Analysis of 14 β -agonists in pork using automated wooden tip based SPME device and UPLC-MS/MS

This study presents an innovative analytical method for the detection and quanti�cation of 14 β -agonists in pork using liquid chromatography-tandem mass spectrometry (LC-MS/MS). The method involves the use of a solid phase microextraction (SPME) probe, consisting of a wooden tip coated with polyacrylonitrile (PAN) and molecularly imprinted polymer (MIP). This probe has been integrated into an automated extraction device, designed in our laboratory to extract the target analytes prior to LC-MS/MS analysis. Compared to the traditional solid-phase extraction (SPE) procedure, the proposed approach signi�cantly reduces the costs related to adsorbent material and solvent consumption, and shortens puri�cation time, thereby reducing both the technical workload and associated expenses. The methodology has been carefully validated in accordance with the National Standard of the People's Republic of China (GB/T 22286-2008), evaluating key parameters including the limit of detection (LOD), limit of quanti�cation (LOQ), matrix effects, linearity, and both intra-day and inter-day precision. The average recovery rates for all β -agonists were observed to lie between 71.6% and 82.2%, with relative standard deviations (RSDs) consistently below 15% for both types of precision. The LOD and LOQ values were found to range from 0.09 to 0.39 μ g/kg and 0.27 to 0.99 μ g/kg, respectively. In terms of accurate detection of positive samples, the developed method demonstrated a closer approximation to the true value of the sample compared to the National Standard method GB/T 31658.22-2022. Overall, the proposed method exhibits substantial promise for the routine assessment of β - agonists in pork and presents a practical alternative for compliance monitoring in regulatory settings.


