Genome-wide screen of Saccharomyces cerevisiae null allele strains identifies genes involved in selenomethionine resistance

Abstract

Selenomethionine (SeMet) is a potentially toxic amino acid, and yet it is a valuable tool in the preparation of labeled proteins for multiwavelength anomalous dispersion or single-wavelength anomalous dispersion phasing in X-ray crystallography. The mechanism by which high levels of SeMet exhibits its toxic effects in eukaryotic cells is not fully understood. Attempts to use Saccharomyces cerevisiae for the preparation of fully substituted SeMet proteins for X-ray crystallography have been limited. A screen of the viable S. cerevisiae haploid null allele strain collection for resistance to SeMet was performed. Deletion of the CYS3 gene encoding cystathionine gamma-lyase resulted in the highest resistance to SeMet. In addition, deletion of SSN2 resulted in both increased resistance to SeMet as well as reduced levels of Cys3p. A methionine auxotrophic strain lacking CYS3 was able to grow in media with SeMet as the only source of Met, achieving essentially 100% occupancy in total proteins. The CYS3 deletion strain provides advantages for an easy and cost-effective method to prepare SeMet-substituted protein in yeast and perhaps other eukaryotic systems.

The complete DNA sequence of many genomes has helped drive the structural genomics field, with the aim of determining many thousands of structures within the next few years. This requires substantial optimization of every step in a crystallographic structure determination. In this aspect, one important technique now routinely used in crystallography is the production of selenomethionine (SeMet) substituted recombinant proteins in Escherichia coli (1, 2). If crystals of such proteins diffract to 2 Å or better and anomalous data are collected at the proper wavelengths, excellent experimental phases can usually be obtained. Even if the maximum resolution is lower and the proteins are larger, anomalous SeMet data are very valuable. A typical protein contains 1 methionine (Met) every 50 residues, so besides providing experimental phases, the selenium positions are valuable when tracing especially large proteins.

Although many proteins can be expressed in E. coli, where the initial work on SeMet derivatives was performed and is now standard, many monomeric proteins do not fold well or are not correctly posttranslationally modified in this organism. Interestingly, more than 90% SeMet substitution has been obtained in mammalian cells (3), but the proteins of interest must be secreted, and obtaining large quantities is more expensive and laborious than in the utilization of microorganisms. Hence, access to other organisms for expression of SeMet proteins is important. The ability to use the yeast Saccharomyces cerevisiae for SeMet derivatives was advanced by the pioneering studies of Kornberg and colleagues on the structure of yeast RNA Polymerase II (4–6). This work defined the conditions to achieve ≈65% occupancy of SeMet in strains that maintained the yeast Met biosynthesis genes (a prototroph). Yeast Met auxotrophs, which cannot synthesize Met, required a ratio of 9:1 SeMet to Met. In all cases, significant growth inhibition was observed; however, this level of incorporation was not optimal for structural determination. The general problem appears to be the toxicity of high selenium concentrations to eukaryotic cells, which is somewhat surprising, because SeMet has been shown previously to be functional when incorporated into (S)-adenosylmethionine (7). The use of an industrial strain of yeast allowed the preparation of a SeMet derivative of translation elongation factor 2 (8). However, although this approach works for a very highly expressed endogenous protein, because of the difficulty in applying molecular genetic approaches in industrial strains it is not applicable to conditions where the chromosomal gene encoding the protein must be deleted or the protein of interest must be expressed from a plasmid. Recent studies have determined that deletion of the 2 genes encoding encoding (S)-adenosylmethionine (AdoMet) synthase in S. cerevisiae results in enhanced tolerance of SeMet (9). This approach, however, requires the deletion of 2 genes, supplementation of the media with (S)-adenosyl methionine, and high concentrations of SeMet to achieve 95% occupancy.

Standardized use of SeMet in yeast would benefit from strains with vigorous growth in SeMet-containing medium, higher incorporation of SeMet into proteins, and the ability to grow in the presence of SeMet as the sole Met source in a Met auxotrophic strain. Because recombinant protein production for large-scale structural genomics projects is essential (10), the ability to use a system such as S. cerevisiae in combination with high-occupancy SeMet protein production would be a major advantage. The availability of a library of viable haploid S. cerevisiae null allele strains (11) has allowed for screening the effects of loss of function on SeMet resistance. Because yeast is a well-understood and genetically tractable organism, this model system provides a unique opportunity to easily screen for molecular and cellular functions involved in selenium metabolism. The goal of this study is to identify genes that when deleted permit growth in high concentrations of SeMet, resulting in fully SeMet-substituted proteins for X-ray crystallography.

