Most investigations into plasma membrane electron transport (PMET) in Saccharomyces cerevisiae have focused on the inducible ferric reductase responsible for iron uptake under iron/copper-limiting conditions. In this paper, we describe a PMET system, distinct from ferric reductase, which reduces the cell-impermeable water-soluble tetrazolium dye, 2-(4-iodophenyl)-3-(4-nitrophenyl)-5-(2,4-disulphophenyl)-2H-tetrazolium monosodium salt (WST-1), under normal iron/copper conditions. WST-1/1-methoxy-phenazine methosulphate reduction was unaffected by anoxia and relatively insensitive to diphenyleneiodonium. Dye reduction was increased when intracellular NADH levels were high, which, in S. cerevisiae, required deletion of numerous genes associated with NADH recycling. Genome-wide screening of all viable nuclear gene-deletion mutants of S. cerevisiae revealed that, although mitochondrial electron transport per se was not required, the presence of several nuclear and mitochondrially encoded subunits of respiratory complexes III and IV was mandatory for PMET. This suggests some form of interaction between components of mitochondrial and plasma membrane electron transport. In support of this, mitochondrial tubular networks in S. cerevisiae were shown to be located in close proximity to the plasma membrane using confocal microscopy.
Plasma membrane electron transport (PMET) is a universal property of all living cells. The best-known PMET system in Saccharomyces cerevisiae is the Fre1p/2p ferric reductase system responsible for iron uptake, induced under iron/copper-limiting growth conditions and extensively inhibited by anoxia (Rosenfeld & Beauvoit et al., 2003) and diphenyleneiodonium (Finegold et al., 1996; Lesuisse et al., 1996; Shatwell et al., 1996). Ferric reductase has also been shown to reduce aromatic azo dyes (Ramalho et al., 2005) and, to a lesser extent, the cell-impermeable dye, 2-(4-iodophenyl)-3-(4-nitrophenyl)-5-(2,4-disulphophenyl)-2H-tetrazolium monosodium salt (WST-1), in the absence of the intermediate electron acceptor 1-methoxy-phenazine methosulphate (PMS) (Knight & Dancis et al., 2006). Here, we report the presence in S. cerevisiae of a PMET system distinct from ferric reductase, which reduces WST-1 indirectly via PMS.
In mammalian cells, PMET functions as an alternative pathway to oxidize NADH, in addition to mitochondrial electron transport and lactate dehydrogenase activity. Thus, mammalian PMET maintains intracellular NADH/NAD+ ratios favourable for continued glycolytic ATP production. PMET is vital for the growth and proliferation of cancer cells that rely on glycolysis for energy production (Herst & Berridge et al., 2006, 2007). In contrast, S. cerevisiae is a metabolically versatile organism with numerous overlapping pathways that can be exploited for recycling NADH (Fig. 1). The presence of yet another NADH-oxidizing pathway in S. cerevisiae in the form of PMET is intriguing. In this paper, we investigate the extent of PMET in wild-type S. cerevisiae cells as well as in mutant strains with compromised NADH oxidation.
NADH-recycling strategies in Saccharomyces cerevisiae. (1) Reoxidation of mitochondrial NADH directly via the internal NADH dehydrogenase (Ndi1p), ubiquinone (CoQ6), cytochrome bc1 (bc1), cytochrome c (Cyt C) and cytochrome c oxidase (COX) (Luttik et al., 1998). (2) Reoxidation of cytosolic NADH directly via the external NADH dehydrogenases 1 and 2 (Nde1p, 2p) (Luttik et al., 1998). (3) Glucose fermentation to ethanol, reoxidation of NADH produced during glycolysis using pyruvate decarboxylases 1, 5 and 6 (Pdc1p,5p,6p) (Bakker et al., 2000). (4) Transport of ethanol across the mitochondrial membrane, reduction to acetaldehyde, NADH reoxidized via Ndi1p (Bakker et al., 2001). (5) Glycerol production, which is a net NADH-oxidizing process (Pahlman et al., 2001). (6) Electrons from glycerol-3-phosphate (G-3-P) can be transferred to ubiquinone Q6, CoQ6 via mitochondrial FAD-dependent glycerol-3-phosphate dehydrogenase (Gut2p) (Overkamp et al., 2002; Rigoulet et al., 2004). (7) Cytosolic and mitochondrial NADH reoxidized via plasma membrane electron transport (tPMET), this paper. Adapted from Bakker (2000).
