Formation of methylselenol, dimethylselenide and dimethyldiselenide in in vitro metabolism models determined by headspace GC-MS

Abstract

The aim of this study was to identify the presence of MeSeH in metabolic reactions. An analytical method based on direct headspace GC-MS, eliminating loss of volatile species during sample pretreatment procedures, was developed for this purpose. The in vitro conversion of selenium compounds to the volatile species methylselenol, MeSeH, dimethyl selenide, DMeSe and dimethyl diselenide, DMeDSe was investigated. The analytical method was evaluated by means of standards of dimethyl diselenide, dimethyl selenide. The corresponding sulfides were found unsuitable as internal standards as they interacted with the selenides. The limit of detection was 0.25 μmol L⁻¹ (20 μg L⁻¹) for the selenide as well as the diselenide. Formation of MeSeH was not observed in significant amount when selenomethionine was incubated with the enzyme L-methionine-γ-lyase; instead large amounts of DMeDSe were formed. In aqueous solution, methylseleninic acid, MeSeA reacted spontaneously with glutathione, GSH to form DMeDSe. In strongly reducing environments, however, MeSeH was also observed. When the formed MeSeH was trapped with iodoacetic acid, no DMeDSe was detected indicating that DMeDSe formation was due to spontaneous oxidation of MeSeH. These findings imply that DMeDSe may be a marker for the production of MeSeH in in vitro models. When MeSeA, Se-methylselenocysteine, Se-MeSeCys and SeMet were incubated with Jurkat cells, DMeDSe formation was only observed in the case of MeSeA. Trace amounts of DMeSe was observed in the vial with MeSeA as well as Se-MeSeCys. When DMeSe and DMeDSe were added to plasma, the sensitivity of only DMeDSe decreased significantly, implicating that DMeDSe underwent a reaction with plasma hindering the volatilization. This emphasizes that results from in vitro selenium metabolism studies may not be uncritically interpreted as consistent with the in vivo reality.

---

Introduction

The essential trace element selenium has for a long time been known to have some cancer protective effect. The effect is, however, more pronounced for some selenium compounds than others although the underlying mechanism is not fully understood.^1–3 The NPC trial was the first clinical study that showed the cancer protective effects of selenized yeast.⁴ In the following SELECT study, prevention of prostate cancer after selenium supplementation could not be proven.⁵ In the first study, selenized yeast was used as selenium source, while selenomethione, SeMet was used in the latter. Besides the major selenium compound, SeMet, selenized yeast contains several other selenium species,⁶ including the methylated species Se-methylselenocysteine, Se-MeSeCys⁷ and gammaglutamyl-MeSeCys.⁸ As different batches of selenized yeast have been shown to be very different in their composition,⁹ the different responses may be due to the different composition of the selenium compounds ingested. This emphasizes the importance of knowing the exact content of the Se product applied.

The generally accepted model for selenium metabolism was first described by Ganther¹⁰ and has in the course of time been further developed. In short, selenium ingested in the form of selenite and the amino acids SeMet and selenocysteine, SeCys, which are the most abundant amino acids in food, are transformed to a pool of selenide from which SeCys is produced and incorporated in the selenoproteins. Another pathway is supposed to go through the key metabolite methylselenol, MeSeH, which can be eliminated from Se-MeSeCys and SeMet by enzymes. Methylseleninic acid, MeSeA, which is often used as a model compound for methylated selenium compounds, is thus believed to be metabolised via the MeSeH pathway. Excess selenium is excreted by methylation, and dimethylselenide, DMeSe has been identified in breath, while the major urinary metabolite is a selenosugar. Metabolism schemes can be found in several recent reviews.^6,11,12

The conversion of SeMet to the active MeSeH demands presence of the enzyme L-methionine-γ-lyase. The catalytic action of this enzyme on SeMet and selenols were examined by Esaki et al. already in 1979.¹³ They concluded that the enzyme catalyses α,γ-elimination of SeMet to yield α-ketobutyrate, ammonia and MeSeH. This finding was based on the measurement of the resulting α-ketobutyrate. Furthermore, the enzyme catalyses γ-replacement reactions with various thiols and thereby produce S-substituted homocysteines.¹³ This conclusion was based on measuring the amount of amino acid formed with ninhydrin after paperchromatography.¹³ This is supported by in vitro experiments where tumor cells were transduced with the enzyme. The cytotoxicity of SeMet was increased up to 1000 fold in these cells compared to nontransduced cells.¹⁴

