Comparison of three different lipid removal cleanup techniques prior to the analysis of sulfonamide drug residues in porcine tissues

Abstract A number of 17 sulfonamides (SNs) determination in porcine tissues using two new materials including Enhanced Matrix Removal for Lipid (EMR‐L) and Oasis PRiME hydrophilic‐lipophilic balance (HLB), and the conventional liquid–liquid extraction with n‐hexane (LLE) sample preparation methods were evaluated and compared. Samples were extracted uniformly with acidified acetonitrile and cleaned up by the three sample preparation methods, and then, analytes were further separated by ultrahigh‐performance liquid chromatography (UHPLC) and detected by a triple quadrupole mass spectrometry (MS/MS) or a quadrupole‐time‐of‐flight tandem mass spectrometry (Q‐TOF/MS). The results showed that the matrix effects from the EMR‐L and HLB were significantly lower than that from LLE, suggesting that these two new materials are superior to n‐hexane in the precipitation of proteins and the adsorption of lipids. Moreover, the recoveries of 17 SNs were quantified by the matrix‐matched calibration curve at spiked level of 5, 10, and 20 μg/kg, and 97.0% of the results satisfied method validation requirements. The samples cleaned up by EMR‐L and HLB achieved the highest average recovery in liver and kidney with high moisture content, and muscle which is high in fat, respectively. In addition, Q‐TOF/MS could play a good role in aided verification based on the result of repeated validation test. In summary, either combination of approaches could be used to achieve monitoring purposes; it is still worthwhile to adopt a specific sample preparation method and MS detector for the quantification in a specific matrix.

Every extraction method has its own advantage, for instance, LLE functions achieve the purpose of extraction and cleanup mainly through a variety of analytes using the mutual insolubility of two different solvents (Li et al., 2019;Zhan et al., 2012). SPE is a method of purifying the extraction based on the physical and chemical properties of each analyte to select the appropriate materials and extraction conditions (Peters, Bolck, Rutgers, Stolker, & Nielen, 2009). Acetonitrile (MeCN) or acidified acetonitrile was served as extractant, C 18 , primary secondary amine (PSA), graphene were typically used as adsorbents for QuEChERS, which is commonly applied to sample preparation of veterinary drugs with enhanced analytical efficiency (Rúbies, Antkowiak, Granados, Companyó, & Centrich, 2015;Wang & Leung, 2012;Zhang et al., 2016). Despite many previous studies, there are still some claimed shortcomings. For example, LLE requires the use of excessive organic reagents, SPE blocks column easily, and QuEChERS has less selectivity for some substances (Garrido Frenich, Aguileraluiz, Martínez Vidal, & Romerogonzález, 2010;Wang et al., 2011;Yoshikawa et al., 2017). Aiming at these problems, a new product from Agilent Technologies called "Enhanced Matrix Removal for Lipid" (EMR-L) could selectively capture the long hydrocarbon chain of lipids (Han, Matarrita, Sapozhnikova, & Lehotay, 2016). Cleanup ability of QuEChERS and protein precipitation extracts were well integrated by EMR-L due to the activated material (López-Blanco et al., 2016).
Meanwhile, Oasis PRiME HLB (HLB) is a popular representative reverse-phase SPE adsorbent developed by Waters, which is commonly used in the cleanup of trace substances. HLB does not require activated and balanced because of special solid-phase packing (Chen et al., 2016;Han et al., 2015). Currently, EMR-L and HLB have been widely used in the detection of veterinary drug residues in animal-derived foods (Zhao, Lucas, Long, Richter, & Stevens, 2018). Nevertheless, to the best of our knowledge, no reports have been available on the profiles for the comparison of performances of EMR-L and HLB methods.
Ultrahigh-performance liquid chromatography coupled to triple quadrupole mass spectrometry (UHPLC-MS/MS) technology provides a broad-based method for the widest range of veterinary drugs and dominates in the field of veterinary drug analysis . It can simultaneously identify and quantify hundreds of targeted analytes to meet the daily monitoring needs of food contaminants (Park et al., 2015). Meanwhile, high-resolution MS (HRMS) instruments such as quadrupole-time-of-flight tandem mass spectrometers (Q-TOF/MS) provide accurate mass measurements, defined as the ability to determine the mass of a compound within a few parts-per-million (ppm) of its theoretical mass (Turnipseed, Lohne, & Boison, 2015). The applicability of Q-TOF/MS for the determination of veterinary drugs in animal food products at or below the levels required by regulatory agencies has been demonstrated (Rodriguez-Cabo, Rodriguez, Ramil, & Cela, 2016).
The aim of this study was to systematically evaluate the analytical and practical performance of three sample preparation methods (EMR-L, HLB, and LLE) by UHPLC-MS/MS in the analysis of 17 SNs spiked into extracts of porcine tissues (muscle, liver, and kidney) at regulatory levels. The feasibility of reducing false-positive results by Q-TOF/MS had also been validated.

