Characterization of the selenocysteine-containing metabolome in selenium-rich yeast

Part 1. Identification of new species by multi-dimensional liquid chromatography with parallel ICP-MS and electrospray Q-TOFMS/MS detection

Abstract

Size-exclusion chromatography (SEC)–strong anion-exchange (SAX) HPLC fractionation of selenium species from the aqueous extract of a selenised yeast sample was optimised under the strict selenium mass balance control by ICP-MS. The SAX HPLC-ICP-MS chromatogram of the most intense SEC fraction produced seven peaks. They were all successfully identified by reversed phase (RP) nanoHPLC-electrospray Q-TOFMS/MS. Eight Se-compounds (derivatives of glutathione) were identified: six of them have not been reported previously. Six of the identified compounds contained selenocysteine (28% of the water-soluble selenium), stabilised by either Se–S or Se–Se bridges. The extensive MS/MS data presented are potentially useful for the optimization of direct LC-ESIMS/MS analyses in the selected (SRM) or multiple (MRM) reaction monitoring modes for the purpose of the authenticity and quality control of Se-rich yeast supplements.

---

Introduction

The putative cancer preventive properties of selenised yeast, reported by Clark et al.,¹ spurred research towards the characterisation of Se-rich yeast in terms of selenium speciation. Selenomethionine (SeMet), accounting for 60–80% of Se and incorporated in the proteins, is by far the most common product of selenite metabolism during Se yeast production.^2,3 The question of the speciation of the remaining selenium is, however, haunting researchers and regulatory bodies.

The composition of the selenium metabolome is characteristic of a yeast strain and fermentation parameters and can be a precious fingerprint of the identity of a particular product and of the reproducibility of the production process.^4,5 The metabolome may contain a particular species showing either a higher therapeutic activity or a higher toxicity. Finally, the knowledge of the full Se-metabolome and pathways of its evolution is essential to reach the desired Se speciation in manufactured Se-rich yeast.⁶ The complete characterization of the Se-metabolome has been the goal of a number of studies, recently reviewed;^2,7 their success, however, has largely been limited.

With regard to the unambiguous identification of Se-species, electrospray MS/MS, proposed, for this purpose, first in 1999 by Casiot et al.,⁸ still seems to be the most suitable analytical technique.^9,10 Two approaches have been proposed. The potentially most attractive one is based on the direct analysis of a Se yeast extract by reversed-phase HPLC-ESI-TOFMS.^11,12 However, using the state-of-the-art instrumentation, Goenaga Infante et al. were able to assign only the two most intense compounds out of a large number of Se species detected by parallel ICP-MS.¹³

The limited success of the “shoot-and-see” approach is due to the ESI-TOFMS response being critically dependent on the purity of the to-be-detected species at the moment of its arrival at the source. The sensitivity is optimum when the analyte molecule arrives at the source unaccompanied by any other easier ionisable species. The sample complexity obviously makes it hard to achieve in a single chromatographic run, which requires multidimensional chromatographic purification beforehand.¹⁴ Two-dimensional size-exclusion (SEC)-reversed phase (RP)-LC was first proposed by McSheehy et al. in the off-line mode.¹⁵ This approach was then refined by building in an additional anion-exchange (SAX) purification step into the procedure, resulting in a 3D (SEC-SAX-RP-HPLC) approach.¹⁶ The RP step was then downscaled to nanoLC and carried out on-line in order to accelerate the analysis and to improve the purity of the analyte species at the moment of its detection.^17,18 The separations were carried out without virtually any mass balance control, the identification of several detected compounds was often uncompleted and the overall outcome of these works can account for the identity of 25% only of the non-protein selenium metabolome.

SEC is clearly the technique of choice for a preliminary fractionation of the yeast aqueous extract. Under optimized conditions the chromatographic recovery reaches 98%.¹⁹ It allows the fractionation of the water-soluble selenium into a protein fraction that could be characterized^20,21 and four metabolite fractions. Only one of the latter, non-specifically retained on the SEC column owing to the presence of adenosyl groups, has been studied.^18,22

The objective of this research was the characterisation of the most intense (accounting for ca. 30% of Se) metabolite fraction in the aqueous extract. For this purpose the second dimension separation was carefully optimised under the strict control of the mass balance by ICP-MS. It was expected that the purity of the Se-compounds reached after 2D separation would be sufficient for the acquisition of the complete ESIMS/MS data for all the compounds using the nanoHPLC sample introduction as it was recently demonstrated for the standardless identification of selenocystathionine and its derivatives in monkeypot nuts, Lecythis minor.²³

Experimental

Reagents and standards

All the chromatographic eluents and reagents were purchased from the Sigma group (Sigma–Aldrich–Fluka–Riedel-de Haën, St. Quentin Fallavier, France). De-ionised water (18.2 MΩ cm; Milli-Q Millipore, Guyancourt, France) was used throughout.

Sample

A selenised yeast sample, Sel-Plex (Alltech, Nicholasville, KY, USA) containing 2.0 ± 0.1 g Se kg^–1, corresponding to a yeast strain Saccharomyces cerevisiae CNCM I-3060, batch ES-453, was used.

Instrumentation

The normal bore HPLC-ICP-MS coupling was achieved by using an Agilent 1100 HPLC system (Agilent Technologies, Palo Alto, CA) connected to an Elan 6000 ICP-MS (PE-SCIEX, Ontario, Canada) for element-specific detection on ³⁴S, ⁷⁷Se, ⁸²Se, and ¹⁰³Rh.