MeSeH is supposed to be the key metabolite in cancer protection, and selenium compounds that are able to produce a steady stream of this compound are considered to be the more efficient protective agents.^15,16 This was recently supported by an animal study. When MeSeA, Se-MeSeCys and SeMet were compared for their inhibitory efficacy against the human prostate cancer cell lines PC-3 and DU145 inoculated in mice, MeSeA and Se-MeSeCys, which are supposed MeSeH forming compounds, were more potent that SeMet and selenite, in spite of a larger tumor Se retention of SeMet.¹⁷

The role of Se in the cell cycle and apoptosis has recently been reviewed,¹¹ but so far only one study on the selenium metabolism in cancer cells has been reported.¹⁸ When MeSeA was incubated with lymphoma B-cells, the major metabolic product was identified as Se-MeSeCys by molecular MS and a minor metabolite was suggested to be SeMet.¹⁸ The same two metabolites were identified in another study on rat liver cells.¹⁹ In both studies, loss of volatile species was observed. DMeDSe was identified by GC-MS as the major metabolite while DMeSe was only a minor metabolite.¹⁸ DMeDSe was also identified by membrane inlet MS as the major reaction product of MeSeA when this compound was incubated in rat hepatocytes or with an aqueous solution of glutathione, GSH.²⁰

The most commonly used analytical method for analysis of volatile selenium species is gas chromatographic separation and ICP-MS detection after solid phase micro extraction, SPME.^21–24 Reported detection limits for this technique are in the sub-ppb range.^21,22 Recently, a headspace stir bar sorptive extraction combined with GC-ICP-MS was used to analyse garlic and onion samples. The detection limit of this method was 33 and 7.1 ng L⁻¹ for DMeSe and DMeDSe, respectively.²⁵ Cryotrapping-cryofocusing gas chromatography ICP-MS has been used for the analysis of DMeSe in breath²⁶ and cryogenic oven cooling GC-TOFMS was used to identify DMeSe and DMeDSe produced by cancer cells.¹⁸ DMeSe, DMeDSe and dimethylsulfurselenium were detected in urine by LC-ICP-MS²⁷ but the study revealed that the inherent advantage of ICP-MS in terms of equal sensitivities for all selenium species was lost as the sensitivity of volatile species was considerably larger compared to the non-volatile species sensitivity.²⁸

In spite of the supposed key role of MeSeH in cancer protection, data on the identification of MeSeH in mammal samples have never been published. The reason for this could be that this compound has been considered extremely reactive and therefore difficult to catch.²⁹ Instead, studies indirectly showing the presence of MeSeH and selenide by conversion to their respective oxidation products after oxidation with hydrogen peroxide have been reported.^30,31 However, results from selenium speciation of coffee, showed identification of MeSeH by GC-TOFMS in roasted coffee beans and brewed coffee from selenium accumulating coffee beans.²³

The aim of this study was to identify the presence of MeSeH as a reaction product in the above mentioned reactions, in which this compound is supposed to be produced. An analytical method based on direct headspace GC-MS, eliminating loss of volatile species during sample pretreatment procedures, was developed for this purpose.

Experimental

Reagents

All reagents were analytical grade. Selenium standards including dimethyl diselenide (DMeDSe), dimethyl selenide (DMeSe), dimethyl disulfide (DMeDS), dimethyl sulfide (DMeS), methylseleninic acid (MeSeA), L-selenomethionine (SeMet) and Se-methylseleno-L-cysteine (Se-MeSeCys) as well as L-methionine-γ-lyase, pyridoxal-5′-phosphate, sodium borohydiride (NaBH₄), ammonium formate and reduced glutathione (GSH) were purchased from Sigma-Aldrich (Steinheim, Germany). Potassium dihydrogen phosphate was purchased from Merck (Darmstadt, Germany) and methanol from VWR-Bie&Berntsen (Herlev, Denmark). Purified water was produced by a MilliQ_Plus system (Millipore, Bedford, USA). 10 mmol L⁻¹ DMeDSe, DMeSe stock solutions were prepared daily in methanol. MeSeH was prepared in situ in vials for HS-GC-MS analysis; 1.2 μl of DMeDSe was dissolved in 1 ml of ethanol and about 20 mg of NaBH₄ was added. The vial was capped with a crimp cap when the most vigorous fizz had declined and the yellow color of DMeDSe had disappeared. All other selenium standards were dissolved in purified water.

