Selenosugar determination in porcine liver using multidimensional HPLC with atomic and molecular mass spectrometry

Abstract

A methodology based on liquid chromatography coupled online with atomic and molecular mass spectrometry was developed for identifying trace amounts of the selenosugar methyl 2-acetamido-2-deoxy-1-seleno-β-D-galactopyranoside (SeGalNAc) in porcine liver, obtained from an animal that had not received selenium supplementation. Sample preparation was especially critical for the identification of SeGalNAc by molecular mass spectrometry. This involved liver extraction using a Tris buffer, followed by sequential centrifugations. The resulting cytosolic fraction was pre-concentrated and the low molecular weight selenium (LMWSe) fraction obtained from a size exclusion column was collected, concentrated, and subsequently analyzed using a tandem dual-column HPLC–ICP-MS system which consisted of strong cation exchange (SCX) and reversed phase (RP) columns coupled in tandem. Hepatocytosolic SeGalNAc was tentatively identified by retention time matching and spiking. Its identity was further confirmed by using the same type of chromatography on-line with atmospheric pressure chemical ionization tandem mass spectrometry operated in the selected reaction monitoring (SRM) mode. Four SRM transitions, characteristic of SeGalNAc, were monitored and their intensity ratios determined in order to confirm SeGalNAc identification. Instrument limits of detection for SeGalNAc by SCX-RP HPLC–ICP-MS and SCX-RP HPLC–APCI-MS/MS were 3.4 and 2.9 μg Se L⁻¹, respectively. Selenium mass balance analysis revealed that trace amounts of SeGalNAc, 2.16 ± 0.94 μg Se kg⁻¹ liver (wet weight) were present in the liver cytosol, corresponding to 0.4% of the total Se content in the porcine liver.

---

Introduction

Selenium (Se) plays a dual role as both an essential and toxic element in a variety of processes relevant to human health. As a result Se metabolism in mammalian systems has attracted considerable interest; however, full knowledge of its biochemical behaviour is far from complete. The majority of studies on Se metabolism generally involve its speciation in urine samples taken from rats receiving high doses of selenite, or from humans exposed to elevated Se intake via dietary supplementation. From these studies the selenosugar methyl 2-acetamido-2-deoxy-1-seleno-β-D-galactopyranoside (SeGalNAc; Fig. 1), has been identified conclusively as the key urinary metabolite in rats fed on low to toxic levels of selenite,^1,2 as well as the main human urinary metabolite produced following selenized yeast supplementation.^3,4 The liver is a vital multifunctional organ in excretory, synthetic and metabolic processes and is considered an indicator of the body burden of trace elements. Current literature, however, despite the importance of this highly complex organ, only contains sparse data on Se species in the liver. Our present knowledge of Se metabolites in the body links them to processes occurring in the liver. Such processes fall into three categories: Se reduction, selenoprotein synthesis and Se methylation.⁵ For example, selenite administered to rats (0.3 mg Se kg⁻¹ body weight (b.w.)) is thought to be transformed in red blood cells, transported to and taken up by the liver for selenoprotein synthesis, or excreted via the urine following methylation. In contrast, selenate (0.3 mg Se kg⁻¹ b.w.) is taken up directly by the rat liver, reduced, and then utilized for the synthesis of selenoproteins, or excreted after methylation.^6,7 The liquid chromatographic retention time of a selenium species, detected in rat urine, matched that of a methylated metabolite (later identified as SeGalNAc)² detected in rat liver 1 h after the rat had received an intravenous injection of selenate at a dose from 10 to 200 μg Se kg⁻¹ b.w.;⁷ yet it disappeared 24 hours after the injection.^1,7 So far this Se species has never been detected in the liver of control rats in the absence of selenium supplementation.⁷ Lack of detection may be largely due to low concentrations of SeGalNAc in the liver. In addition, the complexity of the liver matrix causes interferences and/or suppression effects. Other difficulties arise as a result of inherent biological variations, and even possible Se metabolic pattern variations of individuals. It is thus evident that Se speciation analysis is a challenge in natural biological samples in which supplementation has not taken place.