Mammalian PMET consists of an NADH : oxidoreductase at the inner leaflet of the plasma membrane, membrane ubiquinones and a surface oxidase at the outer leaflet of the plasma membrane (Herst & Berridge et al., 2006). The surface oxidase of mammalian cancer cells has recently been identified as tNOX, a member of the ECTO-NOX family of cell surface NADH oxidases (Chueh, 2002a, b; Scarlett et al., 2005), which has been found to be closely associated with the heat shock protein GRP94/gp96 (Scarlett et al., 2005). However, the S. cerevisiae genome does not appear to contain any obvious tNOX or gp96 orthologues. Therefore, we performed a genome-wide screen of all viable nuclear gene-deletion mutants of S. cerevisiae (Winzeler et al., 1999) to identify mutant strains lacking PMET.
Materials and methods
Strains and growth conditions
All S. cerevisiae strains in the genome-wide screen were diploids, homozygous for deletion of the entire ORF of interest and were derivates of the wild-type strain BY4743 (MATa/MATαura3Δ0/ura3Δ0 leu2Δ0/leu2Δ0 his3Δ1/his3Δ1 met15Δ0/MET15 lys2Δ0/LYS2). All S. cerevisiae strains were maintained on synthetic defined medium (SD) containing 0.17% yeast nitrogen base without amino acids, 0.5% ammonium sulphate, 2% (w/v) d-glucose, supplemented with uracil (22.4 mg L−1), lysine (60 mg L−1), leucine (131 mg L−1), histidine (209 mg L−1) and stored in SD supplemented with 10% (w/v) glycerol at −80 °C.
Saccharomyces cerevisiaeρ° strains were derived from respiratory-competent parental BY4743, CEN.PK113-7D and D273-10B by incubation of exponentially growing cells in SD medium supplemented with 25 μg mL−1 filter-sterilized ethidium bromide (Fox et al., 1991). Cells were allowed to grow to saturation and reinoculated into fresh SD medium supplemented with ethidium bromide. The culture was grown to saturation and streaked on SD medium. Single colonies were replica-plated onto YPEG medium [containing 1% yeast extract, 2% (w/v) Bacto peptone, 1% (v/v) ethanol and 3% (v/v) glycerol] and onto YPD medium. One colony growing on YPD but not on YPEG was further assessed for ρ° status by carbonyl cyanide para-tri-fluoromethoxyphenylhydrazone (FCCP)-insensitive WST-1/PMS reduction (Herst & Berridge et al., 2007).
Construction of the yeast genome deletion library was described previously (Winzeler et al., 1999). Strains were grown in glucose-rich medium (YPD) containing 1% yeast extract, 2% (w/v) glucose and 2% (w/v) bacteriological peptone. Where required, media were solidified by addition of 2% (w/v) agar.
Growth under aerobic and anaerobic conditions was compared with growing cells in a specially designed gassing manifold, using a multi-station magnetic stirrer, combined with a temperature-controlled water bath. WST-1/PMS reduction was measured after incubating 5 mL of a 2 × 107 cells mL−1 culture of exponentially growing S. cerevisiae cells in 20-mL universals under normoxic conditions (20% O2, 5% CO2 and 75% N2), under hypoxic conditions (5% O2, 5% CO2 and 90% N2 and 1% O2, 5% CO2 and 94% N2) and under anaerobic conditions (95% N2 and 5% CO2).
The Cox6-green fluorescent protein (GFP) fusion strain was derived from the wild-type strain ATCC201388 (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0) and was maintained on SD–histidine media. The Sik1-red fluorescent protein (RFP) Mrh1-RFP strain was derived from the wild-type strain Y8205 (MATαCan1Δ∷STE2pr-Sp_his5 Lyp1Δ∷STE3pr-LEU2 his3Δ1 leu2Δ0 ura3Δ0) and was maintained on YPD media supplemented with 100 μg mL−1 nourseothricin.
