Lab de Biotecnología Microbiana, Instituto de Investigaciones Químico-Biológicas, Morelia, Michoacán, México
Correspondence: Jesús Campos-García, Instituto de Investigaciones Químico-Biológicas, Universidad Michoacana de San Nicolás de Hidalgo, Edif. B-3, Ciudad Universitaria, CP 58030, Morelia, Michoacán, México. Tel./fax: +52 443 3265788; e-mail: email@example.com
Ethanol accumulation during fermentation contributes to the toxic effects in Saccharomyces cerevisiae, impairing its viability and fermentative capabilities. The iron–sulfur (Fe–S) cluster biogenesis is encoded by the ISC genes. Reactive oxygen species (ROS) generation is associated with iron release from Fe–S-containing enzymes. We evaluated ethanol toxicity, ROS generation, antioxidant response and mitochondrial integrity in S. cerevisiae ISC mutants. These mutants showed an impaired tolerance to ethanol. ROS generation increased substantially when ethanol accumulated at toxic concentrations under the fermentation process. At the cellular and mitochondrial levels, ROS were increased in yeast treated with ethanol and increased to a higher level in the ssq1∆, isa1∆, iba57∆ and grx5∆ mutants – hydrogen peroxide and superoxide were the main molecules detected. Additionally, ethanol treatment decreased GSH/GSSG ratio and increased catalase activity in the ISC mutants. Examination of cytochrome c integrity indicated that mitochondrial apoptosis was triggered following ethanol treatment. The findings indicate that the mechanism of ethanol toxicity occurs via ROS generation dependent on ISC assembly system functionality. In addition, mutations in the ISC genes in S. cerevisiae contribute to the increase in ROS concentration at the mitochondrial and cellular level, leading to depletion of the antioxidant responses and finally to mitochondrial apoptosis.
Saccharomyces cerevisiae is the most common yeast used in the ethanol-fermentation processes and, more recently, in biofuel production. Successful conversion of glucose to ethanol during fermentation depends mainly on the yeast's ability to counteract the stress factors during the process. Osmotic pressure caused by the initial sugar load, pH, and temperature and stress induced by fermentation products (such as acetaldehyde, acetic acid, methanol, ethanol, and phenolic compounds) constitute the main stressors in the fermentation medium. Although S. cerevisiae possesses inherent ethanol tolerance, ethanol accumulation during the fermentation processes has been described as the main producer of toxic effects (stressor) on yeast viability and fermentative capabilities. These toxic effects halt the fermentation process and reduce ethanol yield. Thus, the selection of yeast strains that tolerate different fermentation stresses is the next goal of the ethanol-production industry (Endo et al., 2008; Marks et al., 2008; Teixeira et al., 2009; Lewis et al., 2010; Stanley et al., 2010). Genomic, proteomic, and functional studies have been conducted to study the ethanol tolerance mechanisms, ethanol toxicity, and stress induced by other ethanol-fermentation-related metabolites. The main ethanol tolerance mechanisms involve changes in the composition of the membrane and cell wall (Chi & Arneborg, 1999; Ma & Liu, 2010), enhanced stress defenses, and general processes associated with the production and storage of protective compounds. Other adaptive changes involve the following: intracellular reorganization, biogenesis, transport processes occurring in vacuoles, peroxisomes, endosomes, cytoskeleton, and transcriptional machinery (Endo et al., 2008; Marks et al., 2008; Teixeira et al., 2009; Lewis et al., 2010; Ma & Liu, 2010; Stanley et al., 2010). However, the molecular mechanisms underlying these observations are poorly understood.