Fig. 1 Methyl 2-acetamido-2-deoxy-1-seleno-β-D-galactopyranoside (SeGalNAc).

In order to identify and characterize Se metabolites in biological samples, analytical techniques with high sensitivity, selectivity and low detection limits are required. High performance liquid chromatographic (HPLC) separation hyphenated to inductively coupled plasma mass spectrometry (ICP-MS) has been applied successfully for a wide variety of elemental speciation analyses.⁸ The techique’s capacity for element specific and sensitive multi-isotope detection provides an advantage for the detection of Se-containing species. The limitation in using HPLC–ICP-MS for the identification of Se-containing metabolites is that it does not provide structural information relevant to the detected species, and thus identification of the detected metabolites is solely based on matching their retention times with those of well characterized Se standards. Unfortunately, such standards are rarely available in the majority of metabolite discovery studies. This drawback can be potentially eliminated by applying a combination of HPLC–ICP-MS with electrospray (ES) and/or atmospheric pressure chemical ionisation (APCI) tandem mass spectrometry (MS/MS). Combined use of these techniques can be a powerful tool for the identification and characterisation of metabolites in body fluids and solid tissue samples.

The objective of the present study was to detect and characterise naturally occurring Se species in porcine liver. More specifically the emphasis was to investigate whether SeGalNAc, the major urinary selenium metabolite, was present in porcine liver containing natural levels of selenium. For this purpose, a methodology suitable for the separation of SeGalNAc from the complex liver matrix and subsequent detection by mass spectrometry was developed. Sample preparation/purification was a required step in order to obtain meaningful results from the HPLC–APCI-MS/MS analyses. Molecular mass spectrometric detection of SeGalNAc in complex biological mixtures, such as liver extracts, appears to require more effective separations, especially for removing the analyte from the majority of the liver matrix. For this reason selenium species were separated and enriched into cellular fractions by means of centrifugation and lyophilisation. The liver cytosol fraction was chromatographed by using size exclusion chromatography (SEC). The low molecular weight Se (LMWSe) fraction obtained from SEC was collected, further concentrated and analysed using HPLC–ICP-MS and HPLC–APCI-MS/MS. An HPLC system involving two columns connected in tandem,⁹ each with a different mode of separation, was developed and used for this purpose. More specifically a strong cation exchange (SCX) column was connected in tandem with a reversed-phase (RP) column. This configuration was used to improve the separation of the target Se analyte from other Se species and also from the liver matrix. This was not sufficiently achieved using a single column approach.

In addition, the total Se content in porcine liver and in its cellular fractions was determined in order to evaluate Se mass balance and the concentration of SeGalNAc.

Experimental

Chemicals

Tris(hydroxymethyl) aminomethane hydrochloride (TRIS) (Assay ≥ 99%) was purchased from Fluka, BioChemika. Ammonium acetate (Puriss. p. a. Reag. ACS ≥ 98%) was purchased from Riedel-deHaën, Sigma-Aldrich Laborchemikalien GmBH. Methanol (G CHROMASOLV for gradient elution, ACS) was obtained from Sigma-Aldrich. SeGalNAc standard which had been prepared synthetically¹⁰ was donated by Professor K. A. Francesconi (Institute of Chemistry-Analytical Chemistry, Karl-Franzens University Graz, Austria). The certified reference material for trace metals DORM-2 (dogfish muscle) was obtained from National Research Council, Canada, and the standard reference material® 1566b oyster tissue from National Institute of Standards & Technology, USA. Water (18 MΩ cm) was obtained from an Ultra Clear Basic water system (SG GmbH, Germany).