INTRODUCTION
β-agonists represent a class of drugs that are characterized by their phenylethanolamine structure, and are primarily used to treat bronchial diseases (Yan et al., 2022).Extensive studies in animal models indicate that administration of β-agonists at levels 5-10 times greater than the recommended therapeutic dose can signi cantly enhance feed conversion rates and lipolysis, while increasing protein synthesis (Juan et al., 2010).Prolonged use of β-agonistssupplemented feeds has been observed to improve feed conversion e ciency while increasing animal body weight and muscle tissue content, thereby boosting the rate of lean meat production (Centner et al., 2014).This demonstrates the ability of β-agonists to redirect nutrients within the organism.Therefore, in the livestock industry, β-agonists are often incorporated into feeds or drinking water to accelerate animal growth and increase lean meat yield.Clenbuterol, ractopamine, and salbutamol are the most used compounds.Recently, with the advent of additional β-agonists, drugs such as semantrol and bambuterol have also been exploited in animal husbandry to stimulate growth.Nevertheless, these compounds may accumulate in the animal's body over time, leading to potential toxicity and adverse side effects from prolonged exposure.Excessive use of β-agonists in pigs, cattle, sheep, and other livestock has led to the bioaccumulation of these compounds in the animals.Consequently, when humans consume food derived from these animals, residues of β-agonists may pose signi cant health risks (Brambilla et al., 2000).In severe cases, acute toxic reactions can even occur, with common symptoms including muscle tremor or pain, tachycardia, headache, dizziness, and nausea.In extremely serious situations, the toxicity may induce shock and potentially lead to death (Giannetti et al., 2016;Guo et al., 2018).
To ensure food safety, the use of clenbuterol as a growth promoter is currently prohibited in several jurisdictions, including China, the United States, and the European Union (Kuiper et al., 1998).European legislation sets a maximum residue limit (MRL) of 0.5 µg/kg for clenbuterol in horse and bovine liver(Commission Regulation (EU) No 37/2010).This threshold is consistent with China's national standard, GB/T5009.192-2003,which stipulates a limit of detection (LOD) for clenbuterol in animal food of 0.5 µg/kg (GB/T5009.192-2003).Moreover, ractopamine (RAC) is also forbidden as a feed additive in the European Union, Japan, China, and most countries worldwide (Zhang et al., 2016).
However, its use is permitted in Canada, the United States, and Brazil, where it is approved for utilization in cattle, turkey, and swine (Zhang et al., 2016).According to the recommendations of the Joint FAO/WHO Expert Committee on Food Additives (JECFA), the MRLs for ractopamine in muscle, fat, liver, and kidney tissues are set at 10, 10, 40, and 90 µg/kg, respectively (Zhang et al., 2016).They also establish the acceptable daily intake (ADI) at a range of 0-1 µg/kg body weight per day (Zhang et al., 2016).Furthermore, China and the European Union have generally proscribed the use of all β-agonists as growth promoters and have instituted several regulations to deter and manage their illicit use in the animal husbandry industry.These policies collectively underscore the global commitment to mitigating potential public health risks associated with the consumption of animal products containing β-agonist residues.Under these circumstances, it is of paramount importance to develop rapid, precise, and reliable method for the detection of residues of β-agonists in animal foods.
In modern analytical chemistry, several instruments, including capillary electrophoresis (CE) (He and Fan, 2019), gas chromatography-tandem mass spectrometry (GC-MS/MS) (Cao et al., 2020), liquid chromatography-tandem mass spectrometry (LC-MS/MS) (Zhu et al., 2019;Kaufmann et al., 2021), and nanosensors (Li et al., 2020) have proven themselves preferred techniques for the detection of β-agonists.Among these technology, LC-MS/MS is characterized by its superior sensitivity, precision, and reproducibility, making it mostly adopted (Yan et al., 2022;Castilla-Fernández et al., 2019;Hsieh et al., 2023).To improve the detection and quanti cation limits of LC-MS/MS and to mitigate matrix effects that could bias the detection results, several sample pretreatment methods have been developed.These include the Quick, Easy, Cheap, Effective, Rugged, and Safe (QuEChERS) method (Yang et al., 2019), solidphase extraction (SPE) (Li-Qi Wang et al., 2012;Liqi Wang et al., 2013;Kulikovskii et al., 2016), disposable Pipette Extraction (DPX) (Mastrianni et al., 2018), and solid-phase micro-extraction (SPME) (Peng et al., 2022).While these strategies are effectively in reducing matrix interferences and enabling higher sample throughput, they are not without challenges.For example, these methods often require signi cant sample and solvent usage, incur high cost for puri cation material costs, and are labor-intensive to perform.
In a previous study, our team presented a rough, cost-effective, and easily fabricated wooden-tip-based SPME probe designed for detecting veterinary drug residues in complex matrices (Gong et al., 2022).This SPME was constructed using a simple dip-coating method, applying adsorbents to the surface of the wooden tip.Compared to the more ubiquitous SPE techniques, our SPME methodology offers several advantages: it streamlines the sample clean-up process, reduces solvent consumption, requires less adsorption material, and lowers the overall cost of the assay.
Moreover, the SPME probe can be reused after cleaning.To further increase the effectiveness of this SPME probe, we have designed and developed an automated apparatus to complement its operations.The aim of this study is to increase detection e ciency and save analysis costs.This is achieved by integrating our automated pre-treatment device with LC-MS/MS for the simultaneous detection of 14 β-agonists in pork samples.

Standard solutions
Standard Stock Solution: Accurately weigh 100 μL of the mixed standard solution accurately and make up the concentration of 1 μg/mL with methanol.A series of working solutions (mixed standard solutions and matrixmatched standard solutions) were prepared at concentrations of 1, 2, 5, 10, 15, 20 ng/L.All stock and working standard solutions were stored at −20°C until further use.

Preparation of PAN/MIP-Coated Wooden-Tip-Based SPME
Dissolve 10 g of PAN powder in 100 mL of dimethylformamide (DMF) at room temperature, stirring constantly to ensure uniform dissolution.Mix 1g MIP and 10 mL of PAN solution to obtain a PAN-MIP slurry.Remove the cotton from the tip of the medicinal swab, dip the wooden stick into the PAN-MIP slurry about 1 cm deep so that the glue is evenly coated on the wooden tip, and then dip and coat again after it has dried slightly.

Sample preparation
For sample preparation, 2 g of a homogenized sample (accurate to 0.01 g) was weighed and transferred to a 10 mL polypropylene centrifuge tube.Thereto was added 5 mL of extraction solution consisting of an aqueous ammonium acetate/acetic acid solu-tion (0.77 g ammonium acetate and 0.03 mL acetic acid) together with 100 μL of βglucuronidase/sulfatase.The mixture was then vortexed at 2000 rpm for 1 minute.Sub-sequently, it was incubated in a water bath at 37℃ for 2 hours, shielded from light.Fol-lowing incubation, the solution was centrifuged 4℃ and 12000 rpm for 5 minutes and 2 mL of upper solution containing the sample was collected for further puri cation.