The iron–sulfur cluster (Fe–S) is a prosthetic group in many prokaryotes and eukaryote enzymes with redox, catalytic, and regulatory functions (Lill & Muhlenhoff, 2006; Lill, 2009; Lill et al., 2012). In bacteria, the Fe–S cluster is mainly synthesized by proteins encoded in the isc gene cluster (iron–sulfur cluster, ISC), although in some bacterial species the SUF and NIF systems are also involved in Fe–S cluster assembly under fixed metabolic conditions (Mühlenhoff et al., 2011). The ISC system has been extensively studied and is involved in the assembly of the Fe–S cluster into apoproteins in both bacteria and yeast (Schilke et al., 1999; Hoff et al., 2000; Dutkiewicz et al., 2003; Conte & Zara, 2011). In eukaryotes, two main systems of Fe–S-protein biogenesis have been described, the cytosol/nucleus (CIA) and mitochondrial (ISC) machineries, functionally the CIA depends on the mitochondrial ISC machinery (Lill & Muhlenhoff, 2006; Lill, 2009). In S. cerevisiae, the homologous genes to the bacterial iscSUA-hscBA-fdx cluster consist of the NFS1, ISU1, ISU2, ISA1, ISA2, JAC1, SSQ1, YAH1, GRX5, and IBA57 genes, which are dispersed in the yeast genome (Schilke et al., 2006; Lill et al., 2012). The ISC assembly machinery for maturation of all cellular Fe–S proteins (mitochondrial, cytosolic, and nuclear) is also involved in iron homeostasis in prokaryotes and eukaryotes (Lill, 2009; Lill et al., 2012). In yeast, the Ssq1p chaperone and the Jac1p J-protein (cochaperone) function together to assist in the biogenesis of Fe–S centers of Fe–S-dependent proteins. The Fe–S cluster assembly in mitochondria is initiated by cysteine desulfurase (Nfs-Isd11), which obtains a sulfur group from a cysteine transferred to the scaffold protein Isu1p and its redundant Isu2p protein assisted by Yah1p. This interaction also involves frataxin (Yfh1p), which acts as an iron donor or activity regulator. The Ssq1p chaperone has ATPase activity that is stimulated by the J-type cochaperone, Jac1p, during interaction with the scaffold protein Isu1p/Isu2p (Lill & Muhlenhoff, 2006; Lill, 2009; Lill et al., 2012). Isu1p is a substrate for both Ssq1p/Jac1p, and Jac1p and Isu1p cooperatively stimulate the ATPase activity of Ssq1p (Dutkiewicz et al., 2003). The subsequent cluster transfer to recipient apoproteins is mediated by the molecular chaperone Ssq1p and its cochaperone Jac1p and is assisted by glutaredoxin Grx5p (Rodríguez-Manzaneque et al., 2002). Recently, the participation of two novel proteins, Isa1p and Isa2p, was described in the yeast mitochondrial ISC assembly machinery; these proteins are involved in the maturation of mitochondrial apoproteins dependent on [4Fe-4S] clusters, and this activity is mediated by the assembly-protein, Iba57p (Mühlenhoff et al., 2011; Lill et al., 2012). Previous studies have suggested that the components of the ISC system could be related to ethanol tolerance in bacteria (Campos-García et al., 2000) and yeast (Yoshikawa et al., 2009; Ma & Liu, 2010). In S. cerevisiae, ethanol toxicity indicates that oxidative stress directly contributes to mitochondrial dysfunction (Costa et al., 1997; Drakulic et al., 2005); however, the underlying mechanisms remain to be elucidated. Therefore, the objective of the present study is to clarify the mechanism of ethanol toxicity and the involvement of the ISC assembly system in mitochondrial reactive oxygen species (ROS) generation.
Materials and methods
Yeast strains, growth conditions, and survival tests
The haploid S. cerevisiae BY4741 (Mat a, his3Δ, leu2Δ0, met15Δ0, ura3Δ0) and its KanMX4 interruption gene mutants, ssq1Δ, isu1Δ, isa1Δ, iba57Δ, and grx5Δ used as representatives of the ISC genes and the sdh2Δ, cox2Δ, and aco1Δ used as control mutants were obtained from Open Biosystem. Tubes or flasks were prepared with 10 or 50 mL of yeast extract/peptone/dextrose (YPD) culture medium and stressor (ethanol) was added at the indicated concentrations. Culture medium was inoculated with overnight yeast cultures that had reached an optical density of 0.1 at 600 nm (OD600 nm) and incubated at 30 °C with low-speed shaking (50 r.p.m.). Yeast growth (biomass) was monitored by measuring OD600 nm. Yeast survival test was carried out in yeast cultures grown on liquid YPD medium, collected in the late exponential growth phase and after adding ethanol 10% (v/v), incubating at 30 °C with low-speed shaking (50 r.p.m.). Cell survival was determined by trypan blue (Sigma) staining, and yeast counts were performed using a Neubauer chamber (Strober, 1997). For ethanol sensitivity or antioxidant protection, ascorbic acid or GSH was added to YPD media plates. Agar YPD plates with ethanol at indicated concentrations were inoculated with yeast cultures at several dilutions, sealed with Parafilm M and incubated at 30 °C.
Batch fermentation test
Fermentations were carried out in modified YPD medium, using glucose at 200 g L−1 (20%). Culture media (50 mL) were inoculated with overnight precultures grown in matching media to initial OD600 nm of 0.1; these were covered with cotton caps and incubated at 30 °C with lightly agitation. The samples collected at respective times were submitted to biomass determination by OD600 nm, and ROS determination by flow cytometry using the cell-permeant fluorescent probe 2′,7′-dichlorodihydrofluorescein diacetate (H2DCFDA; Molecular Probes, Invitrogen) as described later. Ethanol production and sugar consumption in the samples were determined by liquid chromatographic analysis as described (López-Alvarez et al., 2012). The fermented samples were filtered and then separated with an Aminex HPC-87Ca column (Bio-Rad). The operating conditions were as follows: column temperature, 80 °C; deionized water as mobile phase at 0.7 mL min−1 by 20 min running; injection volume, 25 μL. Quantitation was carried out by refraction index detector based on calibration plots using glucose and ethanol as standard compounds (Sigma-Aldrich), obtaining a linear coefficient of determination (R2) of 0.99 for each.
