Identification of selenomethionine in selenized yeast using two-dimensional liquid chromatography-mass spectrometry based proteomic analysis

Abstract

Selenium-enriched yeast has been commonly used as a nutritional supplement. Here we describe a protocol used to investigate the metabolic fate of inorganic selenium in yeast. We provide definitive, mass spectrometry based evidence for the non-specific incorporation of selenomethionine in the yeast proteome involving the replacement of about 30% of all methionine with selenomethionine.

---

It is known that yeast grown in Cr, Zn and Se rich media is able to accumulate enormous quantities of these metallic/metalloid elements. This ability has led to its common use as a nutritional supplement. For Se, an essential trace element, present in a number of mammalian enzymes as selenocysteine (SeCys),¹ supplementation might reduce cancer incidence, aid proper immune function and prevent deficiency syndromes.² The major species of Se in yeast is selenomethionine (SeMet). SeMet is formed from inorganic Se via a route similar to the sulfur metabolic pathway.³ It is commonly believed that the SeMet is then non-specifically incorporated into protein in the place of methionine (Met)² due to the inability of the initiator codon (AUG) to distinguish between the two compounds, previously termed the ‘relaxed’ genetic code.⁴ However, there is only indirect evidences available to support this incorporation theory. This evidence relies on co-migration of radiolabeled ⁷⁵Se incorporated in yeast with various protein fractions in gel electrophoresis.⁵ However, this type of experiment does not probe the nature of the association between proteins and SeMet. Despite the high level of selenium incorporation (0.2%), recent (top down or analytical) approaches using mass spectrometry have yielded limited results with respect to the identification of Se-containing proteins from the whole yeast proteome.^6–8 This can be explained by the fact that the Met residue has a relatively low abundance in the yeast proteome (about 1–2% of all amino acid residues). In addition, we recently reported that about a third of the Met residues are present as SeMet in selenized yeast (this ratio was determined in-house from quantification of Met (0.37 mol kg⁻¹) and SeMet (0.17 mol kg⁻¹) using species-specific isotope dilution.^9–11 Due to this low abundance and the natural distribution of selenium isotopes, the isolation of sufficient quantities of protein or peptide for analysis is a limiting factor in the results obtained. Here we outline an analytical approach based on LC and shotgun proteomics that enables the identification of SeMet in about 30% of the proteins identified.

The challenge of identifying low abundance Se-peptides in a complex matrix is demonstrated in Fig. 1. A comparison of a mass spectrum of a Met-containing peptide (Fig. 1(a)) shows that the most important molecular ion is about 4 times as intense as the equivalent SeMet-containing molecular ion (Fig. 1(b)) due to the ratio of Met to SeMet in the tissue and the effect of isotope distribution of S and Se on the molecular ion pattern. The initial approach adopted to find the Se-modified peptides was to employ conventional data dependent acquisitions (DDA). A tryptic digest of extracted proteins was divided into 7 fractions using preparative size exclusion chromatography (SEC). SEC fractions were concentrated and analyzed by LC MSⁿ with DDA. MS/MS and high-resolution mass spectra acquisitions were triggered for the most intense molecular ion eluting in the LC run and the fragmentation spectra were subjected to a database search. However, a typical chromatogram is very complex and selenopeptides of low abundance relative to other peptides did not usually trigger the DDA. To overcome this problem, we customized our approach, using the DDA and database search to identify Met-containing peptides. The fraction in which a Met-containing peptide was identified was re-analyzed, this time with forced fragmentation of the molecular ion of the Met-containing peptide plus 48 mass units (the difference between the most abundant isotopes of S(32) and Se(80)). This forced fragmentation spectrum was compared with the initial fragmentation spectrum of the Met-containing peptide (e.g., Fig. 1(c) and Fig. 1(d)) and if ‘matching’ spectra were observed, the peptide was confirmed as containing SeMet in the place of Met.

