Technical note: simultaneous determination of amino thiols in pig tissue by ultra-high performance liquid chromatography with fluorescence detection

Abstract

Sulfur amino acid nutrition and metabolism are linked to animal disease. While validated methods for the determination of amino thiol levels in plasma or serum are available, there is a dearth of validated methods for their measurement in tissue. A robust and reproducible ultra-high performance liquid chromatography method has been validated for the simultaneous determination of concentrations of cysteine (Cys), cysteinylglycine (CysGly), homocysteine (Hcys), γ-glutamylcysteine (γ-GluCys), and glutathione (GSH) in pig tissue. Tissue was homogenized and deproteinized with trichloroacetic acid. Amino thiols in the acid-soluble fraction of the tissue homogenate were reduced with tris-(2-carboxyethyl)-phosphine hydrochloride and derivatized with 4-(aminosulfonyl)-7-fluoro-2,1,3-benzoxadiazole (ABD-F). Amino thiols were resolved under reversed-phase gradient conditions on a Waters Acquity BEH C18 column (1.7 µm, 2.1 mm × 100 mm) within 4.5 min and detected with fluorescence. The peak area ratio of analyte to 2-mercaptopropionylglycine internal standard, added to external calibration standards and samples, was used to develop linear calibration curves. Linear calibrations were performed over the range of 15–1,500 nmol/g for Cys, CysGly, Hcys, and γ-GluCys and 150–15,000 nmol/g for GSH. Linearity, lower limit of detection, lower limit of quantitation, accuracy, precision, sample stability, and carryover were evaluated. We demonstrate excellent linearity for all analytes within their respective concentration range (r² > 0.99) and excellent recovery of amino thiols from spiked samples (mean ± SD across tissues; Cys, 100.0 ± 2.2%; CysGly, 95.4 ± 5.1%; Hcys, 96.6 ± 2.0%; γ-GluCys, 102.2 ± 2.7%; and GSH, 100.6 ± 3.3%). The intra-day and inter-day precisions did not exceed 5% and 10%, respectively. Repeated freezing and thawing of tissue homogenate did not affect measured amino thiol concentrations, ABD-labeled amino thiols were stable for 1 wk after derivatization, and there was no sample carryover across consecutive injections. We confirm the identity of each ABD-labeled amino thiol with Orbitrap mass spectrometry. Finally, we apply the method to the determination of amino thiol concentrations in liver and jejunum tissues in newly weaned pigs and show that despite elevated Cys and maintained GSH concentrations in liver, both γ-GluCys and GSH decline in jejunum of weaned pigs.

A reproducible and robust ultra-high performance liquid chromatography method was evaluated for the simultaneous determination of cysteine, cysteinylglycine, homocysteine, γ-glutamylcysteine, and glutathione in solid tissue. The method will facilitate novel insights into tissue-specific and whole-body sulfur amino acid metabolism in livestock animals.

Introduction

Non-protein, low molecular weight thiols, such as cysteine (Cys), cysteinylglycine (CysGly), homocysteine (Hcys), γ-glutamylcysteine (γ-GluCys), and glutathione (GSH), have key roles in animal health and production. Owing to the reversible formation of disulfide bonds, amino thiols, especially GSH, are crucial for antioxidant defense, xenobiotic detoxification, redox signaling, and cellular growth and apoptosis (Gallogly and Mieyal, 2007; Circu and Aw, 2012; Poole, 2015). Among amino thiols, Cys is the most abundant extracellular thiol, whereas GSH is the most abundant intracellular thiol. While Hcys, CysGly, and γ-GluCys are less abundant in biological fluids and tissues, these amino thiols are nonetheless key intermediates in sulfur amino acid metabolism. Homocysteine, derived from methionine (Met), is either remethylated, accepting a methyl group from betaine or N⁵-methyl-tetrahydrofolate to regenerate Met, or undergoes transsulfuration to irreversibly form Cys (Stipanuk, 2004). Cysteinylglycine is a product of γ-glutamyl transferase-mediated degradation of GSH and γ-GluCys is the immediate precursor of GSH. Sulfur amino acid and GSH metabolism are linked to inflammation (Malmezat et al., 1998, 2000; Rakhshandeh et al., 2019) and intestine mucosa atrophy (Bauchart-Thevret et al., 2009). Moreover, the amino thiols γ-GluCys and CysGly exhibit anti-inflammatory properties (Nosworthy and Brunton, 2016; Yang et al., 2019) and N-acetylcysteine, a Cys precursor, promotes intestinal function in pigs challenged with lipopolysaccharide (Yi et al., 2017), soybean allergen β-conglycinin (Wang et al., 2021), and porcine epidemic diarrhea virus (Wang et al., 2017) and attenuates the increase in plasma bile acids in pigs receiving intravenous nutrition (Huber et al., 2019). Thus, the comprehensive determination of tissue amino thiol status, especially in tissues with central roles in amino thiol metabolism, is needed.