Real-time quantification of ROS in S. cerevisiae cultures and suspensions
Intracellular reactive oxygen species (cell-ROS) in yeast cultures or cell suspensions were determined using the oxidant-sensitive, cell-permeant fluorescent probe H2DCFDA (Molecular Probes, Invitrogen) (Du & Takagi, 2007; Kitagaki et al., 2007). H2DCFDA was used at 10 μg mL−1 in cell suspensions; the fluorescent cells and fluorescence intensity were quantified by flow cytometry as described below. For mitochondrial total ROS determination (mit-ROS), yeast suspensions were incubated with 5 μg mL−1 of dihydrorhodamine 123 (DHR123; Sigma) and fluorescence was quantified by flow cytometry. Dihydroethidium (Molecular Probes, Invitrogen) was used for the mitochondrial superoxide () determination, with a fluorescence excitation of 400 nm and an emission of 590 nm, which provided optimal discrimination of superoxide from other ROS. The probe was used at 5 μg mL−1 for flow cytometry as described (Madeo et al., 1999).
Cell cultures were grown to late exponential phase, samples (100 μL) were loaded with the respective fluorescent probe (H2DCFDA, DHR123, or dihydroethidium) and incubated at 30 °C for 2 h in the dark. Then, yeast cell samples were taken to 1 mL with PBS buffer (NaCl 137 mM, KCl 2.7 mM, Na2HPO4·2 H2O 8.1 mM, KH2PO4 1.76 mM, at pH 7.4) and the ROS were immediately quantified by flow cytometry. For ethanol treatments, yeast cultures grown on YPD medium (10 mL) were added with ethanol (10% v/v), and at the respective times, samples (100 μL) were harvested, washed, and suspended in PBS adjusting the volume to 1 mL or 1 × 107 cells mL−1. The cells were loaded with the fluorescent probes by incubating for 2 h. Finally, the fluorescence was determined by flow cytometry using a BD Accuri C6 Flow Cytometer (BD Biosciences). The populations of cells for each of the treatments were gated in the forward scatter and side scatter (SSC) dot plots to eliminate dead cells and cell debris. Populations corresponding to auto- or basal fluorescence were located in the left quadrant and cells with emission of fluorescence increased at least one log unit value were located in the right quadrant from the dot plots. In addition, the percentage of fluorescent cells (PFC) and the median fluorescence intensity (FI) were determined in the monoparametric histograms of fluorescence emission obtained from the dot plots and labeled as percentage of cells and as relative units of fluorescence. The equipment was calibrated using Spherotech 8 peak (FL1-FL3) and 6 peak (FL-4) validation beads (BD Accuri). Fluorescence of H2DCFDA and DHR123 probes was monitored in the emission fluorescence channel FL1 (533/30 nm) and for dihydroethidium probe in the FL2 (587/40 nm). A minimum of 20 000 cellular events were analyzed in each determination point.
Isolation of mitochondria
Mitochondria of the S. cerevisiae yeast and mutant strains were isolated from cultures grown in liquid medium YPD at 30 °C with shaking. Cells were harvested in late exponential growth phase by centrifugation at 2750 g for 15 min at 4 °C (Centrifuge Eppendorf 5810R, rotor F-34-6-38) and washed thrice using distilled water and suspended in buffer (Tris-HCl 50 mM, pH 7.5). Yeast spheroplasts were obtained as described before (Auchere et al., 2008), using a digestion solution (sorbitol 1.2 M, EGTA 1 mM, Tris-HCl 50 mM, DTT 10 mM, at pH 7.5) and adding Lyticase from Arthrobacter luteus (Sigma) at 2 mg g−1 cells by fresh weight. Yeast suspensions were incubated by 60 min at 30 °C. Spheroplasts were washed twice with spheroplast wash buffer (sorbitol 1.2 M, Tris-HCl 50 mM, at pH 7.5). Spheroplasts were suspended in a homogenizing buffer (sorbitol 0.6 M, HEPES-KOH 20 mM, at pH 7.4) and lysed in a potter tube homogenizer and washed thrice with the same buffer. The unruptured cells were removed by centrifugation at 2500 g for 10 min at 4 °C, and yeast mitochondria were harvested from the supernatant by centrifugation at 9600 g for 10 min to 4 °C.
