met the requirements. Moreover, the Qvet-AG column was chosen for the experiment because, compared with conventional SPE columns, it did not need activation, equilibration, elution, and other steps; the one-step operation efficiently shortened the sample purification time by about 80% at the same time as reducing the amount of reagent consumed.

Optimization of Redissolved Solution
The choice of redissolved solution before sample loading directly affects the sensitivity of the method. The 10 µg/L mixed standard working solution was blown dry under nitrogen and redissolved using different volume ratios of 0.1% aqueous formic acid/methanol (90:10, 70:30, 50:50). It was discovered that the increase in the proportion of the organic phase made the target more soluble and significantly improved the response of the sixteen compounds. However, as illustrated in Figure 3a,b, the precision of the quantitative results is easily impacted by the solvent effects when the ratio of organic phase in the redissolved solution is too high. Therefore, 0.1% aqueous formic acid/methanol (90:10, v/v) was finally selected as the redissolved solution.
the spiked recoveries of the sixteen compounds. As shown in Table 4, the sample had the best purification effect on the Qvet-AG column, and the spiking recoveries of all the compounds met the requirements. Moreover, the Qvet-AG column was chosen for the experiment because, compared with conventional SPE columns, it did not need activation, equilibration, elution, and other steps; the one-step operation efficiently shortened the sample purification time by about 80% at the same time as reducing the amount of reagent consumed.

Optimization of Redissolved Solution
The choice of redissolved solution before sample loading directly affects the sensitivity of the method. The 10 μg/L mixed standard working solution was blown dry under nitrogen and redissolved using different volume ratios of 0.1% aqueous formic acid/methanol (90:10, 70:30, 50:50). It was discovered that the increase in the proportion of the organic phase made the target more soluble and significantly improved the response of the sixteen compounds. However, as illustrated in Figure 3a,b, the precision of the quantitative results is easily impacted by the solvent effects when the ratio of organic phase in the redissolved solution is too high. Therefore, 0.1% aqueous formic acid/methanol (90:10, v/v) was finally selected as the redissolved solution.   Figure 4 displays the matrix effects of the sixteen compounds in three livestock substrates. It can be seen that most compounds were inhibited by the three matrices, the inhibition rate was different in size, and the matrix effects of different matrices for the same compound were not all the same. Due to the complexity of food matrices and the requirement for trace analysis of the target, this method used the matrix-matched standard curve method and the isotope internal standard method for quantification, which can reduce the error introduced by the pretreatment to improve the accuracy of quantitative analysis results. hibition rate was different in size, and the matrix effects of different matrices for the same compound were not all the same. Due to the complexity of food matrices and the requirement for trace analysis of the target, this method used the matrix-matched standard curve method and the isotope internal standard method for quantification, which can reduce the error introduced by the pretreatment to improve the accuracy of quantitative analysis results.

Linearity of the Standards Curves, LODs and LOQs
The prepared blank matrix-matched standard solutions were determined online using the established UHPLC-MS/MS method. The results showed a good linearity in the concentration range of 0.1-50 μg/L for the sixteen β-agonists with R 2 ≥ 0.9928, and the linear equations are shown in Table 5. The method LODs ranged from 0.01 to 0.11 μg/kg and the LOQs ranged from 0.04 to 0.38 μg/kg according to the S/N, which was lower than the LODs of β-agonists published in the literature [38,39]. These results further indicated that the present method possessed wide linearity, low LOD, and good precision, which was suitable for the determination of β-agonists in real animal-derived foods.

Linearity of the Standards Curves, LODs and LOQs
The prepared blank matrix-matched standard solutions were determined online using the established UHPLC-MS/MS method. The results showed a good linearity in the concentration range of 0.1-50 µg/L for the sixteen β-agonists with R 2 ≥ 0.9928, and the linear equations are shown in Table 5. The method LODs ranged from 0.01 to 0.11 µg/kg and the LOQs ranged from 0.04 to 0.38 µg/kg according to the S/N, which was lower than the LODs of β-agonists published in the literature [38,39]. These results further indicated that the present method possessed wide linearity, low LOD, and good precision, which was suitable for the determination of β-agonists in real animal-derived foods.

Recovery and Precision
The mean recovery ranges for pork, beef, and lamb at different spiked levels were 62.62-115.93%, 61.35-106.34%, and 62.00-111.83%, respectively, and the interday and intraday precision (RSD) of all compounds in the three matrices was 0.66-9.99% and 2.12%-9.98, respectively (Table A1, Table A2, and Table A3). Tables A1-A3 specifically display the mean recoveries and intraday and interday precision of the sixteen β-agonists in the three matrices, which satisfy the requirements of a recovery range of 60-120% and precision of less than 20% [40]. These results indicate that the method has good accuracy and stability for the detection of the sixteen β-agonists in livestock meat.