| Chemicals and reagents
MeCN, methanol, n-hexane, and formic acid were supplied by Merck.
Ultrapure water was prepared with a Milli-Q system by Millipore. All solvents were in chromatographic grade. were purchased from Sigma-Aldrich. Stock solutions were prepared at a concentration of 100 µg/ml in methanol, and mixed standard solutions were diluted to 1,000 ng/ml. All solutions were stored at −20°C.
Oasis PRiME HLB (6cc, 200 mg) was purchased from Waters. EMR-L product was purchased from Agilent Technologies, which consisted of two 15-ml centrifuge tubes. QuEChERS d-SPE EMR-L contained 1 g EMR-L material, and QuEChERS Final Polish EMR-L contained 1.6 g magnesium sulfate and 0.4 g sodium chloride for salting out.

| Sample preparation
In a comparison of the three methods, 2 g of samples (n = 6 for each) from muscle, liver, and kidney were minced and homogenized before weighed into 50-ml centrifuge tubes spiked with SNs and extracted with 10 ml MeCN containing 0.1% formic acid. The samples were vortex-mixed for 5 min at room temperature using a pulsating vortex platform shaker (Anting). Then, the sample tubes were placed in an ultrasonic bath (Xinzhi) at 70°C for 10 min to extract the veterinary drugs more fully out of the matrix through the ultrasonic-assisted extraction. Extracts were centrifuged at 2,862 rcf for 10 min, and the obtained upper layer was decanted into a new centrifuge tube for the further cleanup as follows.

| Enhanced matrix removal for lipid method
Ultrapure water (5 ml) was added to the QuEChERS d-SPE EMR-L tube and vortex-mixed immediately for 15 s, in order to activate the EMR-L material. The upper layer (5 ml) was transferred into the tube and vortex-mixed for 5 min and ultrasonically cleaned for 10 min, then centrifuged at 2,862 rcf for 10 min.
In the salting-out step, 5 ml of upper layer was transferred into the QuEChERS Final Polish EMR-L tube. These sample tubes were vortexed, ultrasonically cleaned, and centrifuged as described above.
Thereafter, 5 ml of upper layer was pipetted into 15-ml graduated glass centrifuge tubes evaporated to near dryness at 40°C under N 2 flow, and then reconstituted to 1 ml by MeCN.

| Hydrophilic-lipophilic balance method
The upper layer (5 ml) was loaded into the Oasis PRiME HLB tube under gravity and the last few drops of extractant gently pushed through with a pipet bulb. Extracts were evaporated to dryness and reconstituted as described above.

| Liquid-liquid extraction method
MeCN-saturated n-Hexane (10 ml) was added into the polypropylene centrifuge tube that contained the upper layer and vortex-mixed immediately for 30 s. Then, the upper layer n-hexane was transferred, and the above steps were repeated once to clean up the lipid.
The final subnatant was evaporated to dryness and reconstituted in the same step.
The data station operating software was MassLynx ver-  Table 1.   To calculate matrix effects, the slope of the SS calibration curves was compared with the slope of the MM calibration curves

| Comparison of extractives
Comparison of extractives is an important way to evaluate the feasibility of sample preparation method for high-throughput screening. Extractives (1 ml) were, respectively, dropped into a glass tube and dried in the oven at 80°C, and the extractives of muscle, liver, and kidney were processed by the three sample preparation methods ( Figure 1). According to the cleanliness of the bottom of the glass test tube, the EMR-L and HLB gave cleaner matrix extractives than LLE method, due to their specific material. The EMR-L tubes consisted of magnesium sulfate, special polymeric materials, and sodium chloride. The salts in the EMR-L extractives were brought in by salting-out step (Kaufmann, Butcher, Maden, Walker, & Widmer, 2014). Moreover, HLB could effectively adsorb lipids due to the polymerized solid-phase packing of lipophilic divinylbenzene and hydrophilic N-Vinylpyrrolidone . Therefore, samples with higher moisture content such as liver and kidney cleaned up by EMR-L gave the same cleanliness as HLB-cleaned up muscle.

