Lars
Bendahl
* and
Bente
Gammelgaard
Department of Analytical Chemistry, The Danish University of Pharmaceutical Sciences, Universitetsparken 2, DK-2100 Copenhagen, Denmark. E-mail: labe@dfh.dk; Fax: +45 3530 6010
First published on 21st November 2003
The selenium species in nutritional supplement tablets, based on selenized yeast, were separated by capillary zone electrophoresis using capillaries coated dynamically with poly(vinyl sulfonate) and detected by ICP-MS. Sample pre-treatment consisted of cold-water extraction by sonication and subsequent incubation of the cold-water extract with 6 M hydrochloric acid at 110 °C. The total selenium concentration in the cold-water extract was 3.5 mg L−1 and corresponded to 9% of the total selenium content of the tablets. More than 20 different selenium compounds were separated in the cold-water extract within 13 min. The efficiency of the system corresponded to 620000 theoretical plates. When spiking the sample with available standards, co-migration was observed with selenomethionine and selenocystine–Se-methylselenocysteine—the latter species were not separated. When the cold-water extract was hydrolysed in hot hydrochloric acid, 45% of the selenium migrated within a single peak that co-migrated with selenomethionine. Other peaks co-migrated with trimethylselenonium, Se-methylselenomethionine, and selenocystine–Se-methylselenocysteine, respectively. The precision for the analysis of the aqueous extracts expressed as relative standard deviation (n = 3) on peak heights and areas was in the range 1.4–5.3%. Detection limits were better than 15 µg L−1, corresponding to absolute detection limits less than 250 fg.
HPLC is the preferred separation technique for speciation studies and more than 20 selenium compounds have been separated in aqueous yeast extracts by ion-pairing chromatography.6,7 The identities of most of the selenium compounds in yeast samples have been suggested on the basis of co-elution with authentic standards, while the presence of selenomethionine and Se-adenosyl-selenohomocysteine in yeast samples has been confirmed by electrospray ionization (tandem) mass spectrometry.8–10 Additionally, several new selenium compounds have been detected by HPLC and ESI-MS(MS), including a new class of conjugates between glutathione-S and selenium11 and some recently discovered Se-adenosyl compounds.12,13 An overview of the selenium compounds so far suggested to be present in yeast samples by HPLC-ICP-MS and ESI-MS is shown in Table 1.
Identified species | Analytical technique | References |
---|---|---|
a Only applied in reference 9. | ||
Selenomethionine | IPC-ICP-MS and (IPC-ESI-MS)a | 6,7,9,14,22,29,34–37 |
IEC-ICP-MS | 23,38 | |
RPC-ICP-MS | 7 | |
Chiral LC-ICP-MS | 39 | |
Selenocystine | IPC-ICP-MS | 6,7,34–37 |
IEC-ICP-MS | 38 | |
RPC-ICP-MS | 7 | |
Se-methylselenocysteine | IPC-ICP-MS | 6,22,36,37 |
IEC-ICP-MS | 23 | |
Chiral LC-ICP-MS | 39 | |
Selenite | IPC-ICP-MS | 7,22,37 |
IEC-ICP-MS | 38 | |
RPC-ICP-MS | 7 | |
Se-cystathionine | IPC-ICP-MS | 37 |
Se-lantathionine | IPC-ICP-MS | 37 |
Chiral LC-ICP-MS | 39 | |
Selenomethionine-Se-oxide | IEC-ICP-MS | 23 |
Selenate | IEC-ICP-MS | 38 |
RPC-ICP-MS and IPC-ICP-MS | 7 | |
Selenoethionine | RPC-ICP-MS and IPC-ICP-MS | 7 |
Trimethylselenonium | RPC-ICP-MS and IPC-ICP-MS | 7 |
Se-adenosylselenohomocysteine | Chiral LC-ICP-MS | 39 |
SEC/ICP-MS and off-line ESI-MS(MS) | 8 | |
IPC-ESI-MS and IPC-ICP-MS | 9 | |
SEC-IEC-ICP-MS and off-line ESI-MS(MS) | 12 | |
Se-adenosyl-compounds | SEC-ICP-MS and off-line ESI-MS(MS) | 13 |
SEC-IEC-ICP-MS and off-line ESI-MS(MS) | 12 | |
Glutathione-S-selenium conjugates | SEC-IEC-ICP-MS and off-line ESI-MS(MS) | 11 |
The presence of selenocystine has been suggested in several studies (Table 1). Although it could be argued that selenocystine would not be present in selenized yeast as the yeast genome does not code for selenocysteine incorporation in proteins, selenocystine may be formed by other metabolic pathways. Selenoethionine is a synthetic standard and is hardly present in biological samples. On the contrary, this compound is often used as an internal standard.6,14
A large number of the separated selenium species in yeast have not yet been identified and the picture is further complicated by the fact that different fabrications of selenized yeast contain different combinations of different selenium compounds.