Determination of glutathione content in complete cells and mitochondria
Yeast cells cultured in YPD to late exponential growth phase were used for GSH/GSSG determination and for mitochondria isolation. Cell pellets (1 mg) were suspended in buffer (K2HPO4 500 mM, EDTA 5 mM, sulfosalicylic acid 5%, pH 7.0). Cells suspensions were frozen with liquid nitrogen and broken in a mortar. Cell extracts were centrifuged at 10 000 g for 10 min at 4 °C, and supernatants obtained were used for GSH/GSSG determinations. Mitochondria (1 mg) were suspended in homogenizing buffer (sorbitol 0.6 M, HEPES-KOH 20 mM, at pH 7.4) and added sulfosalicylic acid 5% (1 : 3 proportion). Mitochondria were broken by vortexing with frozen and thawing cycles. Extracts were centrifuged at 10 000 g for 10 min at 4 °C, and the supernatants obtained were used for GSH/GSSG determinations. The reduced GSH and oxidized GSSG species were quantified using the glutathione assay kit (Sigma-Aldrich), based on the established enzymatic recycling method previously described using DTNB (5,5′-dithiobis-2-nitrobenzoic acid) in reaction buffer (K2HPO4 500 mM, EDTA 5 mM, DTNB 3 mM, glutathione reductase 0.1 U mL−1, NADPH 2 mM, pH 7.0) to produce the yellow colored compound 5-thio-2-nitrobenzoic acid (TNB) (Shaik & Mehvar, 2006). Quantification of TNB was carried out measuring the absorbance at 412 nm. For GSSG determination, the extracts were incubated 1 h with 4-vinylpyridine (2%) previous to the quantification as described earlier.
Determination of catalase activity
Catalase activity was quantified by oxygen production rate using H2O2 (10 μM) as substrate, cells (25 mg) from S. cerevisiae were placed in 2.5 mL of MES-TEA buffer (pH 6.0) in a sealed glass chamber with constant stirring, quantifying oxygen generation rate with a Clark-type oxygen electrode coupled to a biological oxygen monitor (Zigman et al., 1998).
Determination of cytochrome c integrity
Cytochrome c integrity was determined by detecting the release of cytochrome c from mitochondria using immunocytochemistry in S. cerevisiae spheroplasts obtained as described with minor modifications (Ng et al., 2012). Spheroplast suspensions (1 × 106 cells) were incubated in permeabilization buffer (100 mM KCl, digitonin 50 μg, sorbitol 1.2 M, Tris-HCl 50 mM, at pH 7.5) for 5 min at room temperature and fixed with 3.5% formaldehyde in buffer (sorbitol 1.2 M, Tris-HCl 50 mM, at pH 7.5) for 30 min at room temperature. Excess formaldehyde was removed by washing with PBS buffer and centrifugation. Permeabilized spheroplasts were incubated in 100 μL of blocking media (3% BSA and 0.05% Tween 20 in PBS) for 30 min at room temperature, washed, and incubated with primary anticytochrome c antibody (mouse monoclonal anti-7H8, Santa Cruz Biotechnology, Inc.) at a 1 : 200 dilution in blocking medium at 4 °C overnight. Spheroplasts suspensions were washed twice with PBS and incubated with the secondary antibody, a monoclonal anti-mouse coupled to rhodamine goat anti-mouse IgG (Invitrogen) in blocking medium at a 1 : 200 dilution for 2 h, at room temperature. Finally, spheroplasts were washed twice with PBS and suspended in PBS for flow cytometry analysis. Cytochrome c integrity was quantified by measuring the fluorescence produced in a flow cytometer (BD Accuri C6; BD-Bioscience). The populations of cells were gated in the forward SSC, the fluorescence was quantified in FL1 channel (530 nm) dot plots, and PFC determination into monoparametric fluorescence histograms was carried out as described previously.
