Rapid determination of heterocyclic amines in ruminant meats using accelerated solvent extraction and ultra-high performance liquid chromatograph–mass spectrometry

Graphical abstract


Method details
Heterocyclic amines (HCAs) are one class of harmful compounds that could be produced during meat preparation by grilling. A number of epidemiological studies conducted in rodents have demonstrated the development of cancer in different organs and tissues following dietary exposure to HCAs [2,3]. Most notably, cancers of the colon, liver, oral cavity, hematopoietic system, and lymphoid system [4]. Furthermore, dietary intake of HCAs through red meat consumption has also been shown to increase the risk factor for cancers in humans. Thus, strategies for accurately quantifying HCAs in cooked meats are a major area of research interest to improve the safety of grilled meats.
Analysis of HCAs requires their prior extraction from food matrices. There is a plethora of methods in the literature for extraction of HCAs in cooked foods including liquid-liquid extraction, solid phase extraction (SPE), and tandem solid phase extraction to purify HCAs prior to analysis [5][6][7][8]. These methods require efficient tedious sample pre-treatments to remove interfering meat components such as fats, inorganic salts, proteins, and carbohydrates followed by sample pre-concentration steps to detect HCAs which typically occur at low parts-per-billion-levels in cooked foods. This makes existing sample preparation methods for HCA analysis in cooked foods laborious, time-consuming and unsuitable for multiple sample analysis. Pressurized liquid extraction is a promising alternative to traditional HCA extraction methods, where more polar solvents can be used at elevated temperatures and pressures to selectively and efficiently extract target compounds [9][10][11]. Two analytical methods are at the forefront of HCA analysis in cooked foods including liquid chromatography coupled with mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). GC-MS is sensitive, but prior derivatization of HCAs is necessary for GC analysis, which is time-consuming and not all HCAs can be derivatized completely, limiting the application of GC-MS for HCA analysis. Liquid chromatography-tandem mass spectrometry (LC-MS/MS) is a preferred method, owing to its high selectivity, superior sensitivity and analysis without prior derivatization treatments [12]. To the best of our knowledge, there exist but few methods in the literature that are based on pressurized liquid extraction-liquid chromatography-high resolution accurate mass tandem mass spectrometry (PLE-LC-HRAMS/MS) for HCA analysis [1,5,13]. These methods used dichloromethane/ acetonitrile (1:1, v/v) [1] and dichloromethane/acetone, 50/50 (v/v) [13] for HCA extraction followed by filtration of extracted analytes, identification and quantification of HCAs on a liquid chromatography-mass spectrometer. Significantly, these methods were considered to be fast, reliable and require sample pretreatments using solid phase extraction (SPE), or liquid extraction prior LC/MS analysis.
Our modified pressurized liquid extraction method is based on pure methanol for pressurized solvent extraction of HCAs and UHPLC-HRAMS/MS analysis. In this method, HCA containing samples are digested in basic methanol and selectively extracted on a Dionex ASE 350 extractor (ThemoScientific, Brampton, On, Canada). Extracted analytes are concentrated during extraction, filtered (Mini-UniPrepTM G2 syringeless filter, 0.2 mm, Whatman, Buckinghamshire, UK) and submitted to UHPLC-HRAMS/MS analysis without the need for sample pre-treatments. The method was fast, and showed good accuracy and reliability. Furthermore, use of automated pressurized liquid extractor (Dione ASE 350) would allow for multi-sample analysis of HCAs.