For nanoflow HPLC an Ultimate nanoHPLC system (LC Packings, Amsterdam, The Netherlands) consisting of a binary nanoflow pump and a He-degasser was used. The sample was introduced by means of a 0.3 µl loop of a Model CN2 injection valve (Valco Instruments, Houston, TX, USA) serving as a manual injection port. All connections were made of fused silica capillaries (i.d. 20 µm, Polymicro Technologies, Phoenix, AZ).

For the nanoHPLC-ESI-MS experiments, the nanoHPLC system was connected to a hybrid quadrupole/time-of-flight mass spectrometer Applied Biosystems QSTAR XL (Applied Biosystems, Foster City, CA, USA) used in either full-scan TOF mode with a 1 s integration time or in product ion (MS/MS) mode. The nanoHPLC-ESI-MS coupling was carried out via a nanospray interface (Applied Biosystems), operating in the positive ion mode. The optimum settings were: nanospray needle voltage, 2000 V; declustering potential, 65 eV; focusing potential, 265 eV; collision gas, 0.35 bar N₂. TOF-MS mass spectra were recorded across the range m/z 100–2000. Q-TOF-MS/MS experiments were carried out in the range m/z 50–1000. MS/MS experiments were accomplished with collision energies automatically set by the software. The data recorded were processed with Applied Biosystems/MDS-SCIEX Analyst QS software (Frankfurt, Germany). The instrument was mass calibrated with a mixture of lysine ([M + H]⁺: 147.1128, C₆H₁₅N₂O₂⁺) and cystathionine ([M + H]⁺: 223.0747, C₇H₁₅N₂O₄S⁺).

For total digestion of samples a Multiwave 3000 microwave digestion system (Anton Paar, Courtaboeuf, France) was used.

Procedures

Determination of total selenium. A sample (or aliquots to analyse) was mixed with 4.0 ml of concentrated HNO₃ and 4.0 ml of H₂O₂ in digestion tubes. The temperature was raised to 210 °C within 15 min and held for 20 min. The total Se concentration was determined (Elan 6000) using the ⁷⁷Se and ⁸²Se isotopes by the method of standard additions, using Rh as an internal standard.

Water extraction and size exclusion chromatography (SEC). 0.3 g of selenised yeast was extracted with 10 ml of water in an ultrasonic bath for 1 h. The supernatant was decanted, the residue was re-suspended in water and the extraction was repeated twice. The extracts were centrifuged at 700g for 10 min, pooled, lyophilised and stored under argon in the freezer at –20 °C. Before analysis, the extract was dissolved in 1.5 ml of 0.01 M ammonium acetate (pH 9.5 adjusted with NH₃AQ). The solution was centrifuged at 5800g for 10 min and was then fractionated by SEC. A 10 µl aliquot was retained for the determination of the column recovery.

The protocol of preparative scale SEC, based on the use of a HiLoad 26/60 Superdex 30 pg column (GE Healthcare–Amersham Biosciences, Uppsala, Sweden) was described elsewhere.¹⁹ The fractions selected for further analyses were pooled, frozen and lyophilised.

Anion exchange chromatography (SAX). A PRP-X100 SAX column (250 mm × 4.1 mm × 10 µm; Hamilton, Reno, NV) fitted with a matching guard column filled with the identical phase was used. Gradient elution was made with ammonium acetate (buffer A, 25 mM; buffer B, 250 mM; pH 5.5) delivered at 1.5 ml min^–1. The program was: 0–4 min 100% A, 4–40 min up to 100% B, 40–60 min 100% B. The sample was dissolved in 0.5 ml of buffer A. The injection volume was either 10 µl (for mapping purposes and column recovery determination) or 150 µl (for fraction collection). Fractions were collected every 30 s and 20 µl aliquots were analysed from each fraction to construct the chromatogram. The Se-containing fractions corresponding to the peaks were pooled, frozen and lyophilised.

NanoRPLC-ESI/MS analysis. Nano-scale RPLC-ESI/MS analysis of the purified Se-containing fractions were carried out with an Alltech Ultra IBD nanoLC column (150 mm × 0.075 mm × 5 µm, 100 Å). The compounds were isocratically eluted with 0.05% (v/v) TFA in 50% (v/v) acetonitrile at 0.4 µl min^–1. The sample was dissolved in 0.05% (v/v) TFA in 33% (v/v) acetonitrile. An aliquot of 1.7 µl was necessary to fill the 0.3 µl loop.

Results and discussion

Fractionation of the Se-yeast water extract by SEC

The aqueous extract was found to contain 13 ± 1% of selenium originally present in the sample. The extract was first submitted to preparative scale SEC. Compared with a previous work on this sample providing 98% column recovery of Se,¹⁹ a column recovery of 99 ± 3% was achieved this time owing to the detection of an additional late-eluting peak between 430–450 ml. As was shown in Fig. 1, the elution resulted in several Se peaks with the most abundant one between 195–220 ml. This fraction was found to contain 30 ± 2% of total selenium injected on the column and was collected for further analyses. The other intense peak at 300 ml was extensively characterized elsewhere.^18,22

Fig. 1 Preparative scale SEC-ICP-MS profile of the selenised yeast water extract. Isotopes of ⁸²Se and ³⁴S were monitored. Vertical lines indicate the most abundant fraction collected for further analyses. The ³⁴S signal was intentionally offset for the ease of presentation. For details, see text.