Fig. 1 MS spectra of a Met-containing and SeMet-containing peptide observed in a SEC fraction of yeast. (a) Mass spectrum of the singly charged ion of the peptide DLTQFPAFVTPMGK and (b) the triggered fragmentation spectrum of the most important molecular ion at m/z 1551.9. (c) Mass spectrum of the singly charged ion of the peptide DLTQFPAFVTPXGK (where X denotes SeMet) and (d) the forced fragmentation spectra of the most important molecular ion at m/z 1599.9. It can be observed that the mass difference between the most intense molecular ions of the Met-containing peptide in (a) and the SeMet-containing peptide in (b) is 48 mass units (the difference between the most important isotopes of S and Se). From the comparison of fragmentation spectra of the Met-containing peptide in (c) and the SeMet-containing peptide in (d) two sets of fragments are apparent: fragments with the same m/z in both spectra, which do not contain a Met or a SeMet residue, and fragments (highlighted in bold) observed in (c) with an m/z which is 48 mass units higher in the respective fragment ions in (d) and which contain a Met and a SeMet residue, respectively.

By comparing each set of spectra in this fashion, the presence of SeMet in the peptide can be confirmed. Additionally, a manual search of the chromatogram for the characteristic selenium isotopic distribution permitted the identification of any peptides that were overlooked or incorrectly identified in the database search. Manually sequencing the fragmentation spectra and running partial sequences through BLAST (NCBI) reduced incidences of false positives. The employed 2D LC approach has limited characterization to the most abundant proteins in the proteome.¹² A list of those proteins in which SeMet was found is shown in Table 1. Peptides identified by an alternative approach, based on 1D gel electrophoresis and identification by LC MSⁿ are also listed in Table 1. The ratios of unmodified to modified peptide were estimated from the integrated molecular ion profiles and reflect the ratio determined in-house by LC-ID-MS. Despite the limited number of proteins, the results indicate that the non-specific incorporation of selenium is similar across the whole proteome.

Table 1 List of proteins identified that contain SeMet

Protein	m/z Met–peptide	m/z SeMet–peptide	Sequence of peptide indentified^a	Res #	Ratio Unmod/Mod^b	Biological processes	Cellular component
a Peptides identified, which correspond to smaller fragments of peptides shown in this table, are not listed. b Ratio of unmodified to modified peptide, estimated from peak area of integrated molecular ion profiles. c Peptides identified by 1D gel-electrophoresis LC MS^n.
ADH1	597.3	645.3	AMGYR	219–224	2.1	Glucose fermentation	Cytosol
	1387.8	1435.8	ANGTTVLVGMPAGAK	262–276	2.0
		717.8 (2+)	ANGTTVLVGMPAGAK^c	262–276	10.0
AHP1	890.4	938.4	MPQTVEW	33–39	1.5	Regulation of cell redox homeostasis, response to metal ion	Cytoplasm
DAK 2	761.2	809.2	MDALVGR	546–552	6.3	Glycerol catabolism, response to stress	Cellular component unknown
ENO1	1684.9	1732.8	SIVPSGASTGVHEALEM	32–48	2.5	Glycolysis, gluconeogenesis	Cytoplasm, phosphopyruvate hydratase complex
	521.3	569.3	WMGK	56–59	2.8
ENO2	1684.9	1732.8	SIVPSGASTGVHEALEM	32–48	2.5	Glycolysis, gluconeogenesis	Phosphopyruvate hydratase complex, soluble fraction
	521.3	569.3	WMGK	56–59	2.8
	1297.3	1245.3	WLTGVELADMY	272–282	5.0
FBA1	1931.9	1979.9	LPWFDGMLEADEAYFK	116–131	2.0	Glycolysis, gluconeogenesis	Cytoplasm, cytosol
PDC1	1665.0	1713.0	IDLTQFPAFVTPMGK	235–249	2.5	Ethanol fermentation, pyruvate metabolism	Cytoplasm, nucleus
	1356.9	1404.9	NIVEFHSDHMK	305–315	1.0
RPL11	1524.7	1572.7	YDPSIGIFGMDFY	115–127	4.5	Protein biosynthesis, ribosomal large subunit assembly and maintenance	Cytosolic large ribosomal subunit
SSA1	1607.9	1655.9	NQAAMNPSNTVFDAK	55–69	1.1	SRP-dependent cotranslational membrance targeting and translocation, protein folding, protein-nucleus import and translocation	Cell wall, cytoplasm, nucleus, vacuolar membrane
	1551.0	1599.0	NFTPEQISSMVLGK	111–124	1.2
	1358.8	1406.8	ELQDIANPMSK	593–604	1.0
SSA2	1551.0	1599.0	NFTPEQISSMVLGK	111–124	1.2	Protein folding	Cell wall, cytoplasm, vacuolar membrane
G3PD	1764.9	1812.9	APSSTAPMFVMGVNEVK	121–137	1.5	Not found in SGD
	1846.2	1894.2	VINDAFGIEEGLMTTVH	161–177	1.8
		874.7 (3+)	VINDAFGIEEGLMTTVHSLTATQK^c	161–184	1.43
TDH1	1729.0	1777.0	APSSSAPMFVVGVNHTK	121–137	1.3	Glycolysis, gluconeogenesis	Cell wall, cytosol, lipid particle
	1846.2	1894.2	VINDAFGIEEGLMTTVH	161–177	1.8
TDH2	1846.2	1894.2	VINDAFGIEEGLMTTVH	160–176	1.8	Glycolysis, gluconeogenesis	Cell wall, cytoplasm, cytosol, lipid particle
TDH3	1846.2	1894.2	VINDAFGIEEGLMTTVH	160–176	1.8	Glycolysis, gluconeogenesis	Cell wall, cytoplasm, cytosol, lipid particle
		874.7 (3+)	VINDAFGIEEGLMTTVHSLTATQK^c	161–184	1.4