Most chromatographic methods for amino thiol analysis to date have been developed and validated for biological fluids, such as serum, plasma, or urine, whereas comparatively few methods have been described for tissue (Toyo’oka, 2009). The objective of this work was to validate a rapid, robust, and reproducible method for the simultaneous determination of Cys, CysGly, Hcys, γ-GluCys, and GSH in pig tissue. Since amino thiols do not contain natural chromophores or fluorophores, derivatization is necessary. We describe pre-column derivatization of tissue amino thiols with 4-(aminosulfonyl)-7-fluoro-2,1,3-benzoxadiazole (ABD-F), a derivatization reagent specific to thiols, separation by ultra-high performance liquid chromatography (UHPLC), and detection by fluorescence (FLD). We also confirm the identity and mass spectrum of each ABD-labeled amino thiol with Orbitrap mass spectrometry (MS).

Materials and Methods

All animal procedures were approved by the Institutional Animal Care and Use Committee at Auburn University (PRN 2021-3876).

Chemicals and reagents

L-Cysteine hydrochloride monohydrate and ammonium formate were obtained from BeanTown Chemical (Hudson, NH). DL-Homocysteine was obtained from MilliporeSigma (St. Louis, MO). L-Cysteinylglycine and L-γ-glutamylcysteine were obtained from Bachem Americas (Torrance, CA). L-Glutathione (reduced) and tris-(2-carboxyethyl)-phosphine hydrochloride (TCEP) were obtained from GoldBio (St. Louis, MO). N-(2-mercaptopropionyl)glycine (2-MPG) and 4-(aminosulfonyl)-7-fluoro-2,1,3-benzoxadiazole (ABD-F) were obtained from TCI America (Portland, OR). Formic acid, sodium tetraborate decahydrate, sodium phosphate monobasic, sodium phosphate dibasic, ethylenediaminetetraacetic acid (EDTA), trichloroacetic acid (TCA), HCl, and NaOH were obtained from Thermo Fisher Scientific (Waltham, MA). Ultrapure water was obtained from a Milli-Q IQ 7000 Ultrapure Water System (MilliporeSigma). Acetonitrile (UHPLC-grade) was obtained from J.T. Baker (Phillipsburg, NJ).

Equipment and chromatographic conditions for UHPLC/FLD analysis

The UHPLC system consisted of an Acquity UPLC H-Class PLUS System equipped with a quaternary pump, fluorescence detector, and Empower 3 software (Waters, Milford, MA). Derivatized amino thiols were separated on a C18 analytical column (Waters Acquity UPLC BEH C18, 1.7 µm, 2.1 mm × 100 mm). Mobile phases were 20 mM ammonium formate in water, pH 3 (A) and 100% acetonitrile (B); mobile phase A was filtered through a 0.20-µm cellulose acetate membrane before use. The separation of amino thiols was performed at a flow rate of 600 µL/min with the following gradient: 0–0.3 min, 4% B; 0.3–1.8 min, 10% B; 1.8–2.1 min, 30% B; 2.1–2.7 min, 30% B; and 2.7–3.0 min, 4% B. Column equilibration at 4% B was carried out for 1.5 min resulting in a total run time of 4.5 min. All gradient changes were linear. The autosampler temperature was set at 8 °C and the column temperature was maintained at 50 °C. The sample injection volume was 5 µL. The excitation wavelength was set at 390 nm; the emission wavelength of 510 nm was monitored at 10 Hz.

Equipment and chromatographic conditions for UHPLC/MS analysis

Analysis was performed on a Vanquish Flex Binary UHPLC system with a UV–Vis detector (Thermo Fisher Scientific) and quadrupole Orbitrap mass spectrometer (Orbitrap Exploris 120, Thermo Fisher Scientific) operating in positive electrospray ionization mode and Xcalibur software (version 4.4.16.14; Thermo Fisher Scientific). Chromatographic conditions and gradient were identical to amino thiol analysis by UHPLC/FLD, as described above. Following the injection of derivatized amino thiols, UV absorbance was monitored at 254 nm and at 10 Hz; flow was diverted from the mass spectrometer for the initial 0.7 min to prevent sample buffer from being introduced into the instrument. The optimal collision energy was determined with chromatography using a targeted mass list based on the [M+H]⁺ and retention time windows (Kertesz et al., 2009). Multiple injections and iterations of higher and lower normalized energies were used to approach the optimal collision energy. The following MS parameters were used: scan range, 100–1000 m/z; mass resolution, 60,000; S-lens RF voltage, 70%; maximum ion injection time, automatic; EASY-IC Ion Source, on; positive spray voltage, 3,200 V; ion transfer tube temperature, 320 °C; vaporizer temperature, 285 °C.