The chromatogram of sulfur is noisier because of the approximately 100-fold higher detection limit for ³⁴S than for ⁸²Se. However, the S elution profile shows clearly that the targeted fraction contains the majority of sulfur as well.

Optimization of the 2^nd dimension of the clean-up process, SAX chromatography

Preliminary experiments by reversed-phase (RP) HPLC-ICP-MS showed that the Se-compounds of interest were not adequately retained under the usual de-salting conditions. Most of selenium eluted in the void volume of the C₈ and C₁₈ columns even when an eluent containing as little as 2% (v/v) methanol with 0.05% (v/v) TFA was used. Therefore, a separation mechanism more appropriate for highly polar compounds was investigated. Strong anion-exchange HPLC was reported to allow 100% recovery of Se from enzymatic extracts.³ The optimisation was carried out in this work with the purpose of obtaining the maximum number of peaks in the chromatogram while monitoring the recovery of selenium by ICP-MS. The optimized parameters included pH (in the 4.5–7.0 range), flow rate (1.0 and 1.5 ml min^–1) and buffer concentration (buffer A: 10–25 mM NH₄Ac, buffer B: 100–300 mM NH₄Ac). Figures in the Electronic Supplementary Information† show outcomes of a given series of optimisations. In brief, the use of eluents with more acidic pH, higher buffer concentration and higher flow rate ameliorated the recovery of Se but decreased the quality of separation.

The chromatogram obtained under the optimum conditions is shown in Fig. 2. The column recovery was 96 ± 2%. Seven peaks could be detected of which six were well separated. A minor compound seems to co-elute with the major one within Peak 2. These two peaks could be separated but only at pH ≥ 7.0 and with a flow rate of 1.0 ml min^–1, at the expense of the column recovery dropping below 90% because the compounds eluting as peaks no. 6 and no. 7 then stacked on the column and could not be eluted even with 0.3 M ammonium acetate. Fractions corresponding to all the seven peaks were collected and analysed for the total Se content. Fraction 1 contained a significant amount of sulfur-compounds. Note that the retention of the three most common selenium standards, SeMet, Se(IV) and Se(VI), is by far weaker than that of any of the Se compounds present in the sample. This demonstrates a serious limitation of the optimisation of the separation conditions using the available standards, which is usually not reported in the literature. This also emphasizes the utmost importance of the determination of column recovery during optimization steps.

Fig. 2 Optimised SAX-ICP-MS chromatogram of the collected SEC fraction. The arrows indicate the retention time of three Se-standards. Fractions numbered from 1 to 7 were collected separately for further analyses. The ³⁴S signal was intentionally offset for the sake of clarity of presentation.

Analyses of the SEC-SAX fractions by nanoHPLC-ESI-TOF-MS and -Q-TOFMS/MS

In order to increase the purity of the selenium compounds at the moment of their arrival at the ionization source, nanoRP-HPLC was chosen for sample introduction into ESIMS. This technique was reported in our previous papers to improve the signal-to-noise (S/N) ratio of Se-containing analytes from complex mixtures due to its high sensitivity, possible background correction and the elimination of a number of matrix interferences.^18,23 NanoRP-HPLC-ESI-TOFMS recently allowed the detection of minor Se-species at a concentration of 1 ppm (0.3 ng per injection).²³ In order not to miss any Se-species in this work the injections were carried out at a concentration of at least 10 ppm (ca. 3 ng per injection) of Se in each individual fraction.

Each of the seven fractions was individually analysed with a preliminary full scan ESI-TOFMS acquisition. The full scan spectra were then checked for the presence of Se-containing compounds through the search for the isotopic pattern and mass defect of selenium.^18,23 All the observed Se-species were then subjected to collision induced dissociation (CID) experiments.

SAX-fraction no. 1. The full scan spectrum of this fraction (Fig. 3(a)) is dominated by oxidized glutathione ([M + H]⁺: m/z 613.1, [M + 2H]²⁺: m/z 307.1) and its related in-source fragments ([M + H]⁺: m/z 538.1, 409.1, 355.1, 130.05; [M + 2H]²⁺: m/z 275.6, 269.5). The high abundance of this compound (containing two S atoms) explains the presence of sulfur in the corresponding fractions in SEC and SAX. The search for Se-compounds revealed the presence of a molecule with a monoisotopic mass at m/z 604.07 (see the inset). Its CID fragmentation (Fig. 3(b)) showed several fragments derived from glutathione. The assignment of the Se-containing fragments proved this compound to be a mixed conjugate of either (i) glutathione and γ-glutamoyl-L-selenocysteine or (ii) γ-glutamoyl-selenocysteinylglycine (referred to as Se-glutathione from now on) and γ-glutamoyl-L-cysteine. Especially, the fragment at m/z 167.96 helped the elucidation as it indicates a selenocysteine (SeCys) residue of which the carboxyl group was not incorporated in a peptide bond in one of the conjugates. Fig. 3(c) presents the proposed fragmentation pathway, along with the accurate mass information in Table 1. The detected fragments indicate both types of conjugates were present in the fraction.