Extraction of HCA standards by accelerated solvent extractor (ASE)
Aliquots (50 mg/L) of working standard mixture of HCAs containing internal standard (TriMeIQx = 0.05 mg/mL) were used as quality control for the extraction of samples with the accelerated solvent extractor (Dionex ASE 350, Thermo Scientific, MO, USA). Briefly, 25 uL of 5000 mg/L (5 mg/mL) working standard mixture was mixed with 2.5 mL of 0.5 M NaOH in MeOH/Water (70:30 v/v) and incubated for 2 h at room temperature with stirring in a 5 mL beaker. Diatomaceous earth (1:2 v/w) was added to the mixture and the paste loaded into 10 mL stainless steel ASE extraction cells for extraction with pure methanol. The ASE extraction program was set as follows: 2 cycles, 5 min heating time, 80 C extraction temperature, 160 s purge time, and 1606 psi static pressure. The extracted analytes in 20 mL of methanol were evaporated to dryness with a rotary evaporator and nitrogen gas at room temperature. HCAs were re-suspended in 2.5 mL methanol and filtered (Mini-UniPrepTM G2 syringeless filter, 0.2 mm, Whatman, Buckinghamshire, UK) prior to analysis by UHPLC-HRMS/MS. The ASE extractions were performed in quadruplicates (n = 4) for each working standard mixture.

Ultra-high performance liquid chromatography-high resolution accurate mass tandem spectrometric analysis of HCAs
The UHPLC-HRAMS/MS analysis was performed on a LTQ Orbitrap XL mass spectrometer with an automated Dionex UltiMate 3000 UHPLC system controlled by Chromeleon software (Thermo Scientific, MO, USA). A Luna C18 column (100 Â 2.0 mm I.D., particle size: 3 mm, pore diameter: 100 Å) purchased from Phenomenex (CA, USA) was used for HCAs separation. The solvent system used on the analytical column was: H 2 O, 30 mM ammonium formate (pH 3.2) as solvent A and pure acetonitrile as solvent B. Chromatographic separation was carried out at 20 C with a flow rate of 0.2 mL/min, and 10 mL of sample was injected in the instrument. The gradient used for separation was as follows: 0À1 min 10 % B, 1À3 min 10-20 % B, 3À6 min 20-30 % B, 6À9 min 30-40 % B, 9À12 min 40 % B, 12À13 min 40-50 % B, and re-equilibrated at 90 % A for 2 min. The Orbitrap mass spectrometer was operated in the positive ESI mode. The following optimized parameters were used for the Orbitrap mass spectrometer, sheath gas: 8, auxiliary gas: 2, ion spray voltage: 4.50 kV, capillary temperature: 320 0 C; S-lens RF: 100 V; capillary voltage: 30 V, mass range: 100À1000 m/z; full scan mode at a resolution of 60,000 m/z; top-3 data dependent MS/MS at a resolution of 30,000 m/z and collision incubated for 2 h at room temperature with stirring. Diatomaceous earth (1:2 v/w) was added to the mixture and the paste loaded into 10 mL stainless steel ASE extraction cells, and extracted with an accelerated solvent extractor (Dionex ASE 350, Thermo Scientific, MO, USA). Extracted analytes were recovered after roto-evaporation, resuspended in 2.5 mL pure methanol and analyzed by UHPLC-HRMS/MS on a LTQ Orbitrap XL mass spectrometer (Thermo Scientific, MO, USA) with an automated Dionex UltiMate 3000 UHPLC system as previously described for HCA standards.

Method applicability: HCAs in grilled beef and moose meats
Beef (Bovinae) and moose (Cervidae) striploin steaks (longissimus muscle) were obtained from a local market and from Newfoundland and Labrador Department of Natural Resources respectively. Moose steaks were taken from 4 different animals while beef steaks purchased from 4 different batches of beef were used to mitigate any inherent variability of the meat sources. Ethics approval for this study was granted by Memorial University Animal Care Committee as mandated by the Canadian Council on Animal Care, and all the experiments were performed in accordance with relevant guidelines and regulations. Steaks (1 lb) of beef (B) and moose (M) meat from different batches were cut and divided into four replicates (n = 4), and kept under the same conditions until grilling time [14].

Cooking conditions
Beef and moose samples were grilled at 200À250 C for 25 min on a grill (Cuisinart 1 Gourmet 600B, Canadian Tire, NL) reaching an internal temperature of 75 C. The grill was thoroughly cleaned between samples to avoid any possible cross contamination from meat exudates or trimmings. Meat samples were turned regularly during grilling. After grilling, each replicate was divided into two subsets. One subset was cut into two-inch cubes and used for sensory analysis while the other subset was labeled and stored at À80 C for HCA analysis [14].