Instrumental

GC-MS

An Agilent 6890N gas chromatograph equipped with 5973N Mass Selective Detector and a 7683 Series Injector was used throughout (Agilent Technologies, Waldbronn, Germany). The injector was equipped with a 100 μl syringe; the injected volume was 50 μl. The split/split less inlet was operated in split mode. The inlet temperature was 250 °C and the split ratio was 50 : 1. Unless otherwise stated, the temperature of the autosampler was ambient, 23–23.5 °C. 1.5 ml glass vials with aluminium crimp caps and rubber/PTFE septa were used throughout. To avoid loss of volatiles, sample septa were only pierced once. Air blanks were run frequently and the injection needle was washed with methanol between injections to avoid carry-over between samples. The column was a Phenomenex Zebron ZB-50, 30 m × 0.25 mm ID × df 0.25 μm (Phenomenex, SupWare, Denmark). The flow rate of the helium carrier gas was 1 ml min⁻¹. The oven temperature was 60° isothermal for 1 min; 50° min⁻¹ to 140°; then 1 min isothermal. Ionization was obtained by electron impact and the electron energy was 70 eV. The mass detector was scanning m/z 75–200 or operated in SIM mode monitoring m/z 96, 110, 190, dwell time 50 ms.

LC-ICP-MS

The ICP-MS was a PE Sciex Elan 6000 (Perkin Elmer, Norwalk, CT, USA) equipped with a Micro Mist glass concentric nebulizer (Glass Expansion, Romainmontier, Switzerland) and a PC³ cyclonic spraychamber (Elemental Scientific Inc., Omaha, NE, USA). ICP-MS sampler and skimmer cones were made of platinum. The plasma and auxiliary gas flow rates were 15 L min⁻¹ and 1.2 L min⁻¹, respectively. The nebulizer gas flow, lens voltage and ICP RF power were optimized regularly with a solution of 100 μg Se L⁻¹ in mobile phase.

The data requisition settings were: dwell-time 500 ms, sweeps per reading 1 and readings per replicate were varied corresponding to chromatographic runtime. ⁷⁷Se, ⁷⁸Se and ⁸²Se isotopes were monitored. The chromatographic separation was performed with a G1376A capillary pump, a G1379A degasser and a G1313A autosampler, all from Agilent 1100 series, controlled by ChemStation software (Agilent Technologies, Waldbronn, Germany). The column was a Gemini C18 5 μ 110 Å 2 mm ID × 250 mm (Phenomenex, SupWare, Denmark), ambient temperature. Flow rate 200 μL min⁻¹, injection volume 5 μL. The mobile phase was 18 mmol L⁻¹ ammonium formate pH 7 in 2% methanol.

LC-ESI-MS

The ESI-MS ion trap detector was an Esquire-LC G1979A (Bruker Daltonics GmbH, Bremen, Germany). The electrospray was produced in the negative ionization mode. The ESI Ion Trap MS parameters were; capillary voltage: −4000 V, nebuliser pressure: 30 psi, drying temperature: 350 °C, drying gas flow: 10 L min⁻¹ and the trap drive was 33.7 (arbitrary unit). The ESI-MS was coupled to a LC system consisting of a G1379A degasser, and a G1312A binary pump. The samples were injected with a 7725i manual injection valve from Rheodyne and the injection volume was 5 μl. (Agilent Technologies, Waldbronn, Germany). The column was the same as for LC-ICP-MS analysis; Phenomenex Gemini C18 5μ 110 Å 2 mm ID × 250 mm, ambient temperature. Flow rate 200 μL min⁻¹. The mobile phase was 18 mmol L⁻¹ ammonium formate pH 7 in 2% methanol.

Matrix effects

A solution containing 900 μmol L⁻¹ DMeDSe and 500 μmol L⁻¹ DMeSe was prepared by aqueous dilution of 10 mmol L⁻¹ methanolic stock solutions. 50 μL of this solution was added to 950 μL samples consisting of urine or plasma diluted with water to obtain concentrations of 5, 10, 20, 30, 40, 50, 60, 70, 80 or 95%. The samples were gently shaken and analyzed by HS-GC-MS in SIM mode monitoring m/z 110 and 190.