Ethanol toxicity in S. cerevisiae yeast is increased by mutations in ISC genes
We used haploid yeast S. cerevisiae ISC mutants to study the involvement of the ISC gene-cluster products on ethanol tolerance. Plates with solid YPD medium were incubated with ethanol in the range of 0–10% (v/v) and used to test the ethanol tolerance in S. cerevisiae ISC mutants (SSQ1, ISU1, ISA1, IBA57, and GRX5 genes). In addition, three mutants not involving the ISC system were used as controls, the mutant sdh2Δ (affected in succinate dehydrogenase that forms the complex II from the electron transporter chain, ETC), the cox2Δ (affected in cytochrome c oxidase subunit II of the complex IV from the ETC), and aco1Δ (affected in cis-aconitase, which is involved in the tricarboxylic acid cycle, TCA cycle). The yeast mutants ssq1∆, isa1∆, and iba57∆ showed impaired growth in YPD with ethanol compared with wild type (WT), while the isu1∆ and grx5∆ mutants, as well as the control mutants sdh2Δ, cox2Δ, and aco1Δ showed a similar behavior to the WT (Fig. 1a). With ethanol treatment, using dose–response growth analysis in liquid YPD medium, similar results of ethanol toxicity were observed (Fig. 1b). Under these conditions, the half-maximal inhibitory concentration for ethanol (IC50EtOH) was calculated for each strain. IC50EtOH values were as follows: WT, 8.4 ± 0.3%; isu1∆, 7.9 ± 0.1%; ssq1∆, 6.8 ± 0.1%; isa1∆, 7.1 ± 0.1%; iba57∆, 6.7 ± 0.2%; and grx5∆, 7.4 ± 0.2%; and the control mutants, sdh2Δ, cox2Δ, and aco1Δ, showed (IC50EtOH) values of 8.1 ± 0.2%, 7.6 ± 0.2%; and 8.2 ± 0.2%, respectively. The results indicate that the mutants ssq1∆, isa∆1, and iba57∆ were more sensitive to ethanol (c. 15–20%) than the WT strain or the control mutants. This phenotype was confirmed with survival assays of yeast cultures obtained in the late exponential phase, when cells were treated with a toxic ethanol concentration (10%). WT yeast was able to efficiently tolerate 10% (v/v) ethanol for 6 h similarly to the control mutants sdh2Δ and aco1Δ, showing up to 80% cell survival, lesser extent the cox2Δ with 70% cell survival, while the ssq1∆, isu1Δ isa1∆, grx5Δ, and iba57∆ mutants displayed a drastic decrease in cell survival, which was below 50%. These differences in survival were statistically significant (P <0.05) with respect to WT yeast. Under these assay conditions, the lethal time 50 at 10% (v/v) ethanol concentration [Lt50EtOH(10%)] obtained for the ssq1∆, isa1∆, and iba57∆ mutants were 42–55% lower than WT, with the following values: WT, 11.0 ± 1.5 h; aco1Δ, 10.5 ± 1.5 h; sdh2Δ, 9.5 ± 1.5 h; cox2Δ, 9.5 ± 1.2 h; isu1∆, 7.1 ± 1.8 h; grx5∆, 6.6 ± 1.8 h; isa1∆, 4.8 ± 1.7 h; ssq1∆, 4.1 ± 1.6 h; and iba57∆, 3.7 ± 1.4 h. Thus, the ssq1∆, isa1∆, and iba57∆ mutant yeast strains were the most susceptible to ethanol. Additional tolerance assays were conducted to explore the growth rate in the presence of toxic ethanol concentrations (Fig. 2). The growth kinetic results indicated that for all the ISC mutants tested, the growth was impaired at 8% ethanol, resulting in the following velocity parameters (μ): WT, 0.075 ± 0.008; sdh2Δ, 0.076 ± 0.006; cox2Δ, 0.074 ± 0.006; aco1Δ, 0.076 ± 0.008; isu1∆, 0.068 ± 0.009; grx5∆, 0.055 ± 0.01; isa1∆, 0.053 ± 0.01; iba57∆, 0.047 ± 0.01; and ssq1∆, 0.043 ± 0.01. At a concentration of 10% ethanol, all the ISC mutants and cox2Δ showed a major impaired growth than the sdh2∆ or aco1∆ mutants and the WT (Fig. 2b and c). The ethanol tolerance, survival, and growth rate parameters for the yeast strains under ethanol treatment indicated that the ssq1∆, isa1∆ and iba57∆, and to a lesser degree the grx5∆ and isu1Δ mutants, were more susceptible to ethanol than WT or the control strains, aco1∆ and sdh2∆. These data indicate that mutations in the ISC genes result in an impaired ethanol tolerance in S. cerevisiae, although this effect was less dramatic in the isu1∆ mutant likely due to complementation of the homologous ISU2 gene. It is worth noting that the assays to test this hypothesis could not be carried out due to the lack of a viable phenotype for the haploid jac1∆ mutant.
Increase in ROS generation, induced by ethanol, in S. cerevisiae is exacerbated by ISC genes mutations
The toxic effects caused by ethanol in S. cerevisiae have been associated with an increase in oxidative stress, inactivation of related enzymes, and dysfunctional mitochondrial metabolism (Costa et al., 1997; Kitagaki et al., 2007). However, little is known about the mechanism involved in ROS generation, location/accumulation of ROS, and the kind of species generated by the presence of ethanol. We hypothesized that the protein components involved in Fe–S assembly and the production of ROS induced by ethanol could be related.
To elucidate whether the increased susceptibility to ethanol in S. cerevisiae is related to an increased ROS generation and whether this is associated with the ISC assembly system, real-time quantification of ROS by flow cytometry was carried out using fluorescent ROS indicators. Three fluorescent ROS indicators were used: H2DCFDA for general intracellular ROS, the DHR123 probe for mitochondrial ROS (this compound reacts mainly with H2O2 produced by mitochondrial metabolism), and dihydroethidium to identify mainly intracellular superoxide radicals (). As described previously, all yeast strains were able to tolerate 10% (v/v) ethanol for at least 6 h with up to 50% cell survival (Fig. 1c). Therefore, real-time ROS quantification was carried out at 10% ethanol during the period of 6 h.