External and internal calibration standards and stock solution preparation

Individual stock solutions of Cys, CysGly, Hcys, γ-GluCys, GSH, and 2-MPG at 2.5 mM were prepared in 0.01 M HCl. A series of seven calibration standards, containing Cys, CysGly, Hcys, and γ-GluCys from 1 to 100 µM, GSH from 10 to 1,000 µM, and 2-MPG at 125 µM, were prepared from stock solutions of each amino thiol and divided into 200-µL aliquots. At these concentrations, the calibration standard corresponded to a minimum injection of 0.8 pmol for Cys, CysGly, Hcys, and γ-GluCys and 8 pmol for GSH, and a maximum injection of 80 pmol for Cys, CysGly, Hcys, and γ-GluCys and 800 pmol for GSH, per 5-µL injection onto the analytical column; 100 pmol of 2-MPG was injected at all levels of the calibration standard. The 2-MPG internal ­standard (IS) working solution (125 µM) was prepared by diluting the 2-MPG stock solution 1:20 in 0.01 M HCl. Aliquots of the calibration standard were stored at −80 °C and were not thawed more than once.

A stock solution of TCEP (300 mM, pH 7) was prepared in water, divided in 100-µL aliquots, and stored at −20 °C. A stock solution of ABD-F (1.00 mg/mL) was prepared in borate buffer (125 mM with 5 mM EDTA, pH 9.5), divided in 1-mL aliquots in amber microcentrifuge tubes, and stored at −20 °C. Aliquots of TCEP and ABD-F were not reused after thawing once and stored for no longer than one month.

Sample and calibration standard preparation

Tissue (100 mg; liver and jejunum mucosa collected from eighteen 21- to 23-d old Yorkshire pigs; later section contains full animal details) was homogenized (rotor-stator) and deproteinized in TCA (5% w/v; 1,500 µL TCA per 100 mg tissue) on ice. Following a 15 min incubation on ice to promote protein precipitation, tissue homogenates were centrifuged at 12,000 × g for 10 min and the supernatant, containing acid-soluble amino thiols, was transferred into a clean microcentrifuge tube, briefly centrifuged, and stored at −80 °C until analysis. A portion of the supernatant was pooled per tissue and used for accuracy, intra-day precision, inter-day precision, stability, and carryover analyses.

Tissue homogenate (40 µL), 2-MPG IS (40 µL), NaOH (10 µL; 1 M; added to neutralize TCA in the tissue homogenate), and phosphate buffer (25 µL; 250 mM, pH 7.2) was added to a microcentrifuge tube, vortex-mixed, and briefly centrifuged. TCEP (10 µL; 12.5 mM; prepared from the TCEP stock solution diluted in water) was added to each tube, vortex-mixed, briefly centrifuged, and incubated at room temperature for 15 min. Following reduction of mixed disulfides to free thiols with TCEP, ABD-F (100 µL; 0.25 mg/mL; prepared from the ABD-F stock solution diluted in borate buffer) was added to each tube, vortex-mixed, briefly centrifuged, and incubated at 50 °C for 15 min. Following derivatization, HCl (25 µL; 2 M) was added to each tube and centrifuged at 12,000 × g for 10 min. A 200-µL aliquot of each reduced and derivatized sample was transferred to an amber autosampler vial containing a 250-µL polypropylene insert. The procedure for preparing calibration standards was identical to sample preparation with the exception that the tissue homogenate was replaced with an equal volume of TCA and the 2-MPG IS was replaced with calibration standard (containing 2-MPG IS). Appropriate reaction conditions at both the reduction and derivatization steps were verified by directly measuring pH in a parallel subset of samples (Symphony B10P pH meter, VWR, Radnor, PA). Working concentrations of TCEP and ABD-F were optimized by monitoring peak areas with increasing amounts of TCEP (0, 0.5, 1.0, 2.5, and 5.0 mM) and ABD-F (0.10, 0.25, 0.50, and 1.00 mg/mL) in n = 3 replicates of calibration standard containing 100 µM Cys, CysGly, Hcys, and γ-GluCys, 1,000 µM GSH, and 125 µM 2-MPG.

Method calibration and linearity

Calibration curves were constructed from seven standard concentrations of five amino thiols (1, 2.5, 5, 10, 25, 50, and 100 µM for Cys, CysGly, Hcys, and γ-GluCys; 10, 25, 50, 100, 250, 500, and 1,000 µM for GSH; 125 µM for 2-MPG IS). For each calibration curve, the peak area ratio of each amino thiol to the 2-MPG IS was calculated and plotted against the nominal amino thiol concentration. Calibration curves for each amino thiol were generated by weighted (1/x) linear regression analysis (n = 5) with the regression procedure of SAS (version 9.4; SAS Institute, Cary, NC). Limit of detection (LOD) was calculated as 3.3 × (σ/S), where σ is the standard deviation of the y-residuals of the regression line and S is the slope of the calibration curve. Limit of quantitation (LOQ) was calculated as 10 × (σ/S). The LOD and LOQ are reported as nmol/g tissue.