In vitro metabolism

1 mmol L⁻¹ SeMet or Se-MeSeCys was incubated with 0.02 U ml⁻¹L-methionine-γ-lyase and 10 μmol L⁻¹ pyridoxal-5′-phosphate in 50 mmol L⁻¹ potassium phosphate buffer pH 8 for 2 h at 37 °C. Analysis of vial headspace was done every 15 min in SIM mode (m/z 96, 110, 190). Jurkat cells were obtained from the Faculty of Life Sciences, University of Copenhagen; cells were cultured in Dulbecco's modified Eagle's medium to a cell density of 10⁶ cells ml⁻¹. 1 ml cell suspension was incubated in glass vials with crimp caps at 37° for 24 h. The concentration of the selenium compound (MeSeA, SeMet, Se-MeSeCys) was 100 μmol L⁻¹.

Total selenium determination

Total determination of selenium was obtained with an Elan DRCe ICP-MS equipped with a jacketed cyclonic spraychamber and a MicroMist microconcentric nebulizer (both from Glass Expansion, West Melbourne, Vic, Australia). Methane was used as reaction gas. The nebulizer gas flow, lens voltage and ICP RF power were optimized regularly with a solution of 100 μg Se L⁻¹. Samples were introduced with a flow rate of 1 ml min⁻¹ and the following sample acquisition parameters were used; 25 sweeps/reading, 1 reading/replicate and 5 replicates. Quantification was based on 3 level standard additions.

Results and discussion

Method development

Headspace GC-MS analysis in scan mode was applied for recording mass spectra of standards of MeSeH, DMeDSe and DMeSe. The chromatogram showed well resolved peaks of the three species. DMeDSe and DMeSe are commercially available whereas MeSeH was prepared in situ by reduction of DMeDSe with NaBH₄. Mass spectra of the standards are shown in Fig. 1. The characteristic selenium isotope patterns of compounds containing one or two selenium isotopes were observed. The most abundant ion was m/z 96, 110 and 190 for MeSeH, DMeSe, and DMeDSe, respectively. For all selenium species (DMeDSe, DMeSe and MeSeH) it is difficult to find the selenium isotopic pattern for the EI-fragment ion corresponding to MeSe⁺. The EI ionization method produces fragmentation patterns that are very different from fragmentation patterns observed by soft ionization methods such as electrospray. The disturbance of the isotopic pattern is probably caused by fragment ions corresponding to loss of varying numbers of protons, shifting the mass by one unit. In the case of DMeDSe and DMeSe standards are available to confirm the spectra. However, no MeSeH standard is available and this species is produced in situ as described in the literature. Meija et al.²³ reported the three most abundant fragment ions and their relative intensity; 96 (100), 93 (90) and 80 (80). The figures correspond very well with the mass spectrum obtained in our study and this is considered evidence of formation and detection of MeSeH.

Fig. 1 Analysis of standards. HS-GC-MS (m/z 75–200) of a mixture of DMeDSe (100 μmol L⁻¹) and DMeSe (100 μmol L⁻¹) in water (upper chromatogram). MeSeH was prepared in situ by addition of NaBH₄ to an ethanolic solution of DMeDSe (13 mmol L⁻¹) (lower chromatogram). Compounds 2 and 3 are side products from the vigorous reduction reaction and do not contain selenium. Based on the software integrated spectrum library, compound 2 was tentatively identified as methylated boron; where as the identity of compound 3 is unknown.

When analysing the in situ formed MeSeH, DMeDSe was also observed in the chromatogram getting more abundant over time. This indicates that DMeDSe is the spontaneously formed oxidation product of MeSeH.

Gas was evaporated vigorously upon reduction of DMeDSe with NaBH₄ which made quantitative analysis of MeSeH practically impossible. Hence, the method was only evaluated for determination of DMeDSe and DMeSe in aqueous solution. The mass detector was operated in SIM mode (m/z 110 + 190). For both species, linearity was observed in the concentration range of 0.5 to 1000 μmol L⁻¹. The precision determined as RSD was less than 10% (n = 10). LOD determined as 3σ of analysis of a 0.5 μmol L⁻¹ standard was 0.25 μmol L⁻¹.

The sulfur analogs of DMeDSe and DMeSe, DMeDS and DMeS, respectively were evaluated as potential internal standards. Both the sulfur analogs and the selenium containing metabolites of interest were chromatographically well resolved. However, as dimethylsulfurselenium (DMeSeS) was formed spontaneously in an aqueous mixture of DMeDSe and DMeDS (Fig. 2); DMeDS was not a suitable internal standard for DMeDSe. There was no spontaneous interaction between the monoselenide and the sulfide, but internal standardisation with the sulfides was not further pursued.