Extraction of heterocyclic amines in grilled meats
HCAs were extracted with an accelerated solvent extractor (Dionex ASE 350, Thermo Scientific, MO, USA). Grilled meat (1 g) was ground, spiked with internal standard (TriMeIQx = 0.05 mg/mL), mixed with 0.5 M NaOH in MeOH/Water (70:30 v/v) and incubated for 2 h at room temperature. Diatomaceous earth (1:2 w/w) was added to the mixture and HCAs were extracted with pure methanol. The ASE extraction program was as follows: 2 cycles, 5 min heating time, 80 C extraction temperature, 160 s purge time, and 1606 psi static pressure. The extracted analytes were evaporated to dryness with a rotary evaporator and nitrogen gas at room temperature. HCAs were re-suspended in 2.5 mL methanol and filtered (Mini-UniPrepTM G2 syringeless filter, 0.2 mm, Whatman, Bucking- hamshire, UK) prior to ultrahigh performance liquid chromatography coupled to high resolution accurate mass tandem mass spectrometry (UHPLC-HRMS/MS) analysis [14].

Analysis of heterocyclic amines in grilled meats
The UHPLC-HRMS/MS analysis was performed on a LTQ Orbitrap XL mass spectrometer (Thermo Scientific, MO, USA) as previously described for HCA standards.