First, we determined ROS generation in the yeast population (PFC) and the fluorescence intensity (FI) in cultures treated with or without 10% ethanol. Using the H2DCFDA and DHR123 probes, a similar ROS production behavior was observed, the PFC values in yeast suspensions without ethanol treatment were around 10–40%, while with ethanol treatment the fluorescent yeast population increased to 60–90%, showing significant differences in all the ISC mutant strains tested, with respect to the untreated ethanol suspensions (Fig. 3b and c); this effect was more apparent when dihydroethidium probe was used (Fig. 3d), indicating that in the yeast population treated with ethanol, the generation of ROS occurred. Interestingly, the highest PFC values were observed in the ssq1∆, isa1∆, iba57∆, and grx5∆ mutants (Fig. 3a–d). The unrelated ISC mutants, aco1∆ and sdh2∆, showed a behavior similar to the WT yeast, while that the cox2∆ strain showed a similar behavior to grx5∆ mutant, except for the dihydroethidium probe which was similar to WT.
Additionally, we determined the fluorescence intensity in yeast suspensions treated with ethanol; this parameter may be directly related to the amount of ROS produced in cell suspensions. Our results clearly indicate that the ISC mutations caused an increase in ROS generation with respect to WT yeast (Supporting Information, Fig. S1). Interestingly, the ethanol treatment caused exacerbated ROS generation in all the ISC mutants (Figs 3d and S1). Using the H2DCFDA probe in the ISC mutants (ssq1∆, isa1∆, iba57∆, and grx5∆), we found that the FI values were statistically increased, except for the isu1∆ mutant, compared to control strains (WT, sdh2∆, and aco1∆). Importantly, the FI values in the ssq1∆, isa1∆, iba57∆, and grx5∆ yeast suspensions were c. 10–20-fold higher with ethanol treatment (Fig. S1). In addition, the IF values observed using the DHR123 and dihydroethidium indicators showed behaviors similar to those of the H2DCFDA probe and the differences were statistically significant for the ssq1∆, isa1∆, iba57∆, and grx5∆ mutants, with ethanol addition (Figs 3 and S1). These results indicated that ethanol treatment causes an increase in ROS generation in S. cerevisiae and that this effect is enhanced in ISC gene mutants, confirming that the mechanism of ethanol toxicity involves ROS generation. In addition, the results show that the main ROS generated during ethanol treatment were hydrogen peroxide and superoxide, which increased significantly in the SSQ1, ISA1, IBA57, and GRX5 gene mutations – but increased to a lesser degree or not at all in the isu1∆ mutant. To validate the ROS generation experiments, three control mutants were incorporated in the study, the first one corresponds to the cis-aconitase mutant (aco1∆) and was used as a negative control. As mentioned earlier, it is part of the TCA cycle and should not be related to ROS generation. The mutants sdh2∆ and cox2∆ have a deletion of succinate dehydrogenase of the complex II and cytochrome c oxidase of the complex IV of the ETC, respectively. The results obtained with these yeast strains indicated that the aco1∆ and sdh2Δ mutants did not show significant differences in the ROS levels with respect to the WT yeast (Figs 3 and S1), while the ROS production of the cox2∆ mutant was similar to that of grx5∆ mutant when H2DCFDA and DHR123 probes were used, but not with dihydroethidium, where the ROS contents were similar to those observed for the WT and the mutants aco1∆ and sdh2Δ. The findings obtained with the fluorescent ROS indicators, DHR123 and dihydroethidium, suggest that ROS are mainly generated and accumulate in mitochondria after the treatment with toxic ethanol concentrations, and an additive effect of ROS generation was observed in the ISC gene mutants.
Increases in ROS correlate with maximal ethanol accumulation during the fermentation process and is exacerbated by ISC genes mutations
As mentioned, the addition of ethanol at toxic concentrations provoked a ROS burst mainly generated at the mitochondria level and this burst is exacerbated by dysfunction of the ISC assembly system. However, it has also been proposed that the ethanol yield during the fermentation process is correlated with the ethanol tolerance level of the yeast used. Therefore, real-time ROS determination was carried out in yeast culture during the fermentation assay, using a sugar concentration sufficient to generate a toxic amount of ethanol in the fermentation medium. Fermentative capabilities of the ISC mutants were tested in YPD liquid medium containing 200 g L−1 of glucose (theoretically, this glucose amount produces c. 13% ethanol v/v, under a 100% yield). We have observed that at 240–280 g L−1 sugar concentrations, an arrest of ethanol production occurs at c. 10–11% in the S. cerevisiae BY4741 strain, and the rest of the glucose remains in the fermentation medium. These observations led us to hypothesize that in the S. cerevisiae BY4741 strain, the limit of ethanol yield may be related to its level of ethanol tolerance. Thus, we proposed that 10% ethanol might cause inhibition of yeast fermentative capability, likely due to generalized damage caused by the ROS increase. Therefore, ROS generation was determined in yeast cells harvested throughout the fermentation process (Fig. 4). Our findings indicated that all the yeast strains tested were able to grow efficiently in YPD supplemented with 200 g L−1 of glucose with a similar biomass determined at OD (600 nm) (Fig. 4a). Additionally, the glucose consumption rate and ethanol yield indicated that all the yeast used in this study were able to transform glucose into ethanol with similar yield rates (Fig. 4b). Interestingly, the major glucose consumption and ethanol-production rates among the strains occurred in the first 90 h of fermentation, whereas at 168 h of fermentation, the glucose was totally consumed in all the strains and the maximum ethanol yield was reached and corresponded to a concentration of c. 10–11% (v/v) (Fig. 4b). Thus, in this genetic background (S. cerevisiae BY4741 strain), no significant differences in ethanol yield between the tested strains were found at the end of the test, suggesting that fermentative capabilities were not affected significantly in the ISC mutants, at least not when 200 g L−1 of glucose was utilized.