Fig. 3 (a) NanoLC-ESI-MS full-scan spectrum of the Se-containing SAX fraction no. 1 (cf.Fig. 2). The inset shows the correct isotopic pattern of selenium at m/z 604.07 ([M + H]⁺). (b) Collision-induced dissociation (CID) spectrum of the analyte at m/z 604.07. (c) Proposed fragmentation pathways of the Se-compound. For all accurate mass information see Table 1.

Table 1 Description of the proposed fragment ions formed during the fragmentation of the conjugate of either glutathione and γ-glutamoyl-selenocysteine or selenoglutathione and γ-glutamoyl-cysteine, m/z 604.0704. For the assignment of the fragment numbers please see Fig. 3; for details, please see text

No.	Elemental composition	Theoretical mass/Th^a	Measured mass/Th^a	Difference (ppm)	Notes on losses and fragments
a Th = Thomson, where 1 Th = 1 m/z. All the indicated losses are related to the intact Se-species.
1	C18H30N5O11SSe^+	604.0822	604.0704	20	(Intact species)
2	C18H27N4O11SSe^+	587.0557	587.0470	15	NH3
3	C17H28N5O9SSe^+	558.0767	558.0685	15	Formic acid
4	C17H25N4O9SSe^+	541.0502	541.0283	40	Formic acid and NH3
5	C13H23N4O8SSe^+	475.0396	475.0141	54	γ-Glu
6	C11H18N3O6SSe^+	400.0076	399.9934	35	γ-Glu and Gly
7	C8H16N3O5SSe^+	345.9976	345.9899	22	Two γ-Glu residues
8	C8H13N2O5SSe^+	328.9705	328.9592	34	Glutathione without the heteroatom
9	C10H18N3O6S^+	308.0911	308.0839	23	SeCys and γ-Glu
10	C8H11N2O4Se^+	278.9879	278.9913	–12	Glutathione and H2O from γ-Glu
11	C6H11N2O3SSe^+	270.9650	270.9677	10	Gly and two γ-Glu residues
12	C5H9N2O3SSe^+	256.9494	256.9431	25	Heteroatom-free Cys and two γ-Glu residues
13	C5H9N2O3Se^+	224.9773	224.9745	12	Cys and two γ-Glu residues
14	C3H6NO2Se^+	167.9558	167.9642	–50	Glutathione and γ-Glu
15	C5H8NO3^+	130.0499	130.0477	17	Detected γ-Glu fragment

---

SAX-fraction no. 2. The investigation of the full scan spectrum of this fraction (Fig. 4(a)) allowed the spotting of two selenised compounds. The more abundant monoselenised one at m/z 661.1 ([M + H]⁺), further identified by the MS/MS spectrum (not shown), was recently reported as a conjugate of Se-glutathione and glutathione.¹³ The less abundant one showed the monoisotopic mass of m/z 326.5131 ([M + 2H]²⁺; see the inset) and the isotopic pattern characteristic of the presence of two Se atoms. Its CID spectrum (Fig. 4(b)) showed analogy with that of the monoselenised compound at m/z 604.07 found in the first SAX fraction. According to the proposed fragmentation pathway (Fig. 4(c)), this compound is the conjugate of selenoglutathione and γ-glutamoyl-L-selenocysteine with a theoretical mass of m/z 652.0267 ([M + H]⁺). Note that the monoselenised doubly charged molecule at m/z 322.54, shown in the inset, is an in-source fragment of the abundant Se-compound of m/z 661.1 due to the neutral loss of NH₃. All the relevant accurate mass information is presented in Table 2.

Fig. 4 (a) NanoLC-ESIMS full-scan spectrum of the Se-containing SAX fraction no. 2 (cf.Fig. 2). The inset shows the correct doubly selenised isotopic pattern at m/z 326.5131 ([M + 2H]²⁺). (b) CID spectrum of the analyte at m/z 326.5131. (c) Proposed fragmentation pathways of the Se-compound. For all accurate mass information see Table 2.

Table 2 Description of the proposed fragment ions formed during the fragmentation of the conjugate of selenoglutathione and γ-glutamoyl-selenocysteine, m/z 326.5131 ([M + 2H]²⁺). For the assignment of the fragment numbers please see Fig. 4; for details, please see text

No.	Elemental composition	Theoretical mass/Th^a	Measured mass/Th^a	Difference (ppm)	Notes on losses and fragments
a Th = Thomson, where 1 Th = 1 m/z. All the indicated losses are related to the intact Se-species.
1	C18H30N5O11Se2^+	652.0267	652.0254	2	(Intact species)
2	C7H8NO3Se2^+	313.8829	313.8872	–14	Two γ-Glu residues, two NH3 groups, one formic acid
3	C5H9N2O3Se2^+	304.8938	304.8860	26	Heteroatom-free Cys and two γ-Glu residues
4	C8H13N2O5Se^+	296.9984	297.0033	–16	Selenoglutathione
5	C8H11N2O4Se^+	278.9879	278.9846	12	Selenoglutathione and H2O from a γ-Glu residue
6	C3H6NO2Se2^+	247.8723	247.8668	22	Heteroatom-free glutathione and γ-Glu
7	C5H9N2O3Se^+	224.9773	224.9791	8	SeCys and two γ-Glu residues
8	C5H8NO3^+	130.0499	130.0457	32	Detected γ-Glu fragment

---

SAX-fraction no. 3. The full scan spectrum of this fraction (Fig. 5(a)) shows a compound with two Se atoms at the monoisotopic mass of m/z 709.0361 ([M + H]⁺, see the insets). Its CID mass spectrum (Fig. 5(b)) showed analogy with that of the monoselenised compound at m/z 661.1 found in the second SAX-fraction.¹³ According to the proposed fragmentation pathway and accurate mass analysis (see Fig. 5(c) and Table 3, respectively), this compound is the conjugate of two Se-glutathione molecules with the theoretical mass of m/z 709.0481 ([M + H]⁺).