Method validation
Ultra-high performance liquid chromatograph-high resolution accurate mass spectrometry (UHPLC-HRAMS) with selected ion monitoring (SIM) acquisition was performed for method validation including linear regression, limit of detection (LOD), limit of quantitation (LOQ), precision, accuracy and recovery. Fig. 1A-B shows the extracted ion chromatograms and chemical structures of HCA standards evaluated, and demonstrate the suitability of UHPLC-HRMS/MS to resolve and accurately detect all 7 HCA compounds using a Luna C18 column. The high resolution mass spectrometric capability of the LTQ Orbitrap XL mass spectrometer allowed for accurate and unambiguous identification and quantitation of individual HCA resolved in the chromatograms ( Table 1). All HCAs evaluated in our method were identified based on their mass spectra and fragment ions derived from MS/MS. The two co-eluted peaks at 6.68 min. corresponding to IQ and MeIQx were resolved and quantify by extracted ion chromatography (XIC) based on the exact masses of precursor ions; [M+H] + , which for IQ and MeIQx were different: m/z (IQ) = 199.0978; m/z (MeIQx) = 214.1083. For quantitation, the area under each peak corresponding to each precursor ion exact mass was measured using Xcalibur version 4.0 (Thermo Scientific, MO, USA). Thus, both masses of precursor ion and retention/elution times were used for peak identification and quantitation of IQ and MeIQx.
Significantly, complete resolution/discrimination of co-eluates (IQ and MeIQx) based on their different exact masses was made possible due in part to the high resolving power of the LTQ Orbitrap XL mass spectrometer used in acquiring mass data in our accelerated solvent extraction-ultra high pressure-high resolution accurate mass/mas spectrometric (ASE-UHPLC-HRAMS/MS) method. The LTQ Orbitrap XL mass spectrometer (Thermo Scientific, MO, USA) used in our method has a resolving power (R) up to m/z = 100,000. Mass data for ASE-UHPLC-HRAMS/MS method for HCA analysis  described was acquired on LTQ Orbitrap XL mass spectrometer (Thermo Scientific, MO, USA) operating in positive mode at resolving power (R) m/z = 60,000. Similar applications of accurate mass high resolution mass spectrometry to resolve co-eluted chromatographic peaks based on masses of coeluates have been reported [15]. The exact mass of IQ and MeIQx calculated based on chemical structure and chemical formula using ChemDraw (PerkinElmer, Version 16.0.1.4, Massachusetts, U.S. A.) was 198.09055 and 213.10145 respectively. Furthermore, identification of IQ and MeIQx was corroborated based on the MS/MS spectra which showed distinctly different fragmentation patterns and diagnostic fragments ions which matched the literature MS values for IQ and MeIQx (see Fig. 2A-B). For IQ, the diagnostic product ion at m/ z = 184.0743 formed by loss of methyl unit (ÀCH 3 ) from precursor/parent ion at m/z = 199.0978 [1,5,16]. For MeIQx, the product ions at m/z = 199.0852 resulted from loss of one methyl unit (ÀCH 3 ) from the precursor/parent ion, and m/z = 173.0822 corresponded to elimination of one amino imidazole moiety (-C 2 NH 3 ) from the precursor/parent ion at m/z 214.1083 [1,5,16]. These diagnostic fragment ions were in good agreement with the reported literature values used to identify and quantify IQ and MeIQx [1,5,16].
Calibration curves for the 7 HCA standards showed good linearity over the range 20-100 mg/L (Table 2  this relationship should be perfectly linear. In line with this, our method showed good linearity for all 7 HCA compounds with linear correlation coefficients greater than 0.991 which means there is good correlation between peak areas and concentrations (Table 2 and Fig. 3). Limits of detection (LOD) and quantification (LOQ) were determined using: LOD 3(SD of intercept/Slope) and LOQ 10(SD of intercept/Slope) [17]. LODs ranged from 8.2 to 12 mg/L HCAs, setting a lower limit on HCA concentration in solution and were accompanied by recoveries ranging from 63.5 to 100% suggesting that the ASE extraction method for HCA compounds in accurate and efficient ( Table 2).
Application of UHPLC-HRAMS to grilled beef and moose meats showed results consistent with that reported in the literature (Fig. 4). The most abundant HCA detected was MeIQx consistent with findings in other cooked meat studies [18]. The formation of HCAs in thermally treated meats is based on many factors including cooking methods, temperature, duration of cooking, and type of meat [19].
The intra-day precision and accuracy of the method were evaluated at 3 quality control levels (40, 60 and 100 mg/L) which were above the LOD, LOQ and within linear ranges for HCAs ( Table 3). The intra-day precision was less good and showed accuracies from 45.4 to 97.1%. The detailed values of intra-day precision and accuracy are shown in Table 3. All the values are within an acceptable range. Matrix effects in grilled moose samples caused ion signal suppression for IQ, MeIQ, and Harman whereas the ion signals for MeIQx, PhIP and Nor-harman were enhanced (Table 4). In grilled beef samples, the matrix had suppressing effect on ion signals for IQ, MeIQ, PhIP and Harman whereas the ion signals for MeIQx and Nor-harman were supressed which reduced the signal strengths (Table 5)  meats including pork [16]. In the ASE-UHPLC-HRAMS/MS method described, HCA standard solutions and samples were spiked with TriMeIQx, as the internal standard for quantification, and normalized peaks areas of standards and samples based on internal standard (i.e. ratios) used to calculate concentrations of standard solutions and samples to minimize possible quantification errors produced by the matrix effects. Table 3 Intra day precision and accuracy of 6 HCAs following extraction by accelerated solvent extractors and confirmation of composition and levels by ultra high performance liquid chromatography coupled to high resolution accurate mass tandem mass spectrometry (UHPLC-HRAMS).  The grilled meat matrix effect was determined using the post-extraction method based on the formula below [20]: ME ¼ B A Ã 100 Where, A is the normalized peak area/concentration of an analyte in a standard solution. B is the normalized peak area/concentration of the analyte in grilled meat samples. All standard solutions and grilled meat samples where spiked with internal standard and their corresponding peak areas normalized based on the area ratios. If, ME $ 100 %, there is no matrix effect. If ME < 100 %, an ion-suppression occurred and, if ME > 100 %, ion-enhancement occurred.

Declaration of Competing Interest
The authors declare no conflict of interest.