Fig. 2 Interaction of DMeDSe with the sulfur analog DMeDS. HS-GC-MS (m/z 60–200) analysis of an aqueous mixture of DMeDS (500 μmol L⁻¹) and DMeDSe (500 μmol L⁻¹) and mass spectrum of the formed DMeSeS. The observed EI fragmentation pattern of DMeSeS is consistent with the fragmentation of synthesized reference standards reported by Meija et al.²³

The headspace sensitivity of DMeDSe and DMeSe was investigated with respect to temperature, salt and organic solvents to assess the robustness of the method in complicated sample matrices. As DMeDSe and DMeSe are insoluble in water, stock solutions were prepared in methanol. The influence of decreasing methanol concentration (100% to 5%) on headspace sensitivity was accordingly investigated. At least 400 fold and 15 fold enhancement of the peak area was observed for DMeDSe and DMeSe, respectively when the methanol concentration was reduced from 100 to 5%. Fortunately, methanol and other organic solvents are usually not present in biological samples and samples from in vitro metabolism experiments, and organic solvents will therefore not impose any unwanted sensitivity suppression. Plasma was used as a model matrix for the influence of proteins and biomacromolecular constituents. The plasma matrix strongly influenced the headspace sensitivity of DMeDSe, where as DMeSe sensitivity was not influenced. The headspace sensitivity of DMeDSe decreased with increasing plasma to water ratio and it was not possible to detect any DMeDSe (in a 45 μmol L⁻¹ sample) at plasma ratios above 60%. These results indicate that DMeDSe readily reacts with plasma constituents resulting in the formation of non-volatile selenium species. Human urine was used as a high salt model matrix. The headspace sensitivity of DMeSe was not influenced by the high salt matrix, while randomly large imprecision of 3 replicate samples was induced to the determination of DMeDSe. However, no general signal suppression with increasing urine concentration was observed. The method is not well suited for analysis of DMeDSe in high salt matrices. However, inherently high salt matrices from biological experiments are not relevant due to the very apolar nature of DMeDSe. Increasing the autosampler temperature generally increased the signals. 8-fold and 3-fold increases in signal intensity for DMeSe and DMeDSe, respectively was observed when increasing the autosampler temperature from 20 °C to 60 °C. In the analysis of biological samples the temperature is not elevated as increasing temperature imposes the risk of false results due to sample degradation.

Spontaneous reactions involving DMeDSe

The main objective of this study was to develop a method for catching the reactive MeSeH in in vitro experiments i.e. enzyme mixtures, tissue homogenates, isolated cells or cell cultures. The generally accepted theory is that MeSeH is produced spontaneously from MeSeA by reaction with thiols such as glutathione (GSH).

To confirm the formation of MeSeH, MeSeA, DMeDSe and DMeSe were mixed with large amounts of the reducing agent, NaBH₄. MeSeH was produced by reduction of both MeSeA and DMeDSe, whereas DMeSe was stable in the presence NaBH₄. (Fig. 3a). As MeSeA is in the oxidation state of +2, DMeDSe is in the oxidation state −1 and DMeSe and MeSeH −2, this finding agrees with what would be expected. In a mixture of DMeDSe and a large excess of GSH, MeSeH was also produced (Fig. 3b).

Fig. 3 Spontaneous chemical reductions. HS-GC-MS (m/z 75–200) analysis of (a) NaBH₄ reduced selenium standards (upper) DMeSe (65 mmol L⁻¹), (middle) MeSeA (25 mmol L⁻¹) and lower) DMeDSe (13 mmol L⁻¹). (b) DMeDSe (13 mmol L⁻¹) reduced with GSH (117 mmol L⁻¹). 1 = MeSeH, 2 = DMeSe and 3 = DMeDSe.

The stability of MeSeH prepared from DMeDSe with limited amounts of NaBH₄ was followed by GC-MS for 4 h. Initially, MeSeH was the predominant species in the reaction mixture, followed by rapid disappearance of MeSeH and appearance of DMeDSe until a close to steady-state condition was obtained. Only trace amounts of MeSeH was observed in this steady-state condition. The higher the concentrations of NaBH₄, the longer the presence of MeSeH lasted. In the human body, GSH is ubiquitously present and constantly formed by the action of glutathione reductase in order to provide the required redox state of tissues and cells.³² This means that in vivo GSH may be present in far larger amount than MeSeH or DMeDSe.