Method accuracy and matrix effects

The accuracy of the method was evaluated by determining the recovery of spiked amino thiols in the acid-soluble fractions of liver and jejunum homogenates. Each tissue was spiked with amino thiols equivalent to 15, 75, 150, and 750 nmol/g for Cys, CysGly, Hcys, and γ-GluCys and 150, 750, 1,500, and 7,500 for GSH; each tissue was also analyzed without spiking to determine baseline amino thiol concentrations. The spike solutions contained the 2-MPG IS working solution at 125 µM each and replaced the 2-MPG IS working solution during sample preparation. The concentration of amino thiols in non-spiked and spiked samples was determined in n = 3 replicates per level and tissue. Accuracy was calculated as:

$${\displaystyle \begin{array}{ll}& \mathrm{A}\mathrm{c}\mathrm{c}\mathrm{u}\mathrm{r}\mathrm{a}\mathrm{c}\mathrm{y}\left(\mathrm{\%}\right)=\\ & \frac{\mathrm{O}\mathrm{b}\mathrm{s}\mathrm{e}\mathrm{r}\mathrm{v}\mathrm{e}\mathrm{d}\mathrm{s}\mathrm{p}\mathrm{i}\mathrm{k}\mathrm{e}\mathrm{d}\mathrm{s}\mathrm{a}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n},\mathrm{n}\mathrm{m}\mathrm{o}\mathrm{l}/\mathrm{g}}{\mathrm{N}\mathrm{o}\mathrm{m}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{l}\mathrm{s}\mathrm{p}\mathrm{i}\mathrm{k}\mathrm{e}\mathrm{d}\mathrm{s}\mathrm{a}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n},\mathrm{n}\mathrm{m}\mathrm{o}\mathrm{l}/\mathrm{g}}\times 100\end{array}}$$

where the nominal spiked tissue sample concentration is equal to the sum of the baseline concentration in non-spiked tissue and the corresponding spike concentration. Measured concentration was reported as nmol/g tissue; mean, SD, and CV were calculated. The derivatization efficiency for each amino thiol in the liver and jejunum mucosa (i.e. matrix effect) was calculated as:

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{D}\mathrm{e}\mathrm{r}\mathrm{i}\mathrm{v}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{z}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{e}\mathrm{f}\mathrm{f}\mathrm{i}\mathrm{c}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{y}\left(\mathrm{\%}\right)=\frac{A-S}{S}\times 100}\end{array}}$$

where A is the peak area ratio of each amino thiol in spiked tissue samples and S is the peak area ratio of each amino thiol in the corresponding standard. As an additional check on derivatization efficiency, we compared the peak area of the 2-MPG IS in spiked tissue samples to the peak area of the 2-MPG IS in the corresponding standard.

Method intra-day and inter-day precision

The intra-day and inter-day reproducibility of the method was established by replicate analysis of the pooled acid-soluble fraction of liver and jejunum homogenate. Intra-day precision and accuracy were determined from n = 5 replicates of each tissue within a single day. Inter-day precision and accuracy were determined from n = 5 replicates of each tissue across five consecutive days. Measured concentration was reported as nmol/g tissue; mean, SD, and CV were calculated.

Sample stability and carryover

Stability of the acid-soluble amino thiols was determined by replicate analysis of the pooled liver homogenate after three freeze-thaw cycles from −80 °C to room temperature (n = 5). Stability of the ABD-labeled amino thiols was determined by replicate analysis of pooled liver ­homogenate (n = 5) immediately following derivatization and 1 and 7 d thereafter. All samples were derivatized at the same time and samples that were scheduled for later analysis were stored at 4 °C and protected from light. Sample carryover was determined from a sequence of 5 injections of a water blank, 10 injections of derivatized liver sample, and 5 injections of a water blank (n = 3). The peak area ratio of each amino thiol to the 2-MPG IS was evaluated in the last five injections for each replicate.

Application to weaning stress in pigs

The validated UHPLC/FLD method was applied to study the amino thiol content of the liver and small intestine of weanling pigs. Yorkshire pigs at the Swine Research and Education Center at Auburn University were either weaned 21-d age without access to feed but free access to water for 48 h (W; 6.90 ± 0.81 kg; n = 9, 4 male and 5 female) or were not weaned and remained suckling the sow for 48 h (NW; 6.81 ± 0.65 kg; n = 9, 5 male and 4 female); pigs were obtained from different sows. After 48 h, all pigs were euthanized with an intracardiac injection of saturated KCl under isoflurane anesthesia. Liver and jejunum mucosa were excised from each pig, snap-frozen in liquid nitrogen, and stored at −80°C until analysis. Tissue concentrations of all amino thiols were determined as outlined above. The generalized linear mixed model procedure of SAS 9.4 was used to compare differences in tissue amino thiol concentrations between weaned and non-weaned pigs; treatment was considered a fixed effect and pig was considered a random effect. Normality of residuals was assessed with the Shapiro–Wilk test statistic. Data are presented as least-squares means ± SEM. Differences between groups were considered significant at P < 0.05.