Results and Discussion

A Genetic Screen Of The Haploid Null Collection Identified 99 2-Fold SeMet-Resistant Null Allele Strains.

To apply a genomic approach to the identification of SeMet-resistant or tolerant yeast, the Open Biosystems collection of viable haploid null allele S. cerevisiae strains was used. To optimize the concentration of SeMet for the resistance screen, the Met auxotropic and prototrophic wild-type strains of opposite mating types, BY4741 (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0) and BY4742 (MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0), respectively, were tested for SeMet toxicity in YEPD and synthetic complete C-media. Similar levels of toxicity were observed between the 2 strains in YEPD media with an LD₅₀ of 0.36 mM SeMet (Fig. 1A). Both strains were more sensitive to SeMet in C-media (Fig. 1B), and the strains differed in their ability to grow in the presence of SeMet. In C-media, BY4741 has an LD₅₀ of 0.023 mM SeMet, whereas BY4742 has an LD₅₀ of 0.047 mM SeMet. This difference is likely due to the fact that BY4741 is a Met auxotroph (met15Δ0). Because of the advantages of using a Met auxotropic strain in further studies, the screen was performed in C-plus 0.20 mM SeMet media using the MATa strain collection.

Fig. 1.

A methionine auxotrophic strain is more sensitive to SeMet. Haploid wild-type Met auxotrophic (BY4741, open circle) and prototrophic (BY4742, square) strains were grown overnight at 30°C in YEPD or C- medium and diluted to 0.1 OD₆₀₀. Cells were grown for 20 h at 30°C in triplicate in varying concentrations of SeMet as indicated in (A) YEPD or (B) C- medium in a mitrotiter assay plate. Growth was monitored for triplicate samples by the change in OD₆₀₀ between 0 and 20 h and plotted as percent growth ± SEM normalized to growth in 0 mM SeMet.

The entire collection of 4786 null allele strains were grown in C-media plus 0.20 mM SeMet in microtiter plates at 30°C with constant shaking. The media consisted of 7% YEPD in which the stocks were frozen and 93% C-media. Resistant null strains were defined as growth of at least 2-fold greater than BY4741 after 20 h. A 2-fold increase in SeMet resistance was observed for 103 strains [supporting information (SI) Table S1], in which 62 strains (Table 1) exhibited at least a 3-fold increase in growth over the wild-type strain. Of those strains that grew better than wild type, the cys3Δ strain demonstrated the greatest resistance to SeMet at ≈7-fold. Genes in which null alleles showed resistance to SeMet were grouped by their Gene Ontology category from the Saccharomyces Genome Database (www.yeastgenome.org; Table 1). In addition to the 19 ORFs of currently unknown function (30%), several interesting groups of biological processes emerged. The categories included 7 genes (11%) involved in telomere and chromatin maintenance, 5 genes (8%) in transport, and 4 genes (6%) each in cell ion homeostasis or amino acid metabolism. The other categories had fewer representative genes.

Table 1.

Genes whose null alleles in haploid S. cerevisiae result in a 3-fold or greater increase in SeMet resistance