To verify that DMeDSe was formed via the MeSeH intermediate, DMeDSe was reduced with NaBH₄ in the presence of iodoacetate to trap the selenol group as the acetate seleno ester. In presence of iodoacetate, neither MeSeH nor DMeDSe were observed in the sample headspace (Fig. 4a). Thus, all DMeDSe was reduced by NaBH₄ and the formed MeSeH reacted with iodoacetate to form a non-volatile product. This indicates that DMeDSe observed in the sample headspace without the presence of iodoacetate is a product of spontaneous oxidation of MeSeH. Additional peaks observed were iodomethane and methylated boron and it even seems that DMeSe is a product formed during the rigorous reduction reaction. A similar experiment has earlier been performed by Ganther, who trapped the unstable selenol, glutathione persulfide (GS-SeH) by this reagent.³³ In LC-ICP-MS analysis, two selenium containing compounds were detected (Fig. 4b). The major peak was identified by LC-ESI-MS as 2-(methylseleno)-acetate (Fig. 4c) the theoretical product of the reaction of MeSeH and iodoacetate. The identity of the minor compound was not established. The results of the trapping experiment confirm that DMeDSe is reduced by NaBH₄ and that all MeSeH is trapped by iodoacetate as the peaks are missing in the HS-GC-MS chromatogram; hence the presence of DMeDSe in the non-trapped sample is probably due to the spontaneous oxidation of MeSeH. This indicates that DMeDSe is a spontaneous oxidation product of MeSeH, and DMeDSe observed in metabolism experiments may be used as a marker for MeSeH production. This finding could open the suggestion of trapping MeSeH with iodoacetate to form 2-(methylseleno)-acetate which could easily be analysed by the more sensitive LC-ICP-MS technique for in vitro metabolism experiments. However, iodoacetate may react with endogenous thiols with the risk of inactivating important enzymes and/or disturbing essential functions of the in vitro models.

Fig. 4 Results from experiment on trapping of MeSeH. (a) HS-GC-MS (m/z 75–200) analysis of (upper) 13 mmol L⁻¹ DMeDSe reduced with NaBH₄, (middle) 13 mmol L⁻¹ DMeDSe reduced with NaBH₄ in the presence of 50 mmol L⁻¹ iodoacetate and lower) injection of blank air. (b) LC-⁸²Se-ICP-MS analysis of 13 mmol L⁻¹ DMeDSe reduced with NaBH₄ in the presence of 50 mmol L⁻¹ iodoacetate and (c) LC-ESI-MS spectrum of the major selenium containing peak observed by LC-ICP-MS (Rt 5.4–5.7 min). The spectrum correspond to 2-methylselenol)-acetate.

In vitro metabolism

Enzymatic reactions. When 1 mmol L⁻¹ SeMet was incubated with L-methionine-γ-lyase for 1 h, DMeDSe was the major volatile species detected in the headspace of the sample. A small peak was observed at the retention time of MeSeH as well (Fig. 5a). The data acquisition was performed in the SIM mode (m/z 96 (MeSeH), 110 (DMeSe), 190 (DMeDSe). Surprisingly, DMeDSe production was also observed when incubating Se-MeSeCys under the same conditions. L-methionine-γ-lyase from bacterial sources is known to catalyse α,γ-elimination of SeMet.¹³ The products of the elimination reaction are MeSeH, NH₃ and the corresponding α-keto acid. Thus, the formation of DMeDSe indicates that the MeSeH intermediate was present and the presence of the small amount of MeSeH indicates that an equilibrium between the reduced and oxidized form may exist. This was supported by the finding of small amounts of MeSeH in the DMeDSe standard.

Fig. 5 In vitro formation of volatile selenium metabolites. (A) HS-GC-MS (m/z 96, 110, 190) analysis of 1 mmol L⁻¹ SeMet incubated with 0.02 U ml⁻¹L-methionine-γ-lyase, 10 μmol L⁻¹ pyridoxal-5′-phosphate, 50 mmol L⁻¹ potassium phosphate buffer pH 8, 37°, 1 h. (B) HS-GC-MS (m/z 96, 110, 190) analysis of Jurkat cells (10⁶ cells ml⁻¹) incubated with 100 μmol L⁻¹ of MeSeA (upper) 100 μmol L⁻¹ Se-MeSeCys (lower). Chromatograms in B are blank corrected.