Results and Discussion

High performance liquid chromatography coupled with fluorometric or mass spectrometric detection are the most common methods for the simultaneous determination of multiple amino thiols in biological samples. However, most published methods have been validated in plasma, serum, saliva, or urine (Toyo’oka, 2009) and often measure only the highly abundant thiols Cys and GSH (Guan et al., 2003; Brundu et al., 2016; Kaminska et al., 2018). While mass spectrometry-based quantification of amino thiols is increasingly popular (Forgacsova et al., 2019; Onozato et al., 2020), instrument acquisition, usage, and maintenance costs can rapidly become cost prohibitive for routine analysis of amino thiols (Song et al., 2018). Moreover, investigators whose work focuses on protein and amino acid nutrition and metabolism often have easier access to HPLC systems equipped with fluorescence detectors for analyzing the fluorescent derivatives of amino acids labeled with ο-phthalaldehyde or 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate (Song et al., 2018). Thus, we anticipate that analysis of tissue amino thiols as their fluorescent ABD-F derivatives can be readily adopted by these research groups. Targeted analysis of multiple amino thiols, including the less abundant CysGly and γ-GluCys intermediates, in tissues that play key roles in sulfur amino acid metabolism will confer additional information on the impact of diet, inflammation, or stress on Cys utilization in livestock animals that plasma amino thiol concentrations alone do not provide.

Chromatographic separation

Although amine reactive reagents, such as ο-phthalaldehyde, ninhydrin, phenyl isothiocyanate, dansyl chloride, or 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate, can be used to label amino thiols, we opted for ABD-F derivatization. ABD-F is highly selective toward thiols, enabling rapid resolution of target compounds, and lacks intrinsic fluorescence (Toyo’oka, 2009). ABD-F derivatization also employs milder reaction conditions than the more commonly used ammonium 7-fluoro-2,1,3-benzoxadiazole-4-sulfonate. TCEP was selected as the preferred reducing agent because it does not contain a thiol group. Alternative reducing agents, such as dithiothreitol and β-mercaptoethanol, contain thiol groups and thus react with derivatization reagents, including ABD-F, that specifically label thiols.

Representative fluorescence chromatograms of derivatized amino thiols in the calibration standard at low and high amino thiol concentrations, pooled liver, and pooled jejunum mucosa tissue are shown in Figure 1. The peaks corresponding to ABD-labeled Cys, CysGly, Hcys, γ-GluCys, GSH, and 2-MPG was resolved with retention times of 0.88, 1.12, 1.36, 1.73, 1.98, and 2.95 min, respectively. The CV of the FLD for retention times of ABD-labeled Cys, CysGly, Hcys, γ-GluCys, GSH, and 2-MPG was 0.21%, 0.43%, 0.31%, 0.51%, 1.01%, and 0.06%, respectively. All amino thiols were resolved within 4.5 min. The short runtime is a considerable improvement over published methods that use UV–Vis or fluorescence, for which total analysis time ranged from 6 min (Ferin et al., 2012) to 20 min (Steele et al., 2012). Cysteine peak area was maximized when TCEP was included at 0.5 mM working concentration, whereas peak areas for CysGly, Hcys, γ-GluCys, GSH, and 2-MPG were maximized when TCEP was included at 1.0 mM working concentration (Supplemental Figure S1). Since the relative differences between the 0.5 mM and 1.0 mM concentrations in amino thiol peak areas were less than 2.5%, we selected the 1.0 mM concentration to maximize CysGly, Hcys, γ-GluCys responses because these amino thiols are less abundant in solid tissue than Cys and GSH. Beyond 1.0 mM, there was a significant decline in peak areas such that the relative difference between 1.0 mM and 5.0 mM concentrations were 26.6%, 13.7%, 17.5%, 14.9%, 9.1%, and 12.4% for Cys, CysGly, Hcys, γ-GluCys, GSH, and 2-MPG, respectively (P < 0.05). The working concentration of TCEP selected is equivalent to approximately 50 nmol TCEP per mg tissue. While peak areas for each amino thiol increased from 0.10 mg/mL to 0.25 mg/mL ABD-F (P < 0.05), there was no benefit of further increasing ABD-F from 0.25 mg/mL to 0.50 or 1.00 mg/mL (P > 0.10; Supplemental Figure S1).

Figure 1.

Representative fluorescence chromatogram of ABD-labeled amino thiols in the external calibration standard (low concentration; A), external calibration standard (high concentration; B), liver (C), and jejunum mucosa (D). Peak labels: 1, Cys; 2, CysGly; 3, Hcys; 4, γ-GluCys; 5, GSH; 6, 2-MPG. An inset between 0.7 min and 2.3 min is included to illustrate peaks that correspond to Cys, CysGly, Hcys, and γ-GluCys.