Sample

Fresh porcine liver was purchased, a few hours after the animal had been slaughtered, from a local market store. The whole lobe of porcine liver (about 250–300 g) was cleaned with cold deionised water, homogenised in a blender for 15 s in a cold room and an aliquot was taken for extraction. The animal’s diet was one normally fed to swine, consisting of maize crops, legume crops, grains, etc., without any specific Se supplementation. The Se level in these ingredients was <0.1–0.3 mg Se kg⁻¹ (dry weight) according to our data on the total Se content in foodstuffs.

Preparation of hepatocytosolic fraction from liver sample

The extraction of the liver homogenate was performed using an extractionbuffer consisting of 30 mM TRIS–HCl, pH 7.4 at a w/v (g mL⁻¹) ratio of 1 to 4. The extractions were carried out in a Potter-Elvehjem homogenizer (Jencons, UK) in an ice bath. The resulting suspension was centrifuged (SIGMA 3K20, Germany) at 4 °C and 19 621 × g for 30 min to sediment the nuclei, mitochondria and other coarser nonsuspended material. The supernatant was retained for a second centrifugation (Sorvall Ultra 80, USA) at 4 °C and 150 000 × g for 1.5 h to generate a pellet enriched in microsomes and a supernatant representing the final hepatocytosolic fraction.

Preparation of the selenosugar-containing hepatocytosolic fraction

The hepatocytosolic fraction (≈20 mL) was frozen, lyophilised and then re-suspended in 2.0 mL deionised water. An aliquot (200 μL) of this concentrated fraction was applied to a size exclusion column (SEC; 300 mm × 8 mm i.d. with a guard column 50 mm × 8 mm i.d., HEMA Bio-linear, Germany) and eluted using 5 mM ammonium acetate, pH 7.5 at flow rate of 1.0 mL min⁻¹. The SEC column was calibrated using the following standards: caffeine (M_r 194.20), myoglobin (16 kDa), bovine albumin (66 kDa), and apo-transferrin (76 kDa). A calibration curve of log(M_r) against retention time was plotted and used to assist with the collection of high- and low-molecular weight Se-containing biomolecules. Prior to offline fraction collection, the SEC was connected on-line to ICP-MS (Thermo Electron Elemental Analysis, Winsford, UK) and used to obtain the chromatographic profile of selenium-containing compounds by monitoring signals for ⁷⁷Se⁺, ⁷⁸Se⁺ and ⁸²Se⁺. Subsequently fractions were collected off-line at 100 s intervals, from 350 s to 800 s. The retention time of 650 s corresponded to a M_r cut-off of ∼1 kDa; prior to this retention time a series of high molecular weight Se-containing molecules (HMWSe) were detected (Fig. 2). Column calibration revealed that fractions collected after 650 s (10.8 min) contained low molecular weight selenium (LMWSe) compounds (<1 kDa), which potentially included SeGalNAc. A SeGalNAc standard injected onto the SEC column eluted at 680 s. LMWSe fractions collected from eight chromatographic runs were pooled and lyophilised. The resulting residue was resuspended in 200–250 μL deionised water before being further analysed using tandem dual-column HPLC with atomic and molecular mass spectrometry.

Fig. 2 Size exclusion chromatographic profile of Se species in porcine liver hepatocytosol. Obtained using ICP-MS detection on-line with a HEMA Bio-linear 300 mm × 8 mm i.d. column; mobile phase: 5 mM ammonium acetate, pH 7.5; flow rate: 1.0 mL min⁻¹; injection loop: 50 μL.

To ensure sample stability all pretreatment procedures, including sample homogenation, extraction, centrifugation, and fraction collection, were performed at 0–4 °C and sample preconcentration was carried out using lyophilisation under −60 °C (Christ Alpha 1–4 LSC, Martin Christ Gefriertrocknungsanla-gen GmbH, Germany). Ice baths were used to store samples between the various pretreatment steps, once they were taken out of storage at −80 °C. Only the HPLC separations were conducted under ambient temperature conditions, however, such conditions under the timescales applied have never before been reported to have any effect on selenosugar stability.