Cancer cells. Cells from a T-cell leukemia (Jurkat) cell line were incubated with SeMet, Se-MeSeCys and MeSeA for 24 h (Fig. 5b). Based on the results from the study of the L-methionine-γ-lyase study described above it was reasonable to expect that DMeDSe would be detected if the enzymes were present and functional. No MeSeH was observed in sample headspaces and DMeDSe was observed only upon incubation of MeSeA suggesting that neither Se-MeSeCys nor SeMet were metabolized into MeSeH, whereas MeSeA produced large amounts of MeSeH. Based on total selenium analysis by ICP-MS only 12% of the dosed selenium was recovered indicating a major formation of volatile metabolites. In contrast, merely all dosed Se-MeSeCys and SeMet was recovered unchanged in the growth medium of Jurkat cells, suggesting formation of volatile metabolites only at the trace level. This is in accordance with the commonly accepted metabolism scheme, as MeSeH formation from SeMet and Se-MeSeCys demands the presence of the enzymes methionase and β-lyase, respectively and these enzymes were apparently not present or functional in the present cell model. Jurkat cell selenium metabolism has not been investigated before. However, DMeDSe was also reported as a major metabolite of MeSeA in another lymphoma cell line (DHL-4)¹⁸ and healthy hepatocytes from the rat.²⁰ These findings are also in agreement with results from Ip et al.¹⁶ Although they observed similar cancer inhibitory efficiency of MeSeA and Se-MeSeCys in vivo in a rat mammary cancer model, they were not able to demonstrate equal efficiency in in vitro cancer cell models. Accordingly they concluded that the cancer cell lines may have only modest ability to generate MeSeH from Se-MeSeCys via the enzymatic pathway. No speciation results were reported in the study.

DMeSe was observed in trace amounts in the Jurkat cells treated with MeSeA and Se-MeSeCys, while this metabolite could not be detected in the SeMet incubations. DMeSe is considered a methylated excretion product and this metabolite has been identified in breath from humans²⁶ after ingestion of isotope labeled selenite. The reason why this metabolite only appears in cells incubated with MeSeA and Se-MeSeCys is not readily explained and further experiments are needed to elucidate the mechanism.

In conclusion, a headspace GC-MS method was developed for identifying the volatile selenium species MeSeH, DMeSe and DMeDSe. Mass spectra of in situ MeSeH production from reduction of MeSeA and DMeDSe were presented for the first time. MeSeH was only detected in reducing environment and was readily oxidized to DMeDSe. The production of DMeDSe was demonstrated to proceed via the MeSeH intermediate by trapping of this as the non-volatile acetylseleno ester. When Jurkat cells were incubated with MeSeA, SeMet and Se-MeSeCys, only cells incubated with MeSeA produced DMeDSe indicating that MeSeH was only produced from this species in this in vitro model.

Acknowledgements

The authors thank post doc. Lars Andresen and associate professor Søren Skov, Faculty of Life Sciences, University of Copenhagen for donation of the Jurkat cells.

This project was supported by a grant (271-07-0302) from the Danish Research Council.