Automated Wooden-tip SPME Procedure
Attach the prepared PAN-MIP tips to the automated extraction device and program the device to automatically perform the following sequence: a. Activation: 0.5 mL of solvent consisting of a mixture of acetic acid and methanol in a 1:9 ratio, spinning at 1000 rpm for 1 minute.Subsequently, spin in 0.5 mL of methanol solution at 1000 rpm for 1 minute.b.Equilibration: 0.5 mL of water, spinning at 1000 rpm for 1 minute.c.Adsorption: 2 mL of the sample, spinning at 1000 rpm for 2 minutes.d.Wash: 0.5 mL of water, spinning at 1000 rpm for 1 minute.e. Desorption: 0.5 mL (mixture of acetic acid and methanol in a 1:9 ratio), spinning at 1000 rpm for 2 minutes.
The sequential extraction steps are depicted schematically in Figure 2.After the desorption process, the samples were dried with nitrogen gas.The volume was then brought to 0.5 mL with a solution of 0.1% aqueous formic acid in a 30:70 methanol/formic acid ratio for LC-MS/MS analysis.
A triple quadrupole mass spectrometer (AB Science Qtrap 5500) was used for quanti-cation of the target compounds.Data collection was monitored using the Analyst 1.6.2software.The optimized source parameters are as follows: • spry voltage of 5.5 kV • vaporizer temperature of 450 ℃ • curtain gas (CUR) pressure of 35 psi • ion-source gas 1 (GS1, nebulizer gas) pressure of 55 psi • ion-source gas 2 (GS2, turbo gas) pressure of 55 psi All β-agonists were tested in the MRM mode with positive ESI.The compound de-pendent parameters are listed in Table 1.

Wooden tip SPME surface coating optimization
It has been reported that PAN, a biocompatible polymer, is combined with HLB to provide excellent extraction of a wide range of compounds (Reyes-Garcés et al., 2014).The MIP used in this work exhibit a robust selectivity for βagonists.Therefore, the ratio of PAN to MIP needs to be optimized to enhance the adsorption rate of the target compounds.In this study, the MIP were initially dosed at 1g, and PAN was then mixed with 10 mL of DMF at 1:5, 1:10, 1:20, and 1:30 (g:mL) ratios.It was observed that the 1:5 PAN to DMF mixture was overly viscous, leading to coatings on the tips that were prone to breakage.In contrast, a ratio of 1:30 resulted in a less viscous mixture, and the coating formed could be easily peeled off.A 1:10 ratio gave a suspension of suitable viscosity and allowed the wooden tip to be coated relatively evenly.Thus, the ratio of PAN to DMF was xed at 1:10 (10 mL), and MIP were added at dosages of 0.25g, 0.50g, 0.75g and 1 g.In the dip-coating process, the amount of MIP added signi cantly in uenced the coating characteristics.When the MIP level was 0.25g, 0.50g, or 0.75g, the slurry was thinner, requiring double application of the coating.Conversely, when incorporating 1g of MIP, a single application was su cient.From a visual perspective, it was observed that the coating was more uniform when the MIP addition amount were 0.50g, 0.75g, and 1g (Figure S1).
To further evaluate the adsorption e ciencies of four SPME tips (a: 0.25 g MIP, b:0.50 g MIP, c:0.75 g MIP, d:1g MIP) for all agonists, an experimental investigation was conducted (Figure S1).The blank matrix spiked sample solution, after pretreatment and at a spiked concentration of 20 μg/kg, was divided into two parts.The rst part was analyzed using the method prescribed in GB31658.22-202222- (GB/T 31658.22-2022)),resulting in the quanti cation of agonists content, denoted as W 0 .The second part was initially treated with the method established in this study, followed by further puri cation step as mentioned in GB31658.22-2022, which allowed the calculation of the agonist drugs content, W 1 .The adsorption rate k then derived using the formula k=(W 0 -W 1 )/W 0 ×100%.The results demonstrated that the tip c yielded satisfactory outcomes, and further increase in MIP content did not consistently enhance the adsorption rate of SPME.Consequently, tip c was selected for subsequent studies and the absorption rate ranged from 70% to 89% for all 14 drugs, as detailed in Table S2.The desorption rates of 14 β-agonists adsorbed on tip c, as assessed in mixed solvents, are detailed in Table S2.These rates ranged from 85.6% to 97.0%, signifying that the target substances could be adequately eluted.