Capillary electrophoresis is an attractive alternative separation technique in selenium speciation due to its high separation power and its ability to separate polar and charged compounds, which are difficult to retain on reversed-phase stationary phases even when ion-paring agents are added to the mobile phase.7,14 Extracts of selenized yeast samples contain a large number of selenium compounds in high concentrations and have a low conductivity as compared with urine and plasma samples. Thus, CE-ICP-MS is in many respects a very suitable method for the analysis of yeast samples. However, concentration detection limits for selenium in CE-ICP-MS are generally two orders of magnitude higher than detection limits with HPLC-ICP-MS, and so far the technique has only been applied successfully for analysis of pre-concentrated and selenium enriched biological samples.
Only two applications of CE-ICP-MS for analysis of selenized yeast samples have been reported in the literature.10,15 Day et al. applied CE-ICP-MS for chiral speciation of L- and D-selenomethionine derivatized with 1-fluoro-2,4-dinitrophenyl-5-L-alanine amide in enzymatically digested yeast samples.15 The results supported previously reported results from chromatographic experiments in showing that only the L-form of selenomethionine was present in the samples. Recently, CE-ICP-MS was applied for analysis of the water soluble fraction of selenized yeast by Mounicou et al.10 They used an alkaline electrophoresis medium and reversed polarity detection. However, a number of the selenium compounds in the extract were not baseline separated and 25–30% of the selenium content did not leave the capillary within an analysis time of 30 min: the authors concluded that chromatographic fractionation was necessary prior to analysis by CE-ICP-MS.10 Recently, a dynamic capillary coating procedure with unique separation properties in acidic electrophoresis media was published.16
The aim of this study was to examine the suitability of capillary electrophoresis with ICP-MS detection in dynamically coated capillaries for speciation of selenium in yeast by applying the technique on yeast based selenium nutritional supplement tablets.
CE-ICP-MS | |
---|---|
a External argon supply. | |
Sampling and skimmer cones | Direct injection nebulizer |
Sample introduction system | Platinum |
Argon flow rate | |
Plasma gas | 1.2 L min−1 |
Auxiliary gas | 15 L min−1 |
Nebulisationa | 0.2 L min−1 |
Rf power supply | 1150 W |
Lens voltage | 6 V |
Data acquisition | |
Dwell time | 300 ms |
Sweeps/reading | 1 |
Readings/replicate | 1500 |
Replicates | 1 |
Isotopes monitored | 82Se |
Sheath liquid uptake rate | 10 µL min−1 |
2 mL of the supernatant was hydrolysed by incubation with 2 mL of concentrated hydrochloric acid in a sealed vial for 24 hours at 110 °C. The resulting solution was filtered through a 0.45 µm cellulose acetate membrane and evaporated to dryness on a rotary evaporator at 40 °C. The residue was re-dissolved in 2 mL water and re-evaporated to dryness in order to remove excess hydrochloric acid. Finally, the residue was reconstituted in 2 ml water.
The total selenium content in the extracts was determined by the method of standard additions at five concentration levels using selenite as internal standard.