Selenium mass balance

In order to investigate the extraction efficiency and to estimate the level of the detected hepatocytosolic selenosugar, a mass balance for Se was carried out. Portions (0.1–0.3 g) of the raw porcine liver, liver extract, pellets and supernatant from sequential centrifugations were digested with nitric acid in a SpeedwaveTM MWS-2 microwave digestion system (Berghof Products + Instruments GmbH, Germany). The cooled digests were diluted with water to 10.0 or 15.0 mL, and the Se concentration measured by using ICP-MS (Thermo Electron Elemental Analysis, Winsford, UK). Selenium in the CRM DORM-2 (dogfish muscle) and NIST 1566b oyster tissue was determined for quality assurance purposes: certified Se concentration for DORM-2: 1.40 ± 0.09 mg kg⁻¹, determined: 1.46 ± 0.04 mg kg⁻¹, n = 2; certified Se concentration for NIST 1566b: 2.06 ± 0.15 mg kg⁻¹, determined: 2.15 ± 0.03 mg kg⁻¹, n = 5). The quantification of total Se was based on external calibration using a 6-point calibration curve of Se (0, 1, 5, 10, 25, 50 μg Se L⁻¹) and indium (5 μg In L⁻¹) as the internal standard. All samples were prepared and analysed in duplicate.

Instrumentation

Determination of total selenium was performed using an X-Series ICP-MS (Thermo Electron Elemental Analysis, Winsford, UK) equipped with a concentric nebuliser and an impact bead spray chamber. The typical operation parameters are listed in Table 1.

Table 1 Optimised chromatographic (SCX-RP HPLC) and detector conditions (ICP-MS and APCI-MS/MS)

HPLC
Strong cation exchange (SCX)	Zorbax SCX 300 (150 mm × 4.6 mm i.d., 5 μm particle size)
Reversed phase (RP)	Thermo Scientific BDS Hypersil C18 (150 mm × 4.6 mm i.d., 5 μm)
Mobile phase	5% methanol (MeOH) in 20 mM ammonium acetate; pH adjusted to 4.5.
Injection volume	50 μL
Flow rate	0.5 mL min^−1
ICP-MS
Plasma power	1380 W
Cooling gas	13.5 L min^−1
Auxiliary gas	0.90 L min^−1
Nebuliser gas	1.04 L min^−1
Nebuliser type	Typhoon
Spray chamber	Impact bead
Quadrupole dwell time	200 ms for all isotopes
APCI
Discharge current	4.0 μA
Vaporiser temperature	400 °C
Capillary temperature	300 °C
Collision cell pressure	1.0 mTorr

---

Tandem dual-column HPLC–ICP-MS analysis

The isocratic HPLC system hyphenated to ICP-MS included a Marathon HPLC pump (Rigas Labs, Thessaloniki, Greece) and a 6-port injector (Rheodyne, USA) with a 50-μL loop. The chromatographic separation was performed using a combination of two HPLC analytical columns, i.e., a SCX and a RP column connected in tandem in this order. For compatibility with APCI-MS/MS analysis, a relatively low buffer concentration (20 mM ammonium acetate, Table 1) was used. The outlet of the HPLC column was connected directly to the nebuliser inlet of an X-Series ICP-MS (Thermo Electron Elemental Analysis, Winsford, UK) via a piece of PTFE tubing. The chromatographic profiles for three Se isotopes, ⁷⁷Se, ⁷⁸Se and ⁸²Se, were acquired in the pulse counting mode. Chromatographic conditions used are summarised in Table 1.

The SeGalNAc recovery from the HPLC system was determined using the following approach. First, a standard SeGalNAc solution was injected onto the HPLC system and the chromatographic profile (I) was recorded. Secondly, the injection of the same SeGalNAc solution was repeated but without the HPLC columns, i.e., flow injection mode, and the resulting profile (II) was recorded. The peak area of each profile was used to calculate the percentage chromatographic recovery for SeGalNAc according to the equation: [(Peak area of I)/(Peak area of II)] × 100%.