Professor Jack Pronk (Delft University of Technology, the Netherlands) kindly provided the following strains: S. cerevisiae CEN.PK113-7D wild-type and the gene-deletion mutants: nde1Δ, nde2Δ, nde1Δnde2Δ, gut2Δ, nde1Δgut2Δ, nde2Δgut2Δ, nde1Δnde2Δgut2Δ (Luttik et al., 1998; Overkamp et al., 2000) and nde1Δnde2Δgut2Δpdc1Δpdc5Δpdc6Δ (TAM3KO) (Geertman et al., 2006).
Professor Thomas Fox (Cornell University, NY) kindly provided the following strains: S. cerevisiae D273-10B carrying an ARG8m construct in their mtDNA (Steele et al., 1996), which replaces one of the three mitochondrially encoded cytochrome c oxidase (COX) genes, XPM10b (cox1Δ) (Perez-Martinez et al., 2003), HMD22 (cox2Δ) (Bonnefoy et al., 2001), DFS189 (cox3Δ) (Steele et al., 1996) and the CAB51A (ARG8m) insert control. The mtDNA mutants were maintained and grown in medium lacking arginine to facilitate maintenance of the inserted ARG8m construct and the remaining mtDNA.
Dr Patricia Ramalho (University of Minho, Portugal) kindly provided the following strains: S. cerevisiae CEN.PK113-7D wild-type and the gene-deletion mutants, fre1Δ, fre2Δ and fre1Δ2Δ (Ramalho et al., 2005).
WST-1 and PMS were purchased from Dojindo Laboratories (Kumamoto, Japan). Unless otherwise stated, all other reagents were from Sigma Chemical Company (St. Louis, MO).
PMET activity/genomic screen
Viable nuclear gene-deletion mutants were grown in separate wells in round-bottomed 96-well microplates in YPD for 48 h to the mid-exponential phase. Triplicates of each microplate, containing at least two wild-type and two buffer control wells, were tested for PMET activity (represented by WST-1/PMS reduction) as follows: Microplates were centrifuged for 5 min at 300 g. Cells in wells were washed twice in Hanks balanced salt solution (HBSS) buffer and resuspended in 150 μL HBSS buffer. Dye reduction was initiated by adding 15 μL of a 10 × stock solution of WST-1/PMS in milliQ water to each of the wells (final concentrations: 500 μM WST-1 and 20 μM PMS). Following incubation at 30 °C for 2 h, the plates were centrifuged at 300 g for 5 min, then 100 μL of the supernatant was carefully removed and placed in flat-bottomed microwell plates. The amount of reduced WST-1 was measured as A450 nm in a BMG FLUOstarOptima plate reader. All values were corrected for auto-oxidation (buffer with WST-1 and PMS).
Confocal fluorescent microscopy was used to localize mitochondria-specific and plasma membrane-specific proteins. The MATa Cox6-GFP fusion strain, obtained from the S. cerevisiae GFP (Yeast–GFP) clone library, was used as a mitochondrial inner membrane marker. Mrh1p was used as a plasma membrane marker and Sik1p as a nucleolar marker by chromosomal incorporation of a PCR-amplified carboxy terminal RedStar2 RFP fusion protein cassette into the MATα strain Y8205 (Janke et al., 2004). These strains were mated and diploid cells were selected on SD-histidine+nourseothricin media. Live cells from an overnight culture were photographed using an Evotec Opera confocal microscope with a dual detection camera set-up, × 60 water immersion lens and the appropriate laser and filter configuration for GFP and RFP excitation and detection. Images were taken from multiple z sections and overlaid using acapella software.