When the ROS levels were determined using the general ROS-probe H2DCFDA in cells harvested during the fermentation assay, the results indicate that in the first 24–72 h of fermentation, when the ethanol concentration was c. 2–7%, the ROS level in the ISC mutant strains were less or similar to WT yeast, showing a moderate fluorescence intensity (500–2000 relative units of fluorescence, RU). In some ISC mutants, the intensity was even lower (Fig. 4c, 24−72 h). Importantly, at 96 h of fermentation, when ethanol accumulated to c. 9%, ROS generation was increased significantly in only the ssq1∆, isa1∆, and aco1∆ mutants, but at 120 h, this behavior was also observed in the iba57∆ mutant. Interestingly, at 144 h of fermentation, the ROS levels increased c. 5-fold (when the concentration of ethanol was c. 10%); ROS generation increased showing a maximal value of fluorescence at c. 15 000–20 000 RU in the ssq1∆ and iba57∆ mutants, being significantly higher than the WT, except for the isu1∆ and grx5Δ mutants. These results showed that the ssq1∆, isa1∆, and iba57∆ mutants produced higher ROS levels during fermentation, while isu1∆ and grx5Δ showed a similar behavior to the WT strain (Fig. 4c). As expected, the control mutants of unrelated genes, sdh2Δ, aco1∆, and cox2Δ, showed ROS levels similar to WT. Additionally, we observed that at 120–144 h of fermentation (at an ethanol concentration of 9–10%), the WT strain maintained stabilized ROS levels (4000–5000 RU of fluorescence), while in the ISC mutants, ROS levels were increased 2–4-folds (Fig. 4c). These results support the hypothesis that ethanol accumulation during the fermentation process provokes a significant increase in ROS levels, and that this increase is further pronounced by the ISC gene mutations. The ROS increase in cells, when ethanol in the medium accumulated to 9–10%, suggests that ROS are responsible for the fermentation halt, confirming that the maximal ethanol yield obtained in the fermentation process is dependent on the level of ethanol tolerance of the yeast strain. The findings also suggest that although the ISC assembly system in S. cerevisiae is related indirectly to ROS generation also contributes to ethanol yield and susceptibility.
GSH/GSSG balance is disrupted by ethanol treatment
To determine whether the increased ROS generation caused by ethanol treatment or accumulation in the fermentation medium is related to the mechanism of toxicity, two markers of antioxidant defense were measured; the cellular GSH/GSSG ratio and catalase activity. Our results indicated that in cells of the ssq1∆, isa1∆, grx5∆, and iba57∆ mutants, without ethanol treatment, GSSG was significantly increased, while the GSH content was considerably diminished. The GSH/GSSG ratio clearly indicated that these mutants had a glutathione imbalance (Fig. 5). When the yeast cells were treated with ethanol, a similar result was observed for the ssq1∆, isa1∆, grx5∆, and iba57∆ mutants; a slight increase in GSSG was observed in cells treated with ethanol (Fig. 5). As it has been reported that the mitochondria is a glutathione reservoir that contributes to cell-ROS scavenging, we determined the GSSG and GSH contents in mitochondria isolated from yeast cells. Our results with isolated mitochondria from yeast cultures grown without ethanol did not show any significant differences in GSSG content, but a decrease in the content of GSH was observed in the ssq1∆, isa1∆, and iba57∆ mutants. These data showed that the GSH/GSSG ratio was significantly decreased in the ssq1∆, isa1∆, and iba57∆ mutants (Fig. 5). When mitochondria were isolated from yeast cultures treated with ethanol (10% for 3 h) and used for GSH/GSSG determination, the behavior was similar to mitochondria isolated from cells without ethanol treatment, but the differences observed were higher and significant in the ssq1∆, isa1∆, grx5∆, and iba57∆ mutants (Fig. 5). Interestingly and as expected, the control mutants aco1∆, sdh2∆, and cox2∆ showed similar GSH/GSSG ratios than WT, except for cox2∆ under ethanol treatment, these values were similar to those of the grx5∆ mutant.