The limit of detection (LOD) for SeGalNAc using SCX-RP HPLC–ICP-MS was determined by external calibration. A four-point calibration curve was constructed based on the integrated peak areas from the HPLC–ICP-MS profiles of standard solutions prepared to contain 0, 4.35, 8.70 and 43.5 μg Se L⁻¹ as SeGalNAc.

Tandem dual-column HPLC–APCI-MS/MS analysis

An enhanced resolution triple quadrupole mass spectrometer TSQ Quantum (Thermo Scientific, San Jose, CA) was operated with an atmospheric pressure chemical ionisation (APCI) source in the positive ion mode, using the optimised tune parameters shown in Table 1, in order to determine SeGalNAc. The online HPLC hyphenation comprised of a Thermo Finnigan Surveyor® MS Pump, and a manual injector fitted with a 50-μL loop. The column setup and the applied chromatographic conditions were the same as used for the SCX-RP HPLC–ICP-MS analysis (Table 1). The HPLC eluent was introduced directly into the APCI source at 0.5 mL min⁻¹ without any post-column split.

Results and discussion

Optimisation of HPLC–ICP-MS for the identification of hepacytosolic SeGalNAc

Tentative identification of SeGalNAc in the hepatocytoslic fraction was achieved by matching its retention time with that obtained for the SeGalNAc standard using a tandem dual-column HPLC–ICP-MS system, and by spiking the hepatocytosolic LMWSe fraction with the SeGalNAc standard and observing their co-elution (Fig. 3). Optimisation of the HPLC conditions mainly involved varying the methanol (MeOH) content of the mobile phase. This experimental parameter was selected because it was found to have a pronounced effect on chromatographic performance. When the MeOH content was decreased from 10% to 5%, separation of early eluting peaks was improved, also the peak tentatively identified to be SeGalNAc was retained longer (t_R = 10.3 min versus t_R = 8.6 min), thus providing improved separation from the non-retained matrix components. The latter is critical for the successful use of molecular mass spectrometry, i.e.HPLC–APCI-MS/MS, for further identification and characterisation of SeGalNAc in the hepatocytosolic liver fraction. The identification of the earlier eluting selenium peaks, other than the one identified to be SeGalNAc, is beyond the scope of this paper.

Fig. 3 SCX-RP HPLC–ICP-MS chromatogram for (a) a SeGalNAC standard at concentration of 43.5 μg Se L⁻¹; and (b) hepatocytosolic LMWSe fraction and it’s spiked with 10 μg Se L⁻¹ as SeGalNAC. The chromatographic conditions used are given in Table 1.

The LOD for SeGalNAc by SCX-RP HPLC–ICP-MS was determined to be 3.4 μg Se L⁻¹, calculated from the linear regression of the external calibration curve. The chromatographic recovery of the SeGalNAc standard under the applied HPLC conditions was found to be 71.0% ± 0.1% (n = 2) at levels <10 μg Se L⁻¹ and 84.3% at levels of about 40.0 μg Se L⁻¹ as SeGalNAc.

Identification of hepatocytosolic SeGalNAc by using HPLC–APCI-MS/MS

Because ICP-MS does not provide direct structural information suitable for the conclusive identification of SeGalNAc, a molecular tandem MS technique was also used. APCI was used instead of electrospray (ES) because it has been shown to offer improved robustness and sensitivity for SeGalNAc detection in crude human urine.¹¹ In addition, the low MeOH content of the mobile phase does not seem to affect APCI. HPLC–APCI-MS/MS was performed using the optimised tune conditions shown in Table 1. Selective and sensitive monitoring for SeGalNAc was carried out using selected reaction monitoring (SRM), and four SRM transitions (Fig. 4) were evaluated as described elsewhere.¹¹

Fig. 4 SCX-RP HPLC–APCI-MS/MS chromatograms from a SeGalNAc standard, at a concentration of 25 μg Se L⁻¹, in which four SRM transitions were monitored. The chromatographic conditions used are given in Table 1, except that the MeOH content was 10%.