Fig. 5 (a) NanoLC-ESI-MS full-scan spectrum of the Se-containing SAX fraction no. 3 (cf.Fig. 2). The inset shows the correct doubly selenised isotopic pattern at m/z 709.0481 ([M + H]⁺). (b) CID spectrum of the analyte at m/z 709.0481. (c) Proposed fragmentation pathways of the Se-compound. For all accurate mass information see Table 3.

Table 3 Description of the proposed fragment ions formed during the fragmentation of the conjugate of two selenoglutathione moieties, m/z 709.0361. For the assignment of the fragment numbers please see Fig. 5; for details, please see text

No.	Elemental composition	Theoretical mass/Th^a	Measured mass/Th^a	Difference (ppm)	Notes on losses and fragments
a Th = Thomson, where 1 Th = 1 m/z. All the indicated losses are related to the intact Se-species.
1	C20H33N6O12Se2^+	709.0481	709.0361	17	(Intact species)
2	C19H28N5O10Se2^+	646.0161	646.0257	–15	Formic acid and NH3
3	C18H28N5O10Se2^+	634.0161	634.0557	–62	Gly
4	C15H26N5O9Se2^+	580.0055	580.0369	–54	γ-Glu
5	C15H23N4O9Se2^+	562.9795	562.9915	–21	γ-Glu and NH3
6	C13H21N4O7Se2^+	504.9735	504.9928	–38	γ-Glu and Gly
7	C10H19N4O6Se2^+	450.9635	450.9810	–39	Two γ-Glu residues
8	C10H16N3O6Se2^+	433.9364	433.9558	–45	Two γ-Glu residues and NH3
9	C10H13N2O6Se2^+	416.9099	416.9272	–41	Two γ-Glu residues and two NH3 groups
10	C8H11N2O4Se2^+	358.9044	358.9250	–57	Two γ-Glu residues, Gly, NH3
11	C5H9N2O3Se2^+	304.8938	304.9132	–64	Heteroatom-free glutathione and γ-Glu
12	C9H11N2O4Se^+	290.9879	291.0029	–52	Selenoglutathione, formic acid and NH3
13	C8H11N2O4Se^+	278.9879	279.0040	–58	Selenoglutathione and Gly
14	C10H13N2O6^+	257.0768	257.0911	–56	Selenoglutathione with the diselenide bridge and NH3
15	C5H9N2O3Se^+	224.9773	224.9812	–17	Selenoglutathione and γ-Glu
16	C5H8NO3^+	130.0499	130.0559	–46	Detected γ-Glu fragment
17	C4H6NO^+	84.0444	84.0498	–64	Detected γ-Glu fragment with a loss of formic acid
18	C2H6NO2^+	76.0393	76.0449	–74	Detected Gly fragment

---

SAX-fraction no. 4. The full scan spectrum of this fraction (Fig. 6(a)) shows a Se-species at m/z 563.0503 ([M + H]⁺, see the inset) that was already reported by McSheehy et al.²⁴ and Goenega-Infante et al.,¹³ without structure assignment. As a useful piece of information, it was confirmed to contain a S–Se bond between a glutathione molecule and a Se-containing residue.²⁴ The structure could be elucidated here owing to the high resolution CID mass spectrum (Fig. 6(b)). The presence of a fragment at m/z 167.96 and at m/z 121.95 indicates a SeCys residue with a free carboxyl group. There is a characteristic water loss from a Se-containing fragment at m/z 257.98 that has not been observed in the other glutathione-derived structures. Taking into account the accurate mass information tabulated in Table 4, the structure of a conjugated glutathione and 2,3-dihydroxy-propionyl-L-selenocysteine could be proposed, which is presented with the possible fragmentation pathways in Fig. 6(c). This unusual residue could be possibly formed by the deamination and oxidation of a Ser moiety.

Fig. 6 (a) NanoLC-ESI-MS full-scan spectrum of the Se-containing SAX fraction no. 4 (cf.Fig. 2). The inset shows the correct isotopic pattern of selenium at m/z 563.0503 ([M + H]⁺). (b) CID spectrum of the analyte at m/z 563.0503. (c) Proposed fragmentation pathways of the Se-compound. For all accurate mass information see Table 4.