Characterization of PMET in S. cerevisiae
PMET activity of S. cerevisiae was shown to be cell concentration dependent (Fig. 2). PMET activity of S. cerevisiae per unit surface area was about 1% of those previously measured for human leukemia HL60 cells but similar to that of quiescent mammalian cells (Berridge & Tan et al., 2000; Herst et al., 2004). In HL60, inhibition of mitochondrial electron transport (MET) by potassium cyanide and myxothiazol, and of glycolysis by 2-deoxy-d-glucose and iodoacetamide, increases intracellular NADH levels and thus PMET. Conversely, the uncoupler carbonyl cyanide FCCP increases MET, and decreases intracellular NADH levels and thus PMET (Herst et al., 2004). However, PMET activity by S. cerevisiae was unaffected by these compounds (Table 1). Under anaerobic conditions, mammalian PMET is increased three- to fourfold due to competition between the physiological electron acceptor, oxygen and the artificial electron acceptor, PMS, for electrons from PMET (Herst et al., 2004). In contrast, PMET activity in S. cerevisiae cells was unaffected by oxygen levels from aerobic through to anaerobic (Table 1).
Cell concentration-dependent PMET of Saccharomyces cerevisiae. WST-1/PMS reduction by S. cerevisiae cells (▪) and by HBSS buffer with WST-1 and PMS (◻). Results represent averages±SEM of four separate experiments.
Effect of different conditions on PMET in Saccharomyces cerevisiae and HL60
PMET (% control)
Potassium cyanide (1 mM)
93 ± 8
172 ± 7
Myxothiazol (1 μg mL−1)
97 ± 7
134 ± 6
FCCP (2 μM)
93 ± 3
48 ± 4
Iodoacetamide (2 mM)
92 ± 5
18 ± 6
2-Deoxy-d-glucose (5 mM)
94 ± 9
65 ± 4
5% Oxygen (hypoxic)
98 ± 4†
263 ± 8†
1% Oxygen (hypoxic)
94 ± 5†
331 ± 16†
0% Oxygen (anaerobic)
96 ± 7
345 ± 11
Diphenyleneiodonium (20 μM)
40 ± 5
58 ± 8
↵* PMET activity of the wild type (BY4743) was 90 ± 8 milliA450 h−1 per 106 cells; PMET activity of HL60 was 16.3 ± 2.9 milliA450 min−1 per 105 cells. Results represent averages ± SEM of three† and five separate experiments.
We next examined whether ferric reductase (FRE) could be responsible for WST-1/PMS reduction. Ferric reductase has been shown to be extensively inhibited by anoxia (Rosenfeld & Beauvoit et al., 2003) and diphenyleneiodonium (Finegold et al., 1996; Lesuisse et al., 1996; Shatwell et al., 1996). We found that PMET by S. cerevisiae was only moderately sensitive to diphenyleneiodonium and unaffected by anaerobiosis (Table 1). Ferric reductase activity is mediated by the flavocytochromes Fre1p (Dancis et al., 1992) and to a lesser extent Fre2p (Georgatsou & Alexandraki et al., 1994). The S. cerevisiae genome contains a total of nine FRE homologues; of these, Fre1p, 2p, 3p and 4p are present in the plasma membrane (Yun et al., 2001). In our study, PMET was measured under iron/copper-rich conditions and the fre1Δ, fre2Δ, fre3Δ and fre4Δ single mutants exhibited PMET activities similar to that of the wild-type control (Table 2). Unfortunately, we were unable to establish the extent of PMET in the double mutant, fre1Δfre2Δ. This mutant reduced WST-1 in the absence of PMS and appeared bright yellow under microscopic analysis, which are signs of intracellular dye reduction caused by compromised plasma membrane integrity. Compromised plasma membrane integrity of the fre1Δfre2Δ mutant was confirmed by colocalization of the yellow colour with the membrane integrity indicator dye propidium iodide (results not shown).
PMET activity in fre1Δ, fre2Δ, fre3Δ and fre4Δ mutants
PMET (% control)
87 ± 4
93 ± 5
108 ± 11
80 ± 8†
↵* PMET activity of CEN.PK wild-type was 88 ± 11 milliA450 h−1 per 106 cells. Results represent averages ± SEM of three† and five separate experiments.