References

1. M. P. Rayman, Proc. Nutr. Soc., 2005, 64, 527–542.
2. P. D. Whanger, Br. J. Nutr., 2004, 91, 11–28.
3. C. D. Thomson, Eur. J. Clin. Nutr., 2004, 58, 391–402.
4. L. C. Clark, G. F. Combs Jr, B. W. Turnbull, E. H. Slate, D. K. Chalker, J. Chow, L. S. Davis, R. A. Glover, G. F. Graham, E. G. Gross, A. Krongrad, J. L. Lesher Jr, H. K. Park, B. B. Sanders Jr, C. L. Smith and J. R. Taylor, JAMA, J. Am. Med. Assoc., 1996, 276, 1957–1963.
5. S. M. Lippman, E. A. Klein, P. J. Goodman, M. S. Lucia, I. M. Thompson, L. G. Ford, H. L. Parnes, L. M. Minasian, J. M. Gaziano, J. A. Hartline, J. K. Parsons, J. D. Bearden, E. D. Crawford, G. E. Goodman, J. Claudio, E. Winquist, E. D. Cook, D. D. Karp, P. Walther, M. M. Lieber, A. R. Kristal, A. K. Darke, K. B. Arnold, P. A. Ganz, R. M. Santella, D. Albanes, P. R. Taylor, J. L. Probstfield, T. J. Jagpal, J. J. Crowley, F. L. Meyskens, L. H. Baker and C. A. Coltman, JAMA, J. Am. Med. Assoc., 2009, 301, 39–51.
6. B. Gammelgaard, C. Gabel-Jensen, S. Sturup and H. R. Hansen, Anal. Bioanal. Chem., 2008, 390, 1691–1706.
7. H. G. Infante, G. O'Connor, M. Rayman, R. Wahlen, J. Entwisle, P. Norris, R. Hearn and T. Catterick, J. Anal. At. Spectrom., 2004, 19, 1529–1538.
8. H. G. Infante, G. O'Connor, M. Rayman, R. Wahlen, J. E. Spallholz, R. Hearn and T. Catterick, J. Anal. At. Spectrom., 2005, 20, 864–870.
9. E. H. Larsen, M. Hansen, H. Paulin, S. Moesgaard, M. Reid and M. Rayman, J. AOAC Int., 2004, 87, 225–232.
10. H. E. Ganther, J. Am. Coll. Tocicol., 1986, 5, 1–5.
11. H. W. Zeng, Molecules, 2009, 14, 1263–1278.
12. H. W. Zeng and G. F. Combs, J. Nutr. Biochem., 2008, 19, 1–7.
13. N. Esaki, H. Tanaka, S. Uemura, T. Suzuki and K. Soda, Biochemistry, 1979, 18, 407–410.
14. K. Miki, M. Xu, A. Gupta, Y. Ba, Y. Tan, W. Al-Refaie, M. Bouvet, M. Makuuchi, A. R. Moossa and R. M. Hoffman, Cancer Res., 2001, 61, 6805–6810.
15. C. Ip, J. Nutr., 1998, 128, 1845–1854.
16. C. Ip, H. J. Thompson, Z. Zhu and H. E. Ganther, Cancer Res., 2000, 60, 2882–2886.
17. G. X. Li, H. J. Lee, Z. Wang, H. Hu, J. D. Liao, J. C. Watts, G. F. Combs and J. Lu, Carcinogenesis, 2008, 29, 1005–1012.
18. H. G. Infante, S. P. Joel, E. Warburton, C. Hopley, R. Hearn and S. Juliger, J. Anal. At. Spectrom., 2007, 22, 888–896.
19. C. Gabel-Jensen, J. Odgaard, C. Skonberg, L. Badolo and B. Gammelgaard, J. Anal. At. Spectrom., 2009, 24, 69–75.
20. C. Gabel-Jensen, S. A. Bak, F. R. Lauritsen, H. R. Hansen, L. Badolo and B. Gammelgaard, J. Anal. At. Spectrom., 2009, 24, 949–952.
21. C. Dietz, J. S. Landaluze, P. Ximenez-Embun, Y. Madrid-Albarran and C. Camara, J. Anal. At. Spectrom., 2004, 19, 260–266.
22. J. Meija, M. Montes-Bayon, D. L. Le Duc, N. Terry and J. A. Caruso, Anal. Chem., 2002, 74, 5837–5844.
23. J. Meija, J. M. Bryson, A. P. Vonderheide, M. Montes-Bayon and J. A. Caruso, J. Agric. Food Chem., 2003, 51, 5116–5122.
24. S. Mounicou, M. Shah, J. Meija, J. A. Caruso, A. P. Vonderheide and J. Shann, J. Anal. At. Spectrom., 2006, 21, 404–412.
25. J. Duan, X. Li, C. Yu and B. Hu, J. Anal. At. Spectrom., 2009, 24, 297–303.
26. D. Kremer, G. Ilgen and J. Feldmann, Anal. Bioanal. Chem., 2005, 383, 509–515.
27. D. Juresa, J. Darrouzes, N. Kienzl, M. Bueno, F. Pannier, M. Potin-Gautier, K. A. Francesconi and D. Kuehnelt, J. Anal. At. Spectrom., 2006, 21, 684–690.
28. D. Juresa, D. Kuehnelt and W. Frech, Anal. Chem., 2006, 78, 8569–8574.
29. K. T. Suzuki, K. Kurasaki, S. Ogawa and N. Suzuki, Toxicol. Appl. Pharmacol., 2006, 215, 189–197.
30. Y. Ohta and K. T. Suzuki, Toxicol. Appl. Pharmacol., 2008, 226, 169–177.
31. K. T. Suzuki, K. Kurasaki and N. Suzuki, Biochim. Biophys. Acta, Gen. Subj., 2007, 1770, 1053–1061.
32. A. Pastore, G. Federici, E. Bertini and F. Piemonte, Clin. Chim. Acta, 2003, 333, 19–39.
33. H. E. Ganther, Biochemistry, 1971, 10, 4089–4098.

---