Table 4 Description of the proposed fragment ions formed during the fragmentation of the conjugate of glutathione and 2,3-dihydroxypropionyl-SeCys (or rather, modified Ser–SeCys), m/z 563.0503. For the assignment of the fragment numbers please see Fig. 6; for details, please see text

No.	Elemental composition	Theoretical mass/Th^a	Measured mass/Th^a	Difference (ppm)	Notes on losses and fragments
a Th = Thomson, where 1 Th = 1 m/z. All the indicated losses are related to the intact Se-species.
1	C16H27N4O11SSe^+	563.0557	563.0503	10	(Intact species)
2	C14H22N3O9SSe^+	488.0136	488.0053	17	Gly
3	C11H20N3O8SSe^+	434.0131	434.0030	23	γ-Glu
4	C8H16N3O5SSe^+	345.9976	346.0005	–8	γ-Glu and modified Ser
5	C8H15N2O5SSe^+	330.9861	330.9845	5	γ-Glu, Gly and CO
6	C8H13N2O4SSe^+	312.9756	312.9768	–4	Gly, heteroatom-free Cys and modified Ser
7	C6H12NO5Se^+	257.9875	257.9873	1	Glutathione
8	C6H10NO5Se^+	255.9719	255.9701	7	Glutathione
9	C6H10NO4Se^+	239.9775	239.9769	3	Glutathione and H2O
10	C5H8NO3Se^+	209.9664	209.9659	2	Glutathione and formic acid
11	C3H6NO2Se^+	167.9558	167.9583	–15	Glutathione and modified Ser
12	C5H8NO3^+	130.0499	130.0506	–5	Detected γ-Glu fragment
13	C2H4NSe^+	121.9503	121.9492	9	Glutathione, modified Ser and formic acid

---

SAX-fraction no. 5. The full scan spectrum of this fraction (Fig. 7(a)) reveals the presence of a compound with two Se atoms at m/z 610.9855 ([M + H]⁺, see the inset). The corresponding MS/MS fragmentation data (Fig. 7(b)) shows analogy with the previously described monoselenised compound at m/z 563.0503. It was also observed that in the analysed SEC fraction definite couples of monoselenised and doubly selenised analogues are following each other in the elution order determined by the SAX gradient (cf., m/z 604 and m/z 652, as well as m/z 661 and m/z 709). Therefore, a structure analogue with the molecule at m/z 563.0503 in which the sulfur atom was replaced by selenium could be proposed with the possible fragmentation pathways in Fig. 7(c) and accurate mass data in Table 5.

Fig. 7 (a) NanoLC-ESI-MS full-scan spectrum of the Se-containing SAX fraction no. 5 (cf.Fig. 2(d)). The inset shows the correct doubly selenised isotopic pattern at m/z 610.9855 ([M + H]⁺). (b) CID spectrum of the analyte at m/z 610.9855. (c) Proposed fragmentation pathways of the Se-compound. For all accurate mass information see Table 5.

Table 5 Description of the proposed fragment ions formed during the fragmentation of the conjugate of selenoglutathione and 2,3-dihydroxypropionyl-SeCys (or rather, modified Ser–SeCys), m/z 610.9855. For the assignment of the fragment numbers please see Fig. 7; for details, please see text

No.	Elemental composition	Theoretical mass/Th^a	Measured mass/Th^a	Difference (ppm)	Notes on losses and fragments
a Th = Thomson, where 1 Th = 1 m/z. All the indicated losses are related to the intact Se-species.
1	C16H27N4O11Se2^+	611.0001	610.9855	24	(Intact species)
2	C14H22N3O9Se2^+	535.9681	535.9697	–3	Gly
3	C8H15N2O5Se2^+	378.9306	378.9250	15	γ-Glu, Gly and CO
4	C8H13N2O4Se2^+	360.9203	360.9188	4	Gly, heteroatom-free Cys and modified Ser
5	C6H10NO5Se2^+	335.8884	335.8864	6	Heteroatom-free glutathione
6	C6H8NO4Se2^+	317.8778	317.8772	2	Heteroatom-free glutathione and H2O
7	C8H11N2O4Se^+	278.9879	278.9871	3	SeCys, modified Ser and Gly
8	C6H10NO5Se^+	255.9719	255.9786	–26	Selenoglutathione
9	C3H6NO2Se2^+	247.8723	247.8698	10	Heteroatom-free glutathione and modified Ser
10	C7H9N2O2Se^+	232.9824	232.9809	6	SeCys, modified Ser, Gly and formic acid
11	C3H6NO2Se^+	167.9558	167.9521	22	Selenoglutathione and modified Ser
12	C5H8NO3^+	130.0499	130.0511	–9	Detected γ-Glu fragment

---

SAX-fraction no. 6. The full scan spectrum of this fraction (Fig. 8(a)) presents a monoselenised compound with the monoisotopic mass of m/z 577.0577 ([M + H]⁺, see the inset). The corresponding MS/MS fragmentation data (Fig. 8(b)) shows a characteristic m/z +14 homology with those of the previously described monoselenised compound at m/z 563.0503. Indeed, the fragments at m/z 181.97 and m/z 135.96 indicate a selenohomocysteine residue with a free carboxyl group in the molecule.¹⁸ According to this and the accurate mass analysis (see Table 6), Fig. 8(c) presents a proposed structure along with a possible fragmentation pathway. That is, a glutathione is conjugated with a 2,3-dihydroxypropionyl-L-selenohomocysteine moiety.

Fig. 8 (a) NanoLC-ESI-MS full-scan spectrum of the Se-containing SAX fraction no. 6 (cf.Fig. 2). The inset shows the correct isotopic pattern of selenium at m/z 577.0577 ([M + H]⁺). (b) CID spectrum of the analyte at m/z 577.0577. (c) Proposed fragmentation pathways of the Se-compound. For all accurate mass information see Table 6.