Linearity and limits of detection and quantitation

Linear calibration curves were obtained from 1 to 100 µM for Cys, CysGly, Hcys, and γ-GluCys and from 10 to 1,000 µM for GSH. These concentration ranges corresponded to an injection amount of 0.8 to 80 pmol per injection for Cys, CysGly, Hcys, and γ-GluCys and 8 to 800 pmol per injection for GSH (Table 1; Supplemental Figure S2). After plotting the ratio of the amino thiol peak area to the 2-MPG IS peak area against the injection amount, the regression coefficients for each calibration curve exceeded 0.99 and the CV for the calibration curve slope corresponding to each amino thiol did not exceed 2%. The LOD and LOQ for each amino thiol were less than 1 nmol/g, apart from the LOQ for GSH at 2.41 nmol/g, after correction to the equivalent tissue concentration. When converted to the equivalent amount per injection on the column, this corresponds to an LOD of 13.3, 11.2, 9.1, 15.5, and 42.3 fmol and an LOQ of 40.3, 34.0, 27.6, 47.0, and 128.3 fmol for Cys, CysGly, Hcys, γ-GluCys, and GSH, respectively. These limits are consistent with those reported in previous studies that employ alternative labeling reagents, chromatographic conditions, and detection methods (Cevasco et al., 2010; Zhang et al., 2014; Isokawa et al., 2016; Sun et al., 2016; Kaminska et al., 2018).

Table 1.

Linear regression data for determination of tissue Cys, CysGly, Hcys, γ-GluCys, and GSH concentrations by UHPLC/FLD by the external calibration standard method

	Cys	CysGly	Hcys	γ-GluCys	GSH
Linear range, nmol/g	15–1,500	15–1,500	15–1,500	15–1,500	150–15,000
Slope1	0.00012	0.00077	0.00032	0.00015	0.00033
Slope CV, %	1.69	1.37	0.67	1.66	0.29
Slope r2	0.99	0.99	0.99	0.99	0.99
Intercept	0.00034a	0.00068b	0.00042a	0.00030a	0.02423a
LOD,2 nmol/g	0.25	0.21	0.17	0.29	0.79
LOQ,3 nmol/g	0.76	0.64	0.52	0.88	2.41

¹ Calibration curve parameters calculated from n = 5 replicates of the external calibration standard.

² Limit of detection (LOD) = 3.3 × (σ/S); σ is the standard deviation of the y-residuals of the regression line; S is the slope of the calibration curve.

³ Limit of quantitation (LOQ) = 10 × (σ/S).

^a Intercept differs from 0 (P < 0.05).

^b Intercept does not differ from 0 (P > 0.10).

Accuracy and derivatization efficiency

Known amounts of each amino thiol were added to the homogenate of pooled liver and jejunum tissue to produce samples spiked from 15 to 750 nmol/g tissue equivalent for Cys, CysGly, Hcys, and γ-GluCys and spiked from 150 to 7,500 nmol/g tissue equivalent for GSH (Table 2). Triplicate samples were analyzed per level and the percentage of each amino thiol recovered was calculated using the linear regression equations in Table 1. Averaging across tissues, accuracy ranged from 95.6% to 102.3% (average, 100.0%) for Cys, 89.0% to 101.6% (average, 95.4%) for CysGly, 93.2% to 99.5% (average, 96.6%) for Hcys, 98.6% to 107.8% (average, 102.1%) for γ-GluCys, and 95.5% to 105.8% (average, 100.4%) for GSH. Amino thiol accuracy was within ± 10% of the nominal value apart from CysGly in liver spiked with 75 nmol/g tissue equivalent. The derivatization efficiency of amino thiols across liver and jejunum mucosa averaged 97.7% (Table 2). In liver, however, the derivatization efficiency of CysGly and GSH were both 88.3%, whereas the derivatization efficiency of γ-GluCys was 111.9%, suggesting a marginal effect of liver matrix on the accuracy of the method. The relative difference between peak areas for the 2-MPG IS in spiked tissue samples and corresponding calibration standards were within 10%.

Table 2.

Recovery and derivatization efficiency of Cys, CysGly, Hcys, γ-GluCys, and GSH in spiked pig liver and jejunum mucosa (% ± SD) by UHPLC/FLD

Item	Cys	CysGly	Hcys	γ-GluCys	GSH
Liver1, 2
15	99.8 ± 3.7	94.9 ± 2.9	93.2 ± 1.1	101.6 ± 11.0	99.7 ± 3.5
75	99.9 ± 1.4	90.7 ± 0.9	95.1 ± 1.7	98.6 ± 6.4	100.6 ± 1.1
150	99.5 ± 1.8	89.8 ± 1.2	96.3 ± 0.2	103.0 ± 1.8	100.6 ± 1.0
750	95.6 ± 5.0	89.0 ± 0.5	98.4 ± 0.5	103.0 ± 2.4	95.5 ± 0.7
Derivatization efficiency	94.5 ± 2.0	88.3 ± 1.2	100.4 ± 1.3	111.9 ± 8.8	88.3 ± 3.2
Jejunum mucosa1, 3
15	101.9 ± 1.0	101.6 ± 0.7	97.6 ± 1.0	100.4 ± 1.3	101.9 ± 0.7
75	99.2 ± 5.2	99.4 ± 0.1	97.0 ± 0.5	102.3 ± 5.0	103.6 ± 0.8
150	102.3 ± 6.0	101.4 ± 0.6	99.5 ± 0.4	107.8 ± 0.4	105.8 ± 0.4
750	102.1 ± 0.2	96.0 ± 0.1	95.6 ± 0.2	101.0 ± 0.3	97.0 ± 0.2
Derivatization efficiency	97.9 ± 4.1	97.3 ± 0.6	98.4 ± 2.0	107.0 ± 4.5	92.5 ± 1.0

¹ Liver and jejunum tissue homogenate was spiked with amino thiols equivalent to 15, 75, 150, and 750 nmol/g for Cys, CysGly, Hcys, and γ-GluCys and 150, 750, 1500, and 7500 nmol/g for GSH; n = 3 replicates per level.