Biological process	Gene	ORF	ΔGene/WT
Telomere/chromatin maintenance	PDX3	YBR035C	6.5*
	NA	YOR008C-A	4.6*
	REC107	YJR021C	4.1*
	SSN8	YNL025C	3.5†
	PAA1	YDR071C	3.5
	HHT2	YNL031C	3.3
Transport	TPO4	YOR273C	6.9*
	SUL1	YBR294W	6.2*
	TAT1	YBR069C	3.9
	QDR3	YBR043C	3.1
	MUP1	YGR055W	3.0
Amino acid metabolism	CYS3	YAL012W	7.0*‡
	AAH1	YNL141W	4.1*
	ARO7	YPR060C	3.6
	LYS14	YDR034C	3.0
Cell ion homeostasis	BSD2	YBR290W	4.9*
	HRK1	YOR267C	4.6*
	MMT2	YPL224C	4.3*
	CSG2	YBR036C	3.1
RNA polymerase II promoter/transcription	SSN2	YDR443C	4.2*†‡
	SSN8	YNL025C	3.5†
	SOH1	YGL127C	4.2*
Catabolism	DAL3	YIR032C	3.3
	ASI3	YNL008C	3.2
	SSM4	YIL030C	3.0
Response to stress	SLM1	YIL105C	4.7*†
	GRE2	YOL151W	4.4*
	WAR1	YML076C	3.2
Gluconeogenesis/glycolysis	UBP14	YBR058C	3.4
	NA	YNL108C	3.4
	TDH2	YJR009C	3.1
Translation	TAH1	YCR060W	6.9*
	CGR1	YGL029W	3.0
Regulation of cell growth and actin organization	SLM1	YIL105C	4.7*†
	ARC18	YLR370C	3.8
DNA repair	PSO2	YMR137C	3.4
	RAD28	YDR030C	3.1
Cell wall organization/biogenesis	NA	YOL155C	4.6*
Mitochondrial genome maintenance	RIM1	YCR028C-A	3.1
Bud site selection	BUD23	YCR047C	4.5*
Ergosterol biosynthesis	HMG1	YML075C	3.9
Sterol metabolism	ARE1	YCR048W	3.3
Meiotic DNA synthesis	MUM2	YBR057C	3.3
Unknown	NA	YDR442W	6.1‡
	NA	YJL075C	5.7*
	NA	YMR135W-A	5.1*
	NA	YOR268C	4.1*
	NA	YDR042C	4.0*
	NA	YJL064W	3.7
	NA	YOR364W	3.6
	NA	YDR031W	3.5
	NA	YNL057W	3.5
	NA	YIL110W	3.5
	NA	YIR043C	3.5
	NA	YNL276C	3.4
	NA	YBR071W	3.4
	NA	YDR026C	3.2
	NA	YBR063C	3.1
	MRH1	YDR033W	3.1
	PST2	YDR032C	3.1
	NA	YJL007C	3.1
	NA	YJR024C	3.0

*Among the 22 null allele strains screened by microtiter growth assays with 3-fold better growth than a wild type strain.

^†Genes that fall into more than one biological process.

^‡Genes whose null alleles were confirmed to provide significant SeMet resistance.

More quantitative growth curve analysis over a range of SeMet concentrations was performed on the 22 most resistant null allele strains (greater than 4-fold resistance) in both 0 to 0.77 mM SeMet in YEPD and 0 to 0.1 mM SeMet in C-plus Cys media. Because the cys3Δ strain has lost a component of the cysteine biosythesis pathway, the media was supplemented with 0.12 mM Cys for the remaining experiments. Only 3 null strains, cys3Δ, ssn2Δ, and a potential ORF, ydr442wΔ, showed significant and consistent resistance to SeMet over a range of concentrations in both media (Fig. 2 A and B). The large loss of significant resistance in many strains is likely due to the presence of some YEPD in the media in the initial screen, which could allow for enhanced growth, as seen in the initial analysis of the wild-type strains (Fig. 1). Although ssn2Δ and ydr442wΔ both showed consistent resistance, they are likely similar loss-of-function mutations, because the SSN2 gene and ORF overlap on the 2 strands of the same DNA fragment on chromosome IV, resulting in deletion of a functional SSN2 gene as well. Because the cys3Δ strain demonstrated the highest level of resistance, strains lacking related genes were analyzed in more detail. These included viable haploid null alleles of CYS4, which functions upstream of CYS3, as well as potential paralogs of CYS3 (YFR055W, YJR130C, YLL058W, YML082W, YDR506C, YDR244W, and YBR250W). The cys4Δ strain showed slight sensitivity to SeMet, whereas the strains bearing a null allele of the 7 potential paralogs all showed no change in resistance to SeMet compared with the wild-type strain (data not shown). Thus, based on the consistently high level of resistance observed for the cys3Δ and ssn2Δ strains, we further characterized the nature of SeMet resistance in these strains.

Fig. 2.

A cys3Δ met15Δ0 auxotroph strain results in SeMet resistance. BY4741 (WT, circles) and the 3 most SeMet-resistant null strains (cys3Δ, diamonds; ssn2Δ, filled triangles; and ydr442wΔ, open triangles) were grown overnight at 30°C in (A) YEPD or (B) C plus Cys medium, diluted to 0.1 OD₆₀₀, and grown in triplicate in varying concentrations of SeMet. Growth was monitored as in Fig. 1.