Analyses of Actual Sample
During the determination of 30 commercially available samples, a beef sample (sample number: B-4) revealed the presence of CLB, with a residue of 2.45 ± 0.30 µg/kg and detection rate of 3.33%. The chromatogram and mass spectrogram of the positive clenbuterol sample are shown in Figure 5a,b. The blank samples were used as quality control (QC) samples by spiking during the assay. The CLB was added with 5.0 µg/kg into the QC sample, and the determination value was 4.73 ± 0.08 µg/kg with a good recovery. The identification of the positive sample and the validation results of the QC sample further proved the feasibility and accuracy of the method. detection rate of 3.33%. The chromatogram and mass spectrogram of the positive clenbuterol sample are shown in Figure 5a,b. The blank samples were used as quality control (QC) samples by spiking during the assay. The CLB was added with 5.0 μg/kg into the QC sample, and the determination value was 4.73 ± 0.08 μg/kg with a good recovery. The identification of the positive sample and the validation results of the QC sample further proved the feasibility and accuracy of the method.

Conclusions
β-agonist residues in livestock meat have a negative impact on human health. With the strengthening of the monitoring of veterinary drug residues worldwide, the requirements for the detection of β-agonist residues in animal-derived food have gradually increased as well, and simple, rapid, high-throughput and accurate detection methods have gradually become the development direction. In this study, a UHPLC-MS/MS method was developed for the determination of sixteen β-agonists in livestock meat. Optimized enzymatic conditions during sample pretreatment greatly reduced the time of the assay. Following extraction, the sample was purified using the Qvet-AG one-step column, which improved the detection efficiency while also reducing the consumption of reagents. Finally, the sample was used for UHPLC-MS/MS determination, the matrix-matched standard curve method and isotope internal standard method for quantification. The results showed a good linearity in the range of 0.1-50 μg/L for the sixteen β-agonists, with R 2 ≥ 0.9928. The LODs of the method were 0.01-0.11 μg/kg and the LOQs were 0.04-0.38 μg/kg. In the pork, beef, and lamb matrices, the spiking recoveries of the three levels of the βagonists were 62.62-115.93%, 61.35-106.34%, and 62.00-111.83%, respectively. Both the intraday and interday precision fell under 10%. The established method is sensitive and reproducible, with a short analysis time, and high detection quantity, which is suitable for the routine detection and confirmatory analysis of β-agonists in animal-derived foods and helps in the regulation of β-agonist residues in the market.

Conclusions
β-agonist residues in livestock meat have a negative impact on human health. With the strengthening of the monitoring of veterinary drug residues worldwide, the requirements for the detection of β-agonist residues in animal-derived food have gradually increased as well, and simple, rapid, high-throughput and accurate detection methods have gradually become the development direction. In this study, a UHPLC-MS/MS method was developed for the determination of sixteen β-agonists in livestock meat. Optimized enzymatic conditions during sample pretreatment greatly reduced the time of the assay. Following extraction, the sample was purified using the Qvet-AG one-step column, which improved the detection efficiency while also reducing the consumption of reagents. Finally, the sample was used for UHPLC-MS/MS determination, the matrix-matched standard curve method and isotope internal standard method for quantification. The results showed a good linearity in the range of 0.1-50 µg/L for the sixteen β-agonists, with R 2 ≥ 0.9928. The LODs of the method were 0.01-0.11 µg/kg and the LOQs were 0.04-0.38 µg/kg. In the pork, beef, and lamb matrices, the spiking recoveries of the three levels of the β-agonists were 62.62-115.93%, 61.35-106.34%, and 62.00-111.83%, respectively. Both the intraday and interday precision fell under 10%. The established method is sensitive and reproducible, with a short analysis time, and high detection quantity, which is suitable for the routine detection and confirmatory analysis of β-agonists in animal-derived foods and helps in the regulation of β-agonist residues in the market.

Conflicts of Interest:
Author Jun Ning was employed by the company of the WENS Foodstuff Group Co., Ltd. (Guangdong China). The company had the following involvement with the study: software, val-idation, funding and the decision to submit it for publication. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Appendix A Table A1. Average recovery, intraday precision, and interday precision of the sixteen β-agonists in pork samples (n = 6).

Compound
Added