² Baseline liver amino thiol concentrations: Cys, 361 nmol/g; CysGly, 30.1 nmol/g; Hcys, 6.8 nmol/g; γ-GluCys, 32.5 nmol/g; GSH, 2856 nmol/g.

³ Baseline jejunum amino thiol concentrations: Cys, 117 nmol/g; CysGly, 9.4 nmol/g; Hcys, 5.0 nmol/g; γ-GluCys, 10.2 nmol/g; GSH, 1206 nmol/g.

Precision

The intra-day precision for each amino thiol was less than 5% in pooled liver and jejunum mucosa samples (Supplemental Table S1). The inter-day precision for each amino thiol was less than 10% in pooled liver and jejunum mucosa samples. Among amino thiols, γ-GluCys exhibited the highest CV for intra-day precision (2.77%), whereas CysGly exhibited the highest CV for inter-day precision (8.42%), averaged between liver and jejunum mucosa. However, all measured precisions have a CV less than 10% and represent a considerable improvement over the precision of Cys and GSH analysis in mouse organs (Brundu et al., 2016). Although the concentration difference between the 2-MPG IS and select amino thiols at low concentrations, including Hcys and γ-GluCys, is large, such that peak area ratios are less than 0.01, the measured intra-day precision for these amino thiols is acceptable at low concentrations. The CVs for Hcys and γ-GluCys in liver and jejunum mucosa, in which these concentrations are 10 nmol/g or less and correspond to a peak area ratio less than 0.01, does not exceed 5%. Moreover, the concentration of the 2-MPG IS added during sample preparation can be reduced if these amino thiols are of high importance.

Sample stability and carryover

Repeated freezing and thawing (FT) of the acid-soluble fraction of liver homogenate before derivatization did not affect the measured concentration of Cys, CysGly, Hcys, or γ-GluCys compared to a single FT (Figure 2A; P > 0.10). Although repeated FT decreased the measured concentration of liver GSH from 3,248 to 3,229 nmol/g (P = 0.006), the difference was negligible and only detectable due to the high precision of the method. Time post-derivatization also did not affect the stability of ABD-labeled amino thiols. The measured concentration of Cys, CysGly, Hcys, γ-GluCys, and GSH was not different among analyses performed immediately, 1 d, and 1 wk after derivatization (Figure 2B; P > 0.10). There was no measured carryover of ABD-labeled amino thiols from the liver to blank samples, indicating negligible contamination of consecutive sample injections (Supplemental Figure S3).

Figure 2.

Stability of liver amino thiols following repeated freezing and thawing (A) and stability of ABD-labeled liver amino thiols immediately after, 1 d, and 1 wk after derivatization (B). Data were analyzed by two-factor ANOVA. FT × 1, one freeze-thaw cycle; FT × 3, three freeze-thaw cycles. Values are least-squares means ± SEM; error bars are smaller than the corresponding symbols.

Validation of ABD-labeled amino thiols by Orbitrap mass spectrometry

The external calibration standard was analyzed by Orbitrap mass spectrometry operating in positive electrospray ionization mode. Signal for each amino thiol-ABD derivative was detected at all concentration levels of the standard ranging from 1 to 100 µM for Cys, CysGly, Hcys, and γ-GluCys and from 10 to 1,000 µM for GSH. To our knowledge, this is the first report to characterize the fragmentation mass spectra of ABD-labeled Cys, CysGly, Hcys, γ-GluCys, GSH, and 2-MPG. The exact masses of Cys-ABD, CysGly-ABD, Hcys-ABD, γ-GluCys-ABD, GSH-ABD, and 2-MPG-ABD were 319.0, 376.0, 333.0, 448.1, 505.1, and 361.0, respectively, in agreement with their theoretical exact masses (Table 3). The collision energy for each amino thiol-ABD derivative was optimized to maximize the yield of the intact and unfragmented ion. Optimal collision energies ranged from 17 eV to 20 eV. The fragmentation pattern for each ABD-labeled amino thiol at the selected collision energies is reported in Supplemental Figure S4–S9.

Table 3.