Loss of SSN2 Results in Reduced Levels of Cys3p.

SSN2 has been identified as a part of the Mediator complex that links transcriptional regulators to RNA polymerase II and general transcription factors (12). To determine whether the ssn2Δ phenotype was an indirect result due to reduced levels of Cys3p, Western blot analysis was performed to determine Cys3p levels in a wild-type and ssn2Δ strain. In the absence of a Cys3p antibody, the TAP-tagged CYS3 strain from Open Biosystems was used, as well as the SSN2 gene deleted by PCR and transformation of the ssn2::kanMX locus from the Open Biosystems null strain. Cells were grown in YEPD to mid log phase and then treated with 0.03 mM SeMet for 0, 2, and 4 h. The level of TAP-Cys3p was slightly lower in the ssn2Δ strain, and the level of Cys3p was reduced following treatment with SeMet in both strains, although to a lesser extent in the ssn2Δ strain (Fig. 3). No change in TAP-Cys3p levels was observed in either strain with treatment with Met (data not shown). Thus, the SeMet resistance seen for the ssn2Δ strain was likely due to reduced levels of Cys3p.

Fig. 3.

Deletion of SSN2 results in reduced levels of Cys3p. The wild type and ssn2::kanMX Open Biosystems TAP-tagged CYS3 strains were grown overnight at 30°C in YEPD, diluted to an OD₆₀₀ of 0.5, and grown in YEPD with 0.03 mM SeMet for 0 to 4 h. Total protein was extracted at 0, 2, and 4 h, and 4 μg as determined by the Bradford assay was separated by SDS/PAGE. The Western blot was probed with a polyclonal TAP tag antibody (Open Biosystems) and monoclonal phosphoglycerate kinase antibody (Invitrogen).

SeMet Supports the Growth of the cys3Δ Strain.

Because the library of strains used contain the met15Δ0 allele, making them Met auxotrophs, we determined whether strains with enhanced resistance to SeMet could survive with this form of Met as the sole source in the media. C-Met plus Cys media was prepared with varying ratios of Met to SeMet to a final concentration of 0.14 mM, the standard Met concentration in defined synthetic medium. Growth analysis in a microtiter assay with equimolar Met and SeMet demonstrated that the cys3Δ strain was able to grow at 89% the rate seen in SeMet-free medium and maintained ≈70% of the growth rate in 100% SeMet (Fig. 4A). The ssn2Δ strain was unable to grow in media with SeMet as the major Met source. This also supports the hypothesis that the cys3Δ strain is able to take up SeMet from the medium, and resistance is not due to inhibition of SeMet uptake. Thus, the higher level of SeMet resistance and the ability to grow in SeMet-only medium seen for the cys3Δ strain make it a more attractive candidate for detailed analysis of SeMet metabolism.

Fig. 4.

Loss of cys3Δ confers SeMet resistance and allows viability with SeMet as the sole Met source. (A) Strains of cys3Δ (diamonds), ssn2Δ (filled triangles), and BY4741 (WT, circles) were grown overnight in C plus Cys medium at 30°C, diluted to an OD₆₀₀ of 0.1, and grown in C plus Cys medium with varying ratios of Met to SeMet to a final concentration of 0.14 mM. (B) BY4741 (circles) and cys3Δ (diamonds) strains were transformed with either an empty pRS316 plasmid (URA3, open symbols) or pTKB1017 expressing CYS3 (URA3 CEN, closed symbols), grown in C-Ura plus Cys medium overnight at 30°C, diluted to 0.1 OD₆₀₀, and grown in triplicate in varying concentrations of SeMet in C-Ura-Met media. (C) Strains of cys3Δ (diamonds), sam1Δ (X), sam2Δ (open squares), sam1Δ sam2Δ (closed squares), and BY4741 (circles) grown were grown overnight at 30°C in YEPD, diluted to 0.1 OD₆₀₀, and grown in triplicate in varying concentrations of SeMet in C-Met medium supplemented with Cys (cys3Δ) or AdoMet (sam1Δ sam2Δ). Growth was monitored as in Fig. 1.

CYS3 Overexpression and Complementation.