↵‡‡Not applicable due to increased permeability of cells.
The low PMET activity raises questions about its physiological role in these metabolically versatile yeasts. We, therefore, measured PMET in S. cerevisiae cells harbouring single or multiple mutations in NADH-oxidation pathways (Table 3). PMET activities for all mutants were found to be similar to those of the wild-type strain, with the exception of the nde1Δnde2Δgut2Δpdc1Δpdc5Δpdc6Δ (TAM3KO) mutant, which can neither perform alcoholic fermentation nor oxidize cytosolic NADH via the external NADH dehydrogenases and the glycerol-3-P shuttle (Geertman et al., 2006).
PMET activity in mutants with deletions in genes associated with NADH recycling
PMET (% control)
96 ± 4
95 ± 4
106 ± 3
91 ± 5
104 ± 3
94 ± 4
101 ± 5
196 ± 8†
↵* PMET activity of wild-type CEN.PK was 74 ± 7 milliA450 h−1 per 106 cells. Results represent averages ± SEM of two† and five separate experiments.
Genes involved in PMET in S. cerevisiae
An initial screening of the viable nuclear gene-deletion mutant library (Winzeler et al., 1999) identified 215 out of 4661 mutants with dramatically reduced PMET. Surprisingly, most of these mutants were respiratory incompetent (petites). Dimmer (2002) demonstrated that 341 nuclear ORFs are required for mitochondrial respiratory function and/or normal mitochondrial morphology. However, deletion of only a subset of these 341 ORFS appeared to correlate with low PMET. We re-examined each of the 341 ORFS and any additional ORFS identified by the initial screen in triplicate for PMET activity (Table 4 and supplementary Table 1). Of particular interest was the large number of nuclear gene mutants (19) involved in the expression, maturation and transport of the mitochondrially-encoded structural subunits of the cytochrome c oxidase complex, COX.
We next examined the effect of single gene deletions in any of the three mitochondrially encoded COX genes. The cox1Δ, cox2Δ or cox3Δ mutants were constructed by replacing the structural part of each gene with a gene cassette, which confers arginine prototrophy (Steele et al., 1996). The respiratory-competent ARGm insert control strain exhibited normal PMET levels. Disruption of each of the mitochondrially encoded COX1, COX2 and COX3 genes resulted in low PMET (Table 5). This demonstrates that the presence of Cox1p, Cox2p and Cox3p in the inner mitochondrial membrane is necessary for wild-type levels of PMET in S. cerevisiae. As anticipated, deletion of mtDNA (ρ°) in wild-type cells, isogenic to the strains deleted for COX1, 2 or 3, resulted in low PMET (Table 5). The ρ° phenotype in the other genetic backgrounds used in our experiments had similar low PMET (results not shown). This response in S. cerevisiae cells is in stark contrast to mammalian ρ° cells that exhibit a two- to threefold increase in PMET (Herst et al., 2004).
PMET in mutants lacking the mtDNA-encoded subunits of COX
PMET (% control)
Insert control ARG8m
16 ± 4
10 ± 5
13 ± 3
ρ° derivative of insert control
8 ± 3
↵* PMET of insert control was 72 ± 10 milliA450 h−1 per 106 cells. Results are averages ± SEM of at least four separate experiments.
We then investigated whether the dependence of PMET on mitochondrial membrane components might be explained by a close association between mitochondria and the plasma membrane. We used S. cerevisiae cells that expressed the inner mitochondrial membrane protein, Cox6p, fused to GFP (Cox6-GFP), as well as the plasma membrane and nucleolar proteins, Mrh1p/Sik1p, fused to RFP (Mrh1/Sik1-RFP). Confocal fluorescent images (Fig. 3) clearly demonstrate the close proximity of mitochondrial membrane networks to the plasma membrane.
Confocal microscopy images of Saccharomyces cerevisiae showing the close proximity of the mitochondria (green: Cox6-GFP) to the plasma membrane (red: Mrh1/Sik1-RFP). (a) Top section, (b) middle section and (c) bottom section. The mitochondria appear to form tubular networks that are situated in close proximity to the plasma membrane.