Orbitrap UHPLC/MS characteristics of the ABD-labeled Cys, CysGly, Hcys, γ-GluCys, and GSH

ABD-F derivative	Retention time, min	Collision energy, eV	Formula	Theoretical [M+H]+	Error, ppm	Mass fragments, m/z	Intensity, %	MS fragmentation supplemental figure
Cys	1.3	17	C9H10N4O5S2	319.01564	−0.31	302.0 319.0 273.0 320.0	38.6 38.4 14.0 1.8	S4
CysGly	1.8	17	C11H14N5O6S2	376.03800	−0.57	376.0 273.0 241.0 359.0	30.6 27.2 10.3 9.7	S5
Hcys	2.0	20	C10H12N4O5S2	333.03219	−0.43	333.0 56.0 287.0 316.0	45.8 16.2 10.4 5.4	S6
γ-GluCys	2.6	20	C14H17N5O8S2	448.05913	-0.03	319.0 448.1 302.0 431.0	30.5 28.8 8.4 7.7	S7
GSH	2.8	17	C16H20N6O9S2	505.08059	−1.32	505.1 376.0 273.0 430.0	33.5 30.9 9.3 3.4	S8
2-MPG	3.4	17	C11H12N4O6S2	361.02710	0.38	258.0 241.0 361.0 315.0	21.3 19.5 14.4 7.6	S9

The chromatograms in Figure 1 contain an unknown peak with retention time of approximately 1.23 min. This peak is present when GSH is derivatized alone, whereas the peak is absent when Cys, CysGly, Hcys, and γ-GluCys are analyzed alone. Although the identity of the compound could not be confirmed by MS, it is possibly a side product of the disulfide bond reduction reaction with GSH and TCEP. With increasing concentrations of TCEP (up to 5 mM; data not shown), the peak area ratio of the unknown compound to the 2-MPG IS increases whereas the peak area ratio of the GSH-ABD derivative to the 2-MPG IS decreases.

Application to weaning stress in pigs

We are currently using the method presented here to delineate interactions among low feed intake, sulfur amino acid metabolism, and gut health in newly weaned pigs. Considering that GSH in the gut is a major antioxidant and plays a central role in redox signaling, an adequate supply of sulfur amino acids to the gut may promote intestinal function and mitigate weaning stress in pigs (Martensson et al., 1990; Aw, 2003; Circu and Aw, 2012).

Liver concentration of Cys was 40% higher in weaned compared to non-weaned pigs (P = 0.021), whereas γ-GluCys and GSH concentrations were not different (P > 0.10; Figure 3). Jejunum concentrations of Cys were not different between non-weaned and weaned pigs, whereas γ-GluCys was 33% lower (P = 0.001) and GSH was 20% lower (P = 0.003) in weaned compared to non-weaned pigs. Liver and jejunum concentrations of CysGly and Hcys did not differ between groups (P > 0.10). Jejunum γ-GluCys and GSH levels were lower in weaned pigs despite higher liver Cys and similar jejunum Cys concentrations in weaned compared to non-weaned pigs. Although no feed intake is atypical of commercial conditions, there is still considerable latency between weaning and time to first feed intake, and total feed intake is low in the first 48 h after weaning, in pigs (Bruininx et al., 2001). The weaning-induced depletion of gut GSH may contribute to increased oxidative stress, perturbed intestinal epithelial cell proliferation and differentiation, and intestinal dysfunction in these pigs. Correction of γ-GluCys and GSH levels, independent of feed intake, may be a potential approach to mitigate weaning stress in pigs.

Figure 3.

Liver Cys (A), liver γ-GluCys (B), liver total GSH (C), jejunum Cys (D), jejunum γ-GluCys (E), and jejunum total GSH (F) concentrations in non-weaned (NW; n = 9) and weaned (W; n = 9) pigs. Data were analyzed by one-factor ANOVA. Values are least-squares means ± SEM. *P < 0.05; **P < 0.01; ns, not significant.

Conclusion

In summary, we have described a robust and reproducible UHPLC/FLD method for the simultaneous determination of Cys, CysGly, Hcys, γ-GluCys, and GSH concentrations in solid tissue. While the method employs tissue ­homogenization and pre-column reduction and derivatization of acid-soluble amino thiols, the total time needed for sample preparation does not exceed 2 h. The method demonstrates excellent sensitivity, accuracy, and precision. Moreover, ABD-labeled amino thiols are rapidly resolved and stable for up to one week, enabling high sample throughput. Although the method does not discriminate between reduced glutathione and oxidized glutathione disulfide, the method offers relevant information on tissue levels of substrates for GSH synthesis and products of GSH breakdown. The detection of ABD-labeled amino thiols by MS supports future application in the comprehensive determination of tissue GSH kinetics in vivo with the use of stable isotopically labeled amino acid tracers (Rasch et al., 2020).

Glossary

Abbreviations

2-MPG

2-mercaptopropionylglycine

γ-GluCys

γ-Glutamylcysteine

ABD-F

4-(aminosulfonyl)-7-fluoro-2,1,3-benzoxadiazole

Cys

cysteine

CysGly

cysteinylglycine

EDTA

ethylenediaminetetraacetic acid

FLD

fluorescence detection

GSH

glutathione

Hcys

homocysteine

IS

internal standard

LOD

limit of detection

LOQ

limit of quantitation

Met

methionine

MS

mass spectrometry

TCA

trichloroacetic acid

TCEP

tris-(2-carboxyethyl)-phosphine

UHPLC

ultra-high performance liquid chromatography

Conflict of Interest Statement

The authors declare no real or perceived conflicts of interest.