To confirm that the gene deleted in the cys3Δ strain was, in fact, CYS3, and no secondary alterations in the genome contributed to the SeMet resistance, the wild-type CYS3 gene was used to complement the SeMet-resistant phenotype of the strains. The cys3Δ and BY4741 strains were transformed with either pTKB1017 expressing Cys3p from its authentic promoter on a CEN plasmid or the empty vector control pRS316. The cys3Δ strain shows resistance to SeMet toxicity with the empty plasmid (Fig. 4B). However, the presence of the CYS3 plasmid results in growth of a cys3Δ strain that is essentially the same as the wild-type BY4741 strain with or without the CYS3 plasmid. This confirms that loss of the CYS3 gene is responsible for the SeMet resistance phenotype.

SeMet Incorporation into Proteins in a cys3Δ Null Strain by Total Amino Acid Analysis (TAAA).

To verify that the cys3Δ strain efficiently incorporates SeMet into protein, we determined the SeMet occupancy in total protein extracts from cells grown in the presence of SeMet under several conditions. To compare wild-type and cys3Δ strains, some Met was required in the media to support growth of the wild-type strain. Thus, cells were grown in C-overnight, diluted to an OD₆₀₀ of 0.15, centrifuged, and the pellet resuspended in C-plus Cys plus 0.02 mM SeMet (1:8 SeMet/Met) and grown at 30°C for 2 days. Total protein extracts were sent to the W.M. Keck Foundation Biotechnology Resource Laboratory for TAAA. The percent SeMet incorporation was determined by the ratio of Met to Ala in total protein extracts from cells grown in Met/SeMet media over the ratio of Met to Ala in total protein extracts from cells grown in the presence of only Met. This demonstrated that the cys3Δ strain shows 85.4 ± 0.9% SeMet occupancy compared with 62.7 ± 3.2% for the wild-type strain. This compares with 65% in the work by Bushnell et al. (4). Because the cys3Δ strain is able to grow in the complete absence of Met, protein extracts were made from cys3Δ cells inoculated directly in C-Met plus Cys plus 0.14 mM SeMet and analyzed by TAAA. The results of the cys3Δ strain grown in SeMet media showed levels of Met below the detection level of the method, and thus incorporation of SeMet at essentially 100%.

The cys3Δ and sam1Δ sam2Δ Strains Show Similar Resistance to and Incorporation of SeMet.

Recent results from Malkowski et al. (9) demonstrated that deletion of the SAM1 and SAM2 genes encoding the AdoMet synthase resulted in additive enhanced resistance to SeMet. Resistance was seen following growth of the sam1Δ sam2Δ strain on solid C-media lacking Met or a sam1Δ sam2Δ met6Δ strain on C-complete media with 0.06 mM SeMet. To directly compare the cys3Δ and sam1Δ sam2Δ strains, the Open Biosystems strain background was used. The single sam1Δ and sam2Δ mutants were obtained from the Open Biosystems collection, and the sam1Δ sam2Δ double-mutant strain was prepared by mating the single-mutant strains, sporulation of the resulting diploid, tetrad dissection, and confirmation by PCR and AdoMet dependence. As seen in Fig. 4C, equivalent partial resistance is seen for the sam1Δ or sam2Δ single-deletion strains, and additive resistance was seen for the sam1Δ sam2Δ strain equivalent to the cys3Δ strain up to 0.15 mM. Malkowski et al. (9) demonstrated that deletion of the SAM1 gene resulted in greater resistance than deletion of the SAM2 gene. To determine the possible reasons for this difference, sam1Δ and sam2Δ single-mutant strains as well as a sam1Δ sam2Δ double-mutant strain in the BY4742 (MATα MET15) were analyzed for SeMet resistance. In this strain background, the sam1Δ showed slightly higher resistance than the sam2Δ strain, although the effect of the sam1Δ sam2Δ double mutant was unchanged (data not shown). This may indicate that the status of the MET pathway affects the resistance of the single SAM null allele strains. Thus, the single cys3Δ deletion shows equivalent resistance to SeMet as the sam1Δ sam2Δ strain.