In contrast to mammalian cells, PMET in S. cerevisiae was unaffected by changes in the intracellular NADH flux as evidenced by its insensitivity to KCN, myxothiazol, FCCP, 2DOG, IA and anaerobiosis. This could be explained by the fact that this yeast has multiple NADH-recycling pathways. In this context, it is important to note that the multiple mutant, TAM3KO, which relies on glycerol production for NADH oxidation, showed a twofold increase in PMET activity (Geertman et al., 2006).
Although the genomic screen did not identify a single gene-deletion mutant with completely abolished PMET, 82 nuclear-deletion mutants, three mtDNA mutants and ρ° cells all had low PMET, ≤15% of wild-type levels. The majority of low PMET nuclear mutants were linked to the presence of ubiquinone and several cytbc1 and COX subunits in the inner mitochondrial membrane. Ubiquinone has been shown to be involved in mammalian PMET (Herst et al., 2004; Scarlett et al., 2004) and in ascorbate stabilization in the plasma membrane of S. cerevisiae (Santos-Ocaňa et al., 1995, 1998a, b).
PMET in S. cerevisiae relies on the presence of ubiquinone and components of cytbc1 and COX but not on active MET. This can perhaps be understood by postulating a novel role for these respiratory subunits, as was proposed previously by Dagsgaard (2001), who reported the presence of small amounts of nuclear- and mitochondrially encoded COX subunits in the promitochondria of anoxic cells. The authors question the role of COX subunits under anoxic conditions and hypothesize that the COX subunits may fulfil another role in anoxic cells. We would speculate that WST-1 dye reduction in S. cerevisiae that occurs at the cell surface is facilitated by respiratory subunits via the formation of restricted domains between the outer mitochondrial and plasma membranes. These domains would lead to direct communication between WST-1, PMS and components of the MET chain, with MET being involved in dye reduction in S. cerevisiae in addition to PMET.
In support of such a hypothesis, we have shown that the tubular networks of the mitochondria in S. cerevisiae are situated in close proximity to the plasma membrane, which may facilitate formation of restricted domains between the outer mitochondrial membrane and the PMET system and communication between the mitochondria and components of the PM. In yeasts, regions of endoplasmic reticulum in close proximity to the mitochondrial and plasma membranes are defined by an increased capacity to synthesize lipid/phospholipids and/or enrichment for enzymes involved in lipid/phospholipid metabolism (Ardail et al., 1993; Pichler et al., 2001). This suggests that direct physical contact and/or formation of restricted domains between mitochondrial and plasma membranes may facilitate transfer of reducing equivalents from the mitochondria to the outside of the cell in S. cerevisiae, resulting in dye reduction at the cell surface.
This work was supported by grants from the Cancer Society of New Zealand (M.B.), the Radiation Therapy Department, Wellington School of Medicine and Health Sciences, University of Otago, New Zealand (P.H.), and the Australian Research Council (I.D.). We wish to thank Professor Jack Pronk for supplying S. cerevisiae CEN.PK wt and mutants deficient in NADH-recycling pathways; Dr J.M. Geertman for performing WST-1/PMS reduction experiments on CEN.PK wt and TAM3KO; Professor Thomas Fox for supplying the single mitochondrial gene knockout mutants of S. cerevisiae; Dr Patricia Ramalho for supplying S. cerevisiae CEN.PK wt and the fre1Δ and fre2Δ mutants; and Anthony Fok, who carried out the preliminary pilot work for this study.
(2006) Reduction of the 2,3-bis(2-methoxy-4-nitro-5-sulphenyl)-2H-tetrazolium-5-carboxanilide inner salt (XTT) is dependent on CaFRE10 ferric reductase for Candida albicans grown in unbuffered media. Microbiology 152(Part 8): 2301–2308.
(2001) The yeast glycerol-3-phosphatases Gpp1p and Gpp2p are required for glycerol biosynthesis and differentially involved in the cellular responses to osmotic, anaerobic and oxidative stress. J Biol Chem 276: 3555–3563.