The selenium samples analysed in this study were nutritional supplement tablets, declared to contain 100 µg of selenium in the form of selenized yeast. In order to conserve the identity of the selenium compounds, the tablets were extracted with cold water in an ultrasonic bath followed by centrifugation and ultra-filtration of the pale yellow supernatant through a cellulose membrane with a cut-off at 5 kDa. Protective additives and buffer salts were not added as they could interfere in the stacking process in CE by increasing the conductivity of the sample, resulting in decreased resolution. The selenium concentration in the extracts was found to be 3.5 mg L−1, corresponding to 9% of the total selenium content. This is in accordance with earlier findings.20
Fig. 1 Electropherogram from analysis of the cold water extract of yeast based nutritional supplement tablets by CE-ICP-MS. Inset on the left side of the figure is a trace acquired during flushing of the capillary with 1.5 times capillary volume of buffer after the analysis was completed. Conditions: capillary, 110 cm × 50 µm id coated with poly(vinyl sulfonate); CE buffer, 100 mmol L−1 formic acid, pH 3.0, added 0.01% poly(vinyl sulfonate); sheath liquid, CE buffer added 10% methanol (10 µL min−1); run voltage, 30 kV; hydrostatic sample injection at 9.82 mbar for 100 s corresponding to 16.7 nL (60 pg Se). |
The efficiency of the system, calculated on the basis of the peak at a migration time of 11.85 min, was 620000 theoretical plates. This is, to our knowledge, the best separation efficiency achieved in the separation of selenium compounds by capillary electrophoresis to date. Thus, the resolution is superior to results achieved with reversed-phase, ion-pair and anion-exchange chromatography,6,7 where a column efficiency of 8500 theoretical plates in ion-pair chromatography was obtained. Moreover, the analysis time in capillary electrophoresis was one third of the analysis time in ion-pair chromatography.6
Peak widths less than 1 s at base line were achieved for some of the fastest migrating compounds. This is an order of magnitude less than the peak widths achieved in other studies.10
Inset in the left side of Fig. 1 is the trace acquired during flushing of the capillary with 1.5 times the capillary volume of buffer after the analysis was completed. The trace shows that 2 or 3 selenium species remained undetected within an analysis time of 15 min. This is in contrast to the results achieved when samples were separated in an alkaline CE buffer.10 However, this fraction only represented a small amount of the total selenium and no further effort was made to characterize these compounds.
When extracts were analyzed before the ultra-filtration through the 5 kDa membrane, no additional peaks appeared in the electropherograms; but the amount of selenium which could be flushed out with buffer after the analysis was completed increased slightly (data not shown). Thus, the majority of the selenium compounds in the water-soluble fraction were low molecular weight compounds. This is in accordance with previously reported results.6,21
Fig. 2 Expanded electropherograms from analysis of the cold water yeast extract by CE-ICP-MS and samples spiked with 500 µg Se L−1 selenomethionine and selenocystine, respectively. Conditions are shown in legend to Fig. 1. |
Semi-quantitative determinations of the selenomethionine and selenocystine/Se-methylselenocysteine concentrations in the extract showed that 8.0 ± 0.3% and 17 ± 0.5% (n = 3) of the selenium was associated with compounds that co-migrated with selenomethionine and selenocystine/Se-methylselenosysteine, respectively. This agrees with results from analysis of the same brand of nutritional supplement by reversed-phase chromatography, which showed that selenomethionine and selenocystine accounted for 13% and 14%, respectively, of the total selenium after hot water extraction.7 In addition, the same authors identified minor amounts of trimethylselenonium and selenoethionine in the extract. This difference in presence and distribution of selenium compounds in the extract could be due to batch variations and variations in sample preparation and the resolving power of the methods.