Analysis of SeMet occupancy in a protein purified from the sam1Δ sam2Δ strain showed 83% occupancy in the presence of 0.125 mM SeMet and 95% occupancy in 0.5 mM SeMet (9). To compare the results for an expressed protein in the cys3Δ and sam1Δ sam2Δ strain, a plasmid expressing 6×His tagged S. cerevisiae eEF1Bα was transformed into each strain. Cells were grown in C-Met plus 0.3 mM SeMet, and the eEF1Bα–eEF1A complex was purified by Ni affinity chromatography. The purified proteins were resolved by SDS/PAGE, and the eEF1A band was cut out and subjected to mass spectroscopic analysis by LC-MSMS. Protein purified from the cys3Δ strain showed an average of 88.3 ± 6.7% occupancy over 4 experiments, with a range of 82% to 97% SeMet, whereas the protein from the sam1Δ sam2Δ strain showed an average of 74.4 ± 15.3%, with a range of 59% to 92% occupancy. These data suggest that the utilization of a cys3Δ strain may yield slightly better SeMet occupancy.

The enhanced incorporation of SeMet into proteins in an S. cerevisiae strain lacking the CYS3 gene can be performed in a Met auxotropic strain, requires the deletion of only a single gene, requires standard media supplemented only with cysteine, and requires a low concentration of SeMet for essentially 100% occupancy. Combined with the significantly lower cost of Cys supplementation compared with ADOMET supplementation, the cys3Δ strain has clear advantages over the sam1Δ sam2Δ system. In addition, the cys3Δ paradigm may be transferable to the inhibition of the expression of the CYS3 homologs in other eukaryotes to further our understanding of selenium toxicity in mammals.

The precise mechanisms for selenium toxicity are not known and appear to be highly compound and assay-dependent (reviewed in refs. 13 and 14). One commonly proposed mechanism for selenium toxicity involves the production of hydrogen selenide and/or methylselenol, resulting in the glutathione-dependent production of superoxides. In yeast, the production of extracellular hydrogen selenide by the combination of selenite and reduced glutathione in the growth medium has been shown to be toxic (15). Recent work in fish has suggested that SeMet is only toxic in the presence of a Met gamma lyase activity that would generate methylselenol (16). S. cerevisiae does not harbor a known Met gamma lyase activity. Our results indicate the CYS3 gene encoding cystathionine gamma lyase may metabolize SeMet to methylselenol. Thus, loss of the CYS3 gene, which shares significant sequence similarity to mammalian Met gamma lyase, provides high levels of resistance to SeMet.

Selenium supplementation has been reported to reduce the incidence of prostate, colorectal, and lung cancer using supplements in the form of selenized yeast, which contains ≈85% SeMet (12, 13). However, follow-up analysis showed that only the patients with low baseline selenium levels benefited from increased dietary selenium (17). Selenium supplementation did not prevent the recurrence of skin cancer, and a more recent follow-up has shown that squamous cell carcinoma and total nonmelanoma skin cancer incidence increased in this patient population (18), highlighting the potential danger of toxicity resulting from selenium supplementation (19, 20). Selenium only acts as an antioxidant when used as selenocysteine in a specific group of oxidoreductases (21). In fact, both organic and inorganic forms of selenium are potent oxidants as a consequence of selenium metabolism. This distinction is especially important in yeast because selenium is not incorporated into selenoproteins in fungi or plants, thus potentially limiting the enzymatic reductive capacity of these organisms. Although classic studies revealed that SeMet was efficiently incorporated into the Met pathway (7), the toxicity of SeMet has not been examined in yeast. In contrast, the effects of selenite have been studied by several groups spanning several decades (15, 22–25), underscoring the importance of yeast as a model system for deciphering the mechanisms of selenium toxicity. Indeed, recent data analyzing the role of genes in the RAD9-dependent DNA repair pathway, the RAD6/RAD18 DNA damage tolerance pathway, and the oxidative stress pathway DNA damage tolerance pathway in S. cerevisiae suggest that both selenite and SeMet are likely inducing DNA damage by generating reactive species (19). Thus the study of selenium metabolism in yeast supports an increased understanding of the mechanisms of Se toxicity and resistance, which is an essential part of establishing the extent to which dietary selenium supplementation should be considered.

Materials and Methods

Media, Plasmids, and Strains.

The parental wild-type strains BY4742 (MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0) and BY4741 (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0) or haploid deletions for nonessential genes or TAP-tagged wild-type genes in the MATa BY4741 strain background were obtained form Open Biosystems. Standard yeast genetics methods were used (26). Yeast cells were grown in either YEPD (1% Bacto yeast extract, 2% peptone, and 2% dextrose) or defined synthetic complete media (C-) supplemented with 2% dextrose as a carbon source (27) was used. Yeast cells were transformed by the lithium acetate method (28).