Initially, when analysing the hepatocytosolic liver extracts by HPLC–APCI-MS/MS, matrix effects caused poor chromatographic efficiency and severe sensitivity suppression which did not allow for the detection of SeGalNAc.

In order to avoid these effects it is imperative that SeGalNAc is separated to some extent from the non-retained matrix components. Two experimental approaches were adopted in order to overcome matrix effects. First, an organelle separation protocol (as described in the sample preparation section) was applied to the liver extract by using sequential centrifugation. This method is simple yet effective, allowing for the fractionation of tissue homogenate into its subcellular fractions, i.e., membranes, nuclei, mitochondrial, microsome and cytosol, thus reducing sample complexity. However, even after treatment the resulting HPLC–APCI-MS/MS chromatograms did not reveal the presence of SeGalNAc peaks. In an attempt to further reduce the matrix interferences from the treated samples, the HPLC methanol content was decreased in order to increase the retention of hepatocytosolic SeGalNAc, thus enhancing its separation from other matrix species. Even though this did not completely remove the matrix interferences (both spectrometric and non-spectrometric), it provided sufficient signal-to-noise ratio, thus allowing for the detection of SeGalNAc in the LMWSe fraction (Fig. 5).

Fig. 5 SCX-RP HPLC–APCI-MS/MS chromatograms of LMWSe fraction from the liver sample under optimised HPLC conditions using 5% MeOH, following sample centrifugations for the sequential removal of organelles. Four SRM transitions for SeGalNAc were monitored.

SRM transition intensity ratios were used to verify the identity of SeGalNAc in the LMWSe fraction. First, the ratios of the four SRM transitions for SeGalNAc standard solutions were determined (Table 2). Relative standard deviations (RSDs) of less than 4% (n = 12, measured over 6 experimental days) were obtained for all of them, for concentrations ranging from 10–100 μg Se L⁻¹. These results demonstrated the good repeatability and reproducibility of product ion formation under the applied HPLC–APCI-MS/MS conditions. Subsequently, SRM transition intensity ratios were determined for the SeGalNAc in hepatocytosol (Table 2). The calculated transition intensity ratios, i.e., corresponding to transitions of m/z 300 to the product ions m/z 204, 186 and 144, matched well with those obtained for the SeGalNAc standards. The high uncertainty associated with the sample intensity ratios is attributed with the suppression effects and the low amount of SeGalNAc present. Taking all this into account we consider the values presented in Table 2 to be sufficient for verifying the identity of SeGalNAc in the liver cytosol. It should be noted that it was not possible to determine the 300 → 138 transition intensity, and thus use it for the transition ratio calculations because of its low sensitivity and its high background levels.

Table 2 SRM transition ratios determined for SeGalNAC standard and SeGalNAC detected in hepatocytosolic LMWSe fraction using tandem dual-column HPLC–APCI-MS/MS

Samples analysed	SRM intensity ratios
	m/z 204/186	m/z 204/144	m/z 204/138	m/z 186/144	m/z 186/138	m/z 144/138
SeGalNAc standard (n = 12)	3.03 ± 0.11	3.40 ± 0.07	3.93 ± 0.13	1.12 ± 0.04	1.30 ± 0.03	1.16 ± 0.03
LMWSe fraction (n = 7)	3.0 ± 0.5	2.9 ± 0.4	Not determined (n.d.)	1.0 ± 0.2	n.d.	n.d.