Repeatability [RSD (%)]a (n = 3) | Detection limit | ||||
---|---|---|---|---|---|
Migration time | Peak height | Peak area | Relative/µg Se L−1b | Absolute/fg Sec | |
a Precision expressed as percent relative standard deviations was calculated on the basis of results from analysis of three yeast based nutritional supplement tablets from the same batch. b The concentrations of seleno-amino acids in the extract were determined by internal standardisation with trimethylselenonium and relative detection limits were calculated as concentrations that will give signals equivalent to three times the peak-to-peak noise of the baseline. c Absolute detection limits correspond to an injected sample volume of 16.7 nL. | |||||
Selenomethionine/selenoethionine | 0.52 | 2.9 | 3.4 | 10 | 170 |
Selenocystine/Se-methylselenocysteine | 0.52 | 1.4 | 5.3 | 15 | 250 |
Relative detection limits expressed as the concentration that would give a signal equivalent to three times the peak-to-peak noise of the base line were better than 15 µg Se L−1, corresponding to absolute detection limits less than 250 fg of Se. The relative detection limits in this study were comparable to or better than detection limits for seleno-amino acid standards when analysed by spray chamber based CE-ICP-MS interfaces.10,24 However, relative LODs were 20 times higher than our previously reported detection limits for aqueous selenium standards in CE-ICP-MS using a direct injection nebulizer interface.17 The difference is mainly due to an eight-fold reduction of the injected sample volume. Besides, the background was high at m/z 82 when poly(vinyl sulfonate) was added to the buffer and sheath liquid, which could be due to an isobaric 34S16O3 interference.
The relative detection limits were generally 10–100 times higher with CE-ICP-MS compared with HPLC-ICP-MS due to the small sample volume injected in CE. However, many of the compounds in the aqueous yeast extracts were present in concentrations far beyond the limit of detection of CE-ICP-MS.
Absolute detection limits for CE-ICP-MS in this study were 10 times lower than data reported for HPLC-ICP-MS7,25,26 due to the high efficiency of 620000 theoretical plates with the CE-ICP-MS system.
In order to determine if some of the compounds detected in the cold-water extracts were seleno-amino acid containing proteins or polypeptides, the aqueous fraction was incubated with concentrated hydrochloric acid at 110 °C for 24 h. This procedure hydrolyses proteins to free amino acids and has previously been used to hydrolyse selenium containing plasma proteins27,28 and to extract selenium compounds from yeast samples.29 Electropherograms from analysis of the hydrolysate showed that 45% of the total selenium migrated within a single peak (Fig. 3). Besides, several of the compounds that were detected in the cold-water extract had disappeared and some new compounds had formed. No selenium compounds were observed when the capillary was flushed with buffer after the analysis was completed and comparisons of integration results revealed that all of the selenium in the aqueous extract was recovered after acidic hydrolysis. However, the conditions used to hydrolyse the sample were rather harsh and some of the compounds formed could be due to decomposition by other reaction paths than hydrolysis. Especially, selenocysteine is sensitive to decomposition under the conditions used for acidic hydrolysis, and is often carboxymethylated prior to hydrolysis.30–33
Fig. 3 Expanded electropherograms from analysis of a hydrolysed aqueous extract of yeast and samples spiked with 500 µg Se L−1 trimethylselenonium, selenomethionine, selenocystine and Se-methylselenomethionine, respectively. Conditions are shown in the legend to Fig. 1. |
Thus, a large fraction of the selenium compounds in the cold water extracted yeast sample were predominantly converted into a compound that co-migrated with selenomethionine, suggesting that these compounds might be polypeptides or low-molecular proteins with unspecifically incorporated selenomethionyl residues.
In conclusion, we have demonstrated that capillary electrophoresis in poly(vinyl sulfonate) coated capillaries with online ICP-MS detection is a promising technique to give a finger-print of selenized yeast preparations. More than 20 selenium compounds were separated in aqueous extracts of yeast based nutritional supplement tablets within 13 min. The system had an efficiency of 620000 theoretical plates and detection limits for seleno-amino acids were in the low µg Se L−1 range.
This journal is © The Royal Society of Chemistry 2004 |