Recombinant DNA techniques were performed as described previously (29). Restriction digestion enzymes were obtained from New England Biolabs. The wild-type CYS3 gene was amplified by PCR from total genomic DNA isolated from BY4741 using primers 5′-CTC TTG AAG CTT CCG ACA TCG AGT ACA AGT TCG-3′ and 5′-CTC TTG AAG CTT TTT AGC TTC AGC GCC TCC TG-3′. The resulting fragment was digested with HindIII and cloned into pRS316 (URA3 CEN) to produce plasmid pTKB1017.

Deletion of SSN2 in the Open Biosystems TAP-tagged CYS3 strain was performed by PCR of the ssn2::KanMX locus from Open Biosystems using primers 5′-CCCAATGCCAGTGCTAGGGATGGAA-3′ and 5′-GCGGTCGCAGAAGCGCGCTTTCCTC-3′. The resulting PCR fragment was transformed into the TAP-tagged Cys3p strain and ssn2::KanMX cells selected on YEPD with 200 μg/ml G418 to produce strain TKY1334.

SeMet Growth Assays.

Yeast strains were grown overnight in 5 ml YEPD orC-based media to mid log phase. The OD₆₀₀ was taken by UV-Visible spectroscopy (Thermo Scientific), and cultures were diluted to an OD₆₀₀ of 0.1 in the same media. Concentrations from 0 to 1.5 mM SeMet were made in YEPD or 0 to 0.2 mM SeMet in C-media and filter sterilized. Equal volumes of the SeMet medium and the diluted yeast were added to a 96-well microtiter plate for a final volume of 200 μl. The final OD₆₀₀ was ≈0.05, and the final concentration of SeMet varied from 0.0 to 0.77 mM or 0 to 0.1 mM in YEPD or C-media, respectively. C-media was supplemented with with 0.12 mM cysteine for a cys3Δ strain or 0.06 mM AdoMet for a sam1Δ sam2Δ strain. Plates were incubated at 30°C, and the OD₆₀₀ was taken using the ELx 800 microplate reader (Bio-Tek Instruments) at 0 and 20 h. All samples were analyzed in triplicate.

Screen for SeMet Resistance.

The appropriate SeMet concentration for the screen was determined through SeMet growth assays using the null allele background strains BY4741 and BY4742, as described above. BY4741 was grown overnight in 5 ml YEPD and used as a negative control. Each 96-well microtiter plate of the null collection (Open Biosystems) was defrosted, and 10 μl of each strain was added to 140 μl C-plus 0.204 mM SeMet in a fresh 96-well plate, producing a final concentration of 0.20 mM SeMet. The microtiter plates were grown overnight at 30°C with constant shaking, and the OD₆₀₀ was measured at 0 and 20 h. All results were compared to the growth of isogenic wild-type BY4741 present on the same 96-well plate. The 22 null allele strains that demonstrated 4-fold or more resistance to SeMet were analyzed using 3 colonies in triplicate in SeMet microtiter plate growth assays as outlined above in YEPD and C plus Cys media and compared to the isogenic wild-type strain BY4741.

Analysis of Growth in SeMet/Met Mixture Media.

The ability to use SeMet as the source of Met was determined by mixing SeMet and Met in C-Met plus Cys medium. The final concentration of Met plus SeMet was 0.14 mM, whereas the amounts of SeMet were 0%, 50%, 83%, 91%, and 100%. Cells were grown overnight in C-plus Cys medium at 30°C and diluted to an OD₆₀₀ of 0.1. A total of 1 ml of each diluted culture was transferred to tubes, centrifuged, and the pellet resuspended in the various media. Aliquots of 250 μl, in triplicate, were placed in a microtiter plate and incubated at 30°C. The OD₆₀₀ was measured at 0 and 20 h.

TAAA and Mass Spectroscopy.