Table 6 Description of the proposed fragment ions formed during the fragmentation of the conjugate of glutathione and 2,3-dihydroxypropionyl-Se-homocysteine (or rather, modified Ser–SeHCys), m/z 577.0577. For the assignment of the fragment numbers please see Fig. 8; for details, please see text

No.	Elemental composition	Theoretical mass/Th^a	Measured mass/Th^a	Difference (ppm)	Notes on losses and fragments
a Th = Thomson, where 1 Th = 1 m/z. All the indicated losses are related to the intact Se-species.
1	C17H29N4O11SSe^+	577.0613	577.0577	6	(Intact species)
2	C15H24N3O9SSe^+	502.0393	502.0259	27	Gly
3	C14H22N3O7SSe^+	456.0338	456.0235	23	Gly and formic acid
4	C12H22N3O8SSe^+	448.0197	448.0139	13	γ-Glu
5	C9H17N2O5SSe^+	345.0018	345.0031	–4	γ-Glu, Gly and CO
6	C8H13N2O4SSe^+	312.9756	312.9765	–3	Gly, modified Ser and heteroatom-free homocysteine
7	C8H13N2O3SSe^+	296.9807	296.9811	–1	γ-Glu, NH3, formic acid and modified Ser
8	C7H14NO5Se^+	272.0032	272.0016	6	Glutathione
9	C7H12NO5Se^+	269.9875	269.9847	10	Glutathione
10	C7H12NO4Se^+	253.9926	253.9892	13	Glutathione and H2O
11	C6H10NO3Se^+	223.9802	223.9737	29	Glutathione and formic acid
12	C4H8NO2Se^+	181.9715	181.9720	–3	Glutathione and modified Ser
13	C3H6NSe^+	135.9665	135.9643	16	Glutathione, formic acid and modified Ser

---

SAX-fraction no. 7. The full scan spectrum of this fraction (not shown) contained a monoselenised compound with m/z 693.0581 ([M + H]⁺; [M + 2H]²⁺ = 347.0380). Its mass, along with the MS/MS fragmentation spectrum (not shown) confirms that this molecule is a selenotrisulfide conjugate, selenodiglutathione, reported in a selenised yeast sample by Lindemann and Hintelmann.¹⁷ This compound is unique among the Se-species in the SEC fraction analysed in the sense that it contains no selenoamino acid residue, the Se atom being covalently bound to two cysteinyl S atoms of two glutathione molecules.

Note that a search for the analogue with two Se atoms of the compound with the selenohomocysteine residue (cf. SAX-fraction no. 6) expected at m/z 625.0158 was unsuccessful.

The structures of the SeCys-containing Se-species at m/z 604.07 (Fraction 1), m/z 652.02 (Fraction 2), m/z 709.03 (Fraction 3), m/z 563.05 (Fraction 4), m/z 610.98 (Fraction 5) and the selenohomocysteine-containing species at m/z 577.05 (Fraction 6) have never been presented before.

As tabulated for the analytes and their fragments, the mass accuracy of the assignments varied between 1 and 74 ppm. The manufacturer’s specification of this QTOF instrument is around 5 ppm in ideal conditions; however, several factors can negatively affect this accuracy. First, nanoliquid chromatography is not a robust technique. It suffers from the fluctuation of ionisation which is remarkably influenced by the matrix compounds that are missing in the case of external calibration. The ionisation and detected intensity of a given analyte is changing during nanoLC runs, even in direct infusion mode, because of the continuous degradation and spray instability of nanospray tips.²⁵ Further, as the mass accuracy is a function of signal intensity in the case of such an orthogonal TOF instrument,²⁶ poorer S/N ratios cause poorer mass accuracy. Software or hardware-based tools can be addressed to overcome this problem.²⁷ The usually poor S/N ratios of selenium species can be the reason for the poor accuracy reported during the identification of selenodiglutathione with nanoelectrospray-QTOF (up to 131 ppm mass error in the MS/MS mode)¹⁷ and the detection of Se-adenosylhomocysteine with direct infusion-QTOF (up to 125 ppm mass error in the MS/MS mode).²⁸

Tentative quantification of SeCys-containing Se-species

From the eight compounds identified in the seven SAX fractions, six contain selenium in the form of SeCys. Therefore, aliquots were taken in triplicates from the corresponding fractions and they were analysed for total Se content. The five fractions contained 1.17, 4.73, 1.98, 1.66, and 1.42 µg of Se (RSD < 5%), respectively, that represented 95% of selenium injected onto the SAX column. The two last fractions containing no compounds with a SeCys residue accounted for 0.40 and 0.17 µg, respectively. The presence of minor and undiscovered Se-species in the analysed fractions cannot be totally excluded; however, their individual concentrations would have been at least one order of magnitude lower, due to the sensitivity of the applied nanoLC-ESI-MS configuration.²³ Therefore, taking into account methodological losses (for column recovery determination and mapping) and the extraction efficiency, the amount of SeCys in the analysed yeast aqueous fraction was estimated to be 40 µg Se g^–1 , i.e., ca. 2% of total Se.

Conclusions

The study demonstrates a successful multidimensional method for both the qualitative and quantitative characterisation of a SeCys-rich aqueous fraction of selenised yeast; the task for which the methods reported in the literature largely failed in the past. Also, the success of the approach is clearly due to the use of electrospray-TOFMS/MS, the key to it being the careful optimisation of the sample preparation with strict mass balance control by ICP-MS. The extensive MS/MS data presented are potentially useful for the optimization of direct LC-ESI-MS/MS analyses in the multiple reaction monitoring (MRM) or selected reaction monitoring (SRM) mode²⁹ for the authenticity and quality control of Se-rich yeast supplements.