Method validation
The method proposed in this work has been validated according to the Chinese national standard (GB/T 27404-2008).Validation including matrix effect, linearity, accuracy, precision, limit of detection (LOD), and limit of quanti cation (LOQ).

Matrix effect (ME):
The slope of the solvent calibration curve and the matrix-matched blank extract calibration curve were used to determine the ME in pork samples, according to the equation: ME (%) = [(slope of matrix-matched calibration curve-slope of solvent standard calibration curve)/slope of solvent standard calibration curve] ×100.When the matrix effect (ME) value exceeded 120%, this indicated an enhancing effect, re ecting an augmentation in the analytical signal.On the other hand, if the ME was below 80%, this denoted an inhibitory effect, signifying a suppression of the signal.A ME value in the range of 80% to 120% was considered statistically insigni cant, meaning that the matrix did not signi cantly affect the analytical response.On speci c analysis of the 14 β-agonists, it was found that the ME values were all within this range, con rming that the matrix effect was not a signi cant factor in the observed results (Table 2).Linearity: Calibration curves were established by plotting the peak area ratios of the target compounds versus the concentrations of the corresponding calibration standards at six concentrations (1, 2, 5, 10, 15, 20 ng/L) with six replicates at each level.The linearity was determined by the correlation coe cients of the calibration curves.For each analyte, plots of instrument response versus concentration were generated and the linear equations and correlation coe cients determined by linear regression analysis.For all 14 analytes, regression lines with coe cient of determination (r 2 ) above 0.9993 were obtained as shown in Table 2.
Limits of Detection and Quanti cation: The LODs and LOQs for the method were de ned corresponding to the signalto-noise ratios of 3 and 10, respectively.LODs for the 14 targets ranged from 0.09 to 0.39 μg/kg, and the LOQs varied from 0.27 to 0.99 μg/kg.These ranges are su cient to meet the demands of both qualitative and quantitative analyses.
Chromatograms of 14 kinds of β-agonists are shown in Figure 3.

Analysis of positive sample
To assess the accuracy of the method with respect to positive samples, the method developed in this study was utilized to analyze pro ciency test sample provided by CNAS (No. ACAS-PT-1048) for the determination of 3 βagonists in pork.The results, presented in Table 4, demonstrate that the method described here is in better agreement with the actual values than the Chinese national GB standard method (GB/T 31658.22-2022),which employs a solidphase extraction (SPE) puri cation step.Furthermore, the method pro-posed in this study offers distinct advantages over the GB/T 31658.22-2022standard, particularly in aspects such as the reduction of adsorbent material cost, minimization of sol-vent consumption, decrease in puri cation time, and enhancement in automation, as de-tailed in Table 5.

CONCLUDING REMARKS
In alignment with the principles of green analytical chemistry (GAC), which emphasizes sustainability and environmental responsibility, we have developed an automated SPME technique for the detection of 14 β-agonists in pork.This innovative approach offers several distinct advantages over the conventional national standard method GB31658.22-2022,currently prevalent in China.Speci cally, our method reduces adsorbent material costs, minimizes organic solvent consumption, decreases puri cation time, and lowers labor costs.Additionally, the accuracy of our method has been shown to be closer to true values when testing real samples compared to GB31658.22-2022.
Considering these factors, our method presents a viable and more e cient alternative that could be adopted by regulatory agencies and applied in analytical laboratories. Figures

Table 2
Linear range, correlation coefficient, and calibration curves of 14 β-agonists

Table 3
Average recovery rates of all target compounds spiked in pork at three different concentrations (n=6) Table 5 Comparison of Adsorbent Material Cost, Solvent Consumption, Purification Time, and Automation Between the Proposed Method and the GB/T 31658.22-2022Standard