---

Selenium isotopic patterns resulting from the [SeGalNAc + H]⁺ molecular ion can also be useful for confirming the identity of SeGalNAc. This approach involved, for each of the four SRM transitions monitored, i.e. [SeGalNAc + H]⁺ → 204, 186, 144 and 138, the use of three different molecular ion m/z, each corresponding to a Se isotope. Thus the following SRM transitions were monitored: [⁸²SeGalNAc + H]⁺ (m/z 302) → 204, 186, 144 and 136; [⁸⁰SeGalNAc + H]⁺ (m/z 300) → 204, 186, 144 and 136; and [⁷⁸SeGalNAc + H]⁺ (m/z 298) → 204, 186, 144 and 136. The m/z of the product ions originating from the three different molecular ions, i.e. 302, 300 and 298, are identical as a result of the loss of the neutral moiety CH₃SeH upon CID at low energies.¹⁰ For each molecular ion m/z the four SRM transitions were monitored and their intensities used to calculate Se isotope ratios for both SeGalNAc standards and the LMWSe fraction. In the case of the hepatocytosolic fraction, only the most, and second most, abundant isotopes, i.e.⁸⁰Se and ⁷⁸Se, corresponding to molecular ions with m/z 300 and 298, were sensitive enough to provide useful analytical data for the extract sample. Also, only the two most sensitive product ions 204 and 186 were used. Thus the intensity ratio for [300 → 204]/[298 → 204] was determined to be 2.05 for the SeGalNAc standard and 2.05 ± 0.01 (n = 2) for hepatocytosolic SeGalNAc. Whereas, the ratio for [300 → 186]/[298 → 186] was found to be 2.03 for the SeGalNAc standard and 2.06 ± 0.15 (n = 2) for hepatocytosolic SeGalNAc. Thus a good match between the Se isotopic ratios in the standard and the LMWSe fraction was obtained, once again verifying the identity of SeGalNAc in liver cytosol. In addition, this data shows that the abundance of ⁸⁰Se is approximately double (2.06) that of ⁷⁸Se, which is in close agreement with IUPAC value¹² of 2.0872, taking into account the technique’s limited performance for measuring isotope ratios.

Even though the HPLC–APCI-MS/MS method was not applied in the present study for quantitative analysis , as this was done by HPLC–ICP-MS, a comparable LOD with ICP-MS detection was found for this method with an instrument LOD of 2.95 μg Se L⁻¹.

Based on all these results, the conclusive identification of SeGalNAc in liver samples has been demonstrated by use of tandem dual-column HPLC–ICP-MS and HPLC–APCI-MS/MS analyses.

Mass balance study and the estimation of SeGalNAc concentration in heptatocytosol

Mass balance determinations were carried out using ICP-MS and HPLC–ICP-MS for the analysis of whole liver samples and different fractions obtained during the liver extraction and the HPLC separations. In particular, the concentrations of SeGalNAc in biological samples were calculated based on the peak area from the dual-column HPLC system by using an external calibration curve method. The porcine liver used in this study had a water content of 71.1% ± 0.02% (n = 2). The total selenium content was 0.544 ± 0.09 mg Se kg⁻¹ wet weight (w.w.) (n = 4) and 1.76 ± 0.07 mg Se kg⁻¹ dry weight (n = 4) determined using ICP-MS and calculated using the external calibration method. The standard additions method was also applied and a value of 0.536 mg Se kg⁻¹ wet weight was obtained.

With the use of an aqueous buffer (30 mM TRIS–HCl, pH 7.4), 73.3% ± 0.03% (n = 3) of the total selenium in porcine liver was extracted (i.e., water soluble) in a single extraction process. Of the extractable selenium species, 74.2% ± 11.4% (n = 2) was present in the hepatocytosol (hepatocytosolic Se concentration in this study: 0.44 ± 0.12 μg Se L⁻¹ of cytosol (n = 2)). Combined with the Se content remaining in the pellet, the recovery of total selenium in the extraction procedure ranged from 87.4% to 108.4%. The concentration of SeGalNAc in porcine liver was determined to be 2.16 ± 0.94 μg Se kg⁻¹ (w.w.) (n = 3), which accounted for 0.42% of the selenium content in the whole porcine liver and 0.56% in the liver cytosol.