| Matrix effects
Matrix effects are common in LC-MS analysis of complex samples, especially in ESI + mode (Dasenaki & Thomaidis, 2015). Appropriate acidity can promote the mass spectrometer ionization efficiency of ESI + (Gómez Pérez, Romero-González, Martínez Vidal, & Garrido, 2014), and MeCN is a suitable solvent for extracting SNs and has the function of precipitating proteins to improvement of the ionization efficiency of the target analyte. In order to facilitate the quantification of this experiment, the method of using acidified MeCN extraction followed by cleanup with different adsorbents was adopted.
Matrix-matched calibration, which was widely used to reduce matrix effects in practice, was performed in this study. Matrix effects for all compounds which were detected by UHPLC-MS/MS are presented in Table 2. It can be seen that matrix effects for most of analytes prepared using three different methods rarely exceed ±20%, and most analytes produced a slight ionization suppression. However, SD and SIT for LLE method were subjected to significant signal suppression in muscle and liver, respectively. The matrix effects of LLE were more severe than the EMR-L and HLB cleanup methods. For a sample individual, the matrix effects of muscle, liver, and kidney are not regular. For example, Tiele, Maiara, Osmar, Martha, and Renato (2016) used the LLE method to detect veterinary drug residues in liver and kidney, and the matrix effect of liver samples was higher than that of kidneys. Granelli and Branzell (2007) developed an LLE method for detecting antibiotics in muscle and kidney. For some antibiotics, the matrix effect in muscle was higher than in the kidney.

| Linearity, LODs, LOQs, Recoveries, and RSDs
MM calibration was used to better quantitate these analytes when evaluating sample preparation methods in this study (Valese et al., 2017). LODs and LOQs were determined by separately considering  (Deng et al., 2011;Xie et al., 2012). This means that these three methods for detecting SNs residues meet industry standards.   (Lehotay et al., 2012). In any case, the overall recovery meets the methodological validation, showing the feasibility of porcine matrix analyzed by UHPLC-MS/MS.

| Repeated validation test by Q-TOF/MS
The mass of ions could be accurately determined by using the Q-TOF/ MS, which is helpful for rapid screening. As a broad-spectrum detec-  Hence, Q-TOF/MS is capable of performing high-throughput screening for daily monitoring of those compounds with higher residue limits (Turnipseed, Storey, Clark, & Miller, 2011).

| Comparison with the national standard method of China
In order to further verify the reliability of the method, the researched methods in this study were compared with the national standard method

| CON CLUS IONS
In this study, the analytical and practical performance of three sample preparation methods (EMR-L, HLB, and LLE) and two analytical methods (UHPLC-MS/MS and Q-TOF/MS) in the analysis of 17 SNs spiked into extracts of porcine tissues (muscle, liver, and kidney) were systematically evaluated. According to our results, EMR-L and HLB were recommended to pretreat high moisture content or lipid matrix, respectively, as the result shown in matrix effects. The overall recovery met the methodological validation. LLE showed lower recovery and higher matrix effects in the SNs analysis compared with the other sample preparation methods, although it could be used as a rapid screening.
In addition, Q-TOF/MS results were confirmed that it can play a good role in aided verification based on the result of repeated validation test.
This study, therefore, demonstrates the feasibility of application of commercial product, HLB and EMR-L, compared with the traditional LLE technology.

ACK N OWLED G M ENTS
This work was financially supported by Jiangsu Agriculture Science

CO N FLI C T O F I NTE R E S T
No conflict of interest declared.

E TH I C A L A PPROVA L
This study does not involve any human or animal testing.