Acknowledgements

M.D. acknowledges a Marie Curie fellowship (MEIF-CT-2003-501297). The authors thank Dr Hugues Preud’homme for his help with the HPLC-ESI-QTOF-MS experiments and Professor Judit Kosáry for structure assignments. The financial support of the MS platform via the CPERs 20.6 and 21.6 by the Region of Aquitaine is acknowledged.

References

1. L. C. Clark, G. F. Combs, 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, H. K. Park, B. B. Sanders, C. L. Smith and J. R. Taylor, JAMA, J. Am. Med. Assoc., 1996, 276, 1957–1963.
2. G. N. Schrauzer, Pure Appl. Chem., 2006, 78, 105–109.
3. A. Polatajko, B. Banas, J. R. Encinar and J. Szpunar, Anal. Bioanal. Chem., 2005, 381, 844–849.
4. A. Polatajko, M. Sliwka-Kaszynska, M. Dernovics, R. Ruzik, J. R. Encinar and J. Szpunar, J. Anal. At. Spectrom., 2004, 19, 114–120.
5. G. N. Schrauzer, J. Am. Coll. Nutr., 2001, 20, 1–4.
6. H. Goenaga Infante, R. Hearn and T. Catterick, Anal. Bioanal. Chem., 2005, 382, 957–967.
7. A. Polatajko, N. Jakubowski and J. Szpunar, J. Anal. At. Spectrom., 2006, 21, 639–654.
8. C. Casiot, V. Vacchina, H. Chassaigne, J. Szpunar, M. Potin-Gautier and R. Lobinski, Anal. Commun., 1999, 36, 77–80.
9. K. A. Francesconi and M. Sperling, Analyst, 2005, 130, 998–1001.
10. S. McSheehy and Z. Mester, Trends Anal. Chem., 2003, 22, 210–224.
11. M. Montes-Bayón, D. L. LeDuc, N. Terry and J. A. Caruso, J. Anal. At. Spectrom., 2002, 17, 872–879.
12. A. P. Vonderheide, S. Mounicou, J. Meija, H. F. Henry, J. A. Caruso and J. R. Shann, Analyst, 2006, 131, 33–40.
13. H. Goenaga Infante, G. O'Connor, M. Rayman, R. Hearn and K. Cook, J. Anal. At. Spectrom., 2006, 21, 1256–1263.
14. J. Szpunar and R. Lobinski, Anal. Bioanal. Chem., 2002, 373, 404–411.
15. S. McSheehy, W. Yang, F. Pannier, J. Szpunar, R. Lobinski, J. Auger and M. Potin-Gautier, Anal. Chim. Acta, 2000, 421, 147–153.
16. S. McSheehy, F. Pannier, J. Szpunar, M. Potin-Gautier and R. Lobinski, Analyst, 2002, 127, 223–229.
17. T. Lindemann and H. Hintelmann, Anal. Chem., 2002, 74, 4602–4610.
18. J. F. García-Reyes, M. Dernovics, P. Giusti and R. Lobinski, J. Anal. At. Spectrom., 2006, 21, 655–665.
19. J. R. Encinar, M. Sliwka-Kaszynska, A. Polatajko, V. Vacchina and J. Szpunar, Anal. Chim. Acta, 2003, 500, 171–183.
20. J. R. Encinar, L. Ouerdane, W. Buchmann, J. Tortajada, R. Lobinski and J. Szpunar, Anal. Chem., 2003, 75, 3765–3774.
21. P. Giusti, D. Schaumlöffel, H. Preud'homme, J. Szpunar and R. Lobinski, J. Anal. At. Spectrom., 2006, 21, 26–32.
22. J. F. García-Reyes, M. Dernovics, P. Ortega-Barrales, A. R. Fernandez-Alba and A. Molina-Diaz, J. Anal. At. Spectrom., 2007, 22, 947–959.
23. M. Dernovics, T. Garcia-Barrera, K. Bierla, H. Preud'homme and R. Lobinski, Analyst, 2007, 132, 439–449.
24. S. McSheehy, P. Pohl, J. Szpunar, M. Potin-Gautier and R. Lobinski, J. Anal. At. Spectrom., 2001, 16, 68–73.
25. B. A. Sinnaeve and J. F. V. Bocxlaer, J. Chromatogr., A, 2004, 1058, 113–119.
26. K. F. Blom, Anal. Chem., 2001, 73, 715–719.
27. H. C. Köfeler and M. L. Gross, J. Am. Soc. Mass Spectrom., 2005, 16, 406–408.
28. S. McSheehy, J. Szpunar, V. Haldys and J. Tortajada, J. Anal. At. Spectrom., 2002, 17, 507–514.
29. S. Letsiou, V. Nischwitz, P. Traar, K. A. Francesconi and S. A. Pergantis, Rapid Commun. Mass Spectrom., 2007, 21, 343–351.

---

Footnote

† Electronic supplementary information (ESI) available: optimisation of the analytical scale SAX separation of the collected SEC fraction monitored with ICP-MS. See DOI: 10.1039/b708294k

---