Wild-type and cys3Δ strains were grown in C- plus Cys medium overnight and diluted to an OD₆₀₀ of 0.15 the following day. Cells were centrifuged, and the pellet was resuspended in C-Met plus Cys plus 0.02 mM SeMet or C plus Cys medium and incubated at 30°C for 2 days. The cys3Δ strain was inoculated in triplicate in C- plus Cys medium or C-Met plus Cys plus 0.14 mM SeMet and grown at 30°C overnight to an OD₆₀₀ of 1.3. Total yeast protein extracts were prepared using the Y-PER Plus, Dialyzable Yeast Protein Extraction Reagent kit (Pierce Biotechnology), and samples were sent to the W.M. Keck Foundation Biotechnology Resource Laboratory at Yale University. Aliquots of each sample were dried and oxidized by incubation with performic acid at 4°C for 16 h to convert Met and Met-sulfoxide to Met-sulfone. The performic acid was dried off, and the samples were resuspended in 100 μl water and dried. Aliquots were hydrolyzed in vacuo with 6 N HCl/0.2% phenol (with 2 nmol norleucine/100 μl as an internal standard) for 16 h at 115°C. The hydrolysis acid was dried off and the aliquots dissolved in 100 μl sodium citrate loading buffer (Beckman Na-S with 20 μM homoserine as a second internal standard). Analysis was done on a Beckman 7300 Amino Acid Analyzer using an ion-exchange column to separate the amino acids and then derivatize them with ninhydrin/hydrindantin for detection at 570 nm and 440 nm. The percent of SeMet incorporation was determined by the difference in the ratio of Met-sulfone to other amino acids between the SeMet and control samples. Standard deviation is based on 4 independent analyses.

The cys3Δ and sam1Δ sam2Δ strains were transformed with pTKB339 expressing 6×His-tagged eEF1Bα. Cells were grown in C-Ura media overnight at 30°C. Cells were diluted to an OD₆₀₀ of 0.1 in 50 ml C-Ura-Met plus 0.3 mM Met media and C-Ura-Met plus 0.3 SeMet media with 0.12 mM cysteine for a cys3Δ strain or 0.06 mM AdoMet for a sam1Δ sam2Δ strain, and were grown overnight to an A₆₀₀ of 0.5 to 1. Cells were harvested by centrifuge at 5,000 × g and lysed with glass beads in LEW buffer (50 mM NaH₂PO₄ 300 mM NaCl, pH 8.0). The purified 6×His-tagged eEF1Bα–eEF1A complex was purified with the PrepEase His-tagged protein purification kits (USB Corp.). The complex was run on SDS/PAGE, and the eEF1A band was cut from the gel and subjected to standard in-gel digestion, including reduction of disulfide bonds with DTT and alkylation of sulfohydro groups on cysteine with iodoacetamide before tryptic digestion and peptide extraction. All LC-MSMS experiments were performed using U3000 (Dionex) in nano-LC mode on line with LTQ (ThermoFisher). Extracted peptides from in-gel digestion were first solubilized in 0.1% TFA and loaded on to a self-packed 75 μm × 12 cm emitter column packed with Magic C18AQ, 3 μm, 200 Å (Michrom Bioresources Inc.) and eluted with a linear gradient of 2% to 45% mobile phase B in 30 min (mobile phase A: 0.1% formic acid/water; mobile phase B: 0.1% formic acid/acetonitrile). Mass spectrometry data were acquired using a data-dependent acquisition procedure with a cyclic series of a full scan followed by zoom scans and MSMS scans of the 5 most intense ions with a repeat count of 2 and the dynamic exclusion duration of 30 sec. The LC-MSMS data were searched against yeast database using an in-house version of the GPM (GPM Extreme; Beavis Informatics Ltd.) with fixed modification of iodoacetamide modification of cysteine (+57 Da) and potential modifications of oxidation of methionine (+15.99) and SeMet (+47.9 Da).

To quantitate the incorporation rate of SeMet, LC-MSMS was conducted again with the same HPLC program as a data-dependent run, except the mass spectrometry was set to acquire targeted MS/MS of the control peptide (IGGIGTVPVGR m/z = 513.7) and methionine peptide (VETGVIKPGMVVTFAPAGVTTEVK m/z = 811.3). The peak area of each peptide was calculated by peak integration of 2 prominent fragment ions. The percent of SeMet incorporation was determined by the ratio of peak area of Met peptide to the peak area of the control peptide from cells grown in SeMet media over the ratio of peak area of Met peptide to the peak area of the control peptide from cells grown in Met media.