---

In summary, the analytical procedures allowed us to confirm that the yeast can support a high degree of SeMet incorporation, seemingly with no induced stress mechanisms. The approach has been restricted to higher abundance proteins with a relatively high Met content in the sequence.

Acknowledgements

The authors are grateful to Institute Rosell-Lallemand (Montreal, QC, Canada) for supporting this research and providing the yeast sample used in this study. S. McSheehy would like to thank NSERC for financial support.

Notes and references

1. J. F. Atkins and R. F. Gesteland, Nature, 2000, 407, 461.
2. Selenium: Its Molecular Biology and Role in Human Health, ed. D. L. Hatfield, Kluwer Academic Publishers, Boston/Dordrecht/London, 2001.
3. G. N. Schrauzer, J. Nutrition, 2000, 130, 1653–1656.
4. N. Budisa, C. Minks, S. Alefelder, W. Wenger, F. Dong, L. Moroder and R. Huber, FASEB, 1999, 13, 41–51.
5. M. Korhola, A. Vainio and K. Eldelmann, Ann. Clin. Res., 1986, 18, 65–68.
6. J. R. Encinar, L. Ouerdane, W. Buchmann, J. Tortajada, R. Lobinski and J. Szpunar, Anal. Chem., 2003, 75, 3765–3774.
7. H. Chassaigne, C. C. Chery, G. Bordin, F. Vanhaecke and A. R. Rodriguez, J Anal. At. Spectrom., 2004, 19, 85–95.
8. S. McSheehy, F. Pannier, J. Szpunar, M. Potin-Gautier and R. Lobinski, Analyst, 2002, 127(2), 223–229.
9. S. McSheehy, Z. Mester and L. Yang, Anal. Chem., 2004, 76, 5149–5156.
10. L. Yang, R. E. Sturgeon, S. McSheehy and Z. Mester, J. Chroma A, 2004, 1055, 177–184.
11. L. Yang, R. E. Sturgeon, W. R. Wolf, R. J. Goldschmidt and Z. Mester, J Anal. At. Spectrom., 2004, 19, 1448–1453.
12. B. Futhcher, G. I. Latter, P. Monardo, C. S. McLaughlin and J. I. Garrels, Mol. Cell. Biol., 1999, 19, 7357–7368.

---