Conclusions

According to the metabolic pathway for selenium in rats proposed by Suzuki and co-workers, once selenium is taken up by the mammalian body, as an essential element, it is transported to the cell for biosynthesis of selenoproteins.¹³ Selenoproteins are degraded in the cell, and selenosugars seem to be produced in cells in the liver and SeGalNAc excreted into the bloodstream, and then into urine. So far, there has been a lack of detection of SeGalNAc in mammalian liver tissues under nonsupplemented selenium conditions. Our identification and quantification of naturally occurring SeGalNAc in porcine liver reveals the presence of trace amounts of SeGalNAc between 1.22–3.10 μg Se kg⁻¹ (w.w.). This finding further validates the proposed Se metabolism models involving the presence of SeGalNAc in the liver, even under normal Se dietary levels. To achieve this, a combination of analytical approaches were developed and optimised. The combined use of ICP-MS and APCI-MS/MS provided a powerful tool for selenium metabolite detection and identification in complex biological samples. The developed method is not only applicable to liver samples, but also holds potential for the analysis of other tissue types, such as biological fluids. However, it is stressed that both sample preparation and chromatographic separation are crucial for the successful identification of low level compounds in biological material. The present results indicated that signal suppression by coeluting matrix components may be a significant limitation factor for molecular mass spectrometry analysis. Organelle enrichment/separation techniques and a multidimensional separation strategy work efficiently towards the reduction of sample complexity and thus reduce matrix effects during MS detection. Application of efficient sample preparation and separation procedures allows HPLC–APCI-MS/MS to not only provide high selectivity, but also sensitivity comparable to ICP-MS detection. In addition, monitoring two or more SRM transitions for each analyte allows for the calculation of their intensity ratios which is particularly useful for verifying identification and therefore improving overall quality control. In addition this approach takes advantage of the fact that Se is a multi-isotopic element and thus its isotope pattern can be monitored by APCI-MS/MS thus providing additional data for the confident identification of selenium species.

Acknowledgements

The authors would like to acknowledge the funding of a Marie Curie Excellence grant (MEXT-CT-2003-002788) by the European Commission.

References

1. Y. Kobayashi, Y. Ogra, K. Ishiwata, H. Takayama, N. Aimi and K. T. Suzuki, PNAS, 2002, 99, 15932.
2. Y. Ogra, K. Ishiwata, Takayama, H. Takayama, N. Aimi and K. T. Suzuki, J. Chromatogr. B, 2002, 767, 301.
3. B. Gammelgaard, K. G. Madsen, J. Bjerrum, L. Bendahl, O. Jøns, J. Olsen and U. Sidenius, J. Anal. At. Spectrom., 2003, 18, 65.
4. B. Gammelgaard and L. Bendahl, J. Anal. At. Spectrom., 2004, 19, 135.
5. K. T. Suzuki, Y. Shiobara, M. Itoha and M. Ohmichib, Analyst, 1998, 123, 63.
6. Y. Shiobara, Y. Ogra and K. T. Suzuki, Analyst, 1999, 124, 1237.
7. Y. Kobayashi, Y. Ogra and K. T. Suzuki, J. Chromatogr. B, 2001, 760, 73.
8. C. Casiot, J. Szpunar, R. Łobinski and M. Potin-Gautier, J. Anal. At. Spectrom., 1999, 14, 645.
9. S. McSheehy, F. Pannier, J. Szpunar, M. Potin- Gautier and R. Lobinski, Analyst, 2002, 127, 223.
10. P. Traar, F. Belaj and K. A. Francesconi, Aust. J. Chem., 2004, 57, 1051.
11. S. Letsiou, V. Nischwitz, P. Traar, K. A. Francesconi and S. A. Pergantis, Rapid Commun. Mass. Spectrom., 2007, 21, 343.
12. IUPAC Commission on Atomic Weights and Isotopic Abundances, Pure Appl. Chem., 1997, 69, 2384.
13. K. T. Suzuki, J. Health Sci., 2005, 51, 107.

---