Catalase constitutes one of the main antioxidant defenses in cells; therefore, the activity of this enzyme was measured in yeast cells subjected to ethanol treatment. Our results showed that catalase activity increased significantly in the control mutants sdh2∆ and cox2∆, also as in the ssq1∆, isa1∆, grx5∆, and iba57∆/mutants – except for the isu1∆ mutant – with respect to WT yeast or the control strain aco1∆ (Fig. 6a). While exposure to ethanol induced a significant increase in catalase activity only in the WT yeast and the aco1∆ mutant, catalase activity remained at similar levels in the other ISC mutants, except for the isu1∆ mutant (Fig. 6a).
To protect yeast cells against the toxic effects of ethanol (10%), ROS scavengers such as ascorbic acid and GSH were added to the medium. Our findings indicated that these antioxidants were able to protect the WT and the ISC mutants against the toxic effects of ethanol (Fig. 6b). These data also indicated that ethanol exposure is indeed responsible for ROS generation and that this effect is mainly exacerbated in the ssq1∆, isa1∆, and iba57∆ mutations. These results also suggest that ROS production, induced by ethanol, is occurring mainly in the mitochondria, more likely at the ETC level, and that this effect is increased in an additive manner in mutants with a dysfunctional ETC. These results confirm that the mechanism of toxicity of ethanol involves a burst of ROS production, and that this effect may be due to an imbalance of the GSH/GSSG ratio brought about by ethanol treatment. Importantly, the induction of catalase, as well as other cellular antioxidant defenses, was significantly induced in the WT yeast treated with ethanol, while in the ISC mutants, catalase expression increased in the absence and presence of ethanol.
Ethanol triggers mitochondrial apoptosis
We monitored mitochondrial oxidative damage by measuring cytochrome c (Cytc) release. Assays of Cytc integrity, in yeast monitored in the absence of ethanol, indicated that after 6 h, Cytc release (c. 10–20% of the yeast population) occurred in all the mutants. However, when the yeast cultures were treated with ethanol (10%), an increase in Cytc release was observed in all strains. Cytc release was higher in the ssq1∆, isa1∆, and iba57∆ mutants (c. 70% of the yeast population), while for the isu1∆ and grx5∆ mutants, the effect was similar to the control mutants aco1∆, sdh2∆, and cox2∆ (Fig. 7). These findings indicate that exposure of S. cerevisiae to toxic concentrations of ethanol triggers a mitochondrial apoptosis response and that this effect is increased in dysfunctional mitochondria, which in this instance could be caused indirectly by a dysfunctional ISC system probably associated with a defective incorporation or recycling of Fe–S clusters.
Several genes involved in stress caused by ethanol have been described in S. cerevisiae (Marks et al., 2008; Teixeira et al., 2009; Lewis et al., 2010; Stanley et al., 2010). These studies have reported changes in the expression of genes related to intracellular organization, biogenesis, and transport processes, as well as changes in vacuole function, ion homeostasis, signaling, lipid metabolism, and energy reserve. In addition, transcriptomic approaches, using wine yeast, have described the participation of several groups of molecular chaperones that are also related to global heat-shock stress responses (Marks et al., 2008). Alternatively, studies of ethanol toxicity have found that ethanol may produce ROS via generation of the radical hydroxyethyl (CH3-CH2-OH−), but the generation of this unstable intermediate is inefficient, and therefore, the compound is expeditiously counteracted by antioxidant cellular mechanisms before it can cause cellular damage (Albano et al., 1996).
Iron metabolism has also been linked to the mechanism of ethanol toxicity. It has been suggested that the ISC system, which is essential for Fe–S cluster biogenesis and recycling, is involved in ethanol tolerance and the preservation of mitochondria functionality (Schilke et al., 1999). The ISC gene-cluster protein products in S. cerevisiae carry out their function in the mitochondrial matrix and are involved in the formation/assembly or repair of Fe–S clusters of numerous apoenzymes (see Lill, 2009; Lill et al., 2012). Specifically, mutations in the bacterial gene hscA (Campos-García et al., 2000) or the genes SSQ1 (Schilke et al., 1999) and ISA1 (Yoshikawa et al., 2009) in S. cerevisiae have been proposed to impair ethanol tolerance and to induce a deficient growth in its presence. Similarly, the gene GRX5 has been related to protection against oxidative damage (Rodríguez-Manzaneque et al., 1999). Therefore, we investigated the mechanism of toxicity caused by ethanol and the involvement of the protein components implicated in mitochondrial Fe–S assembly.
Ethanol treatment during dose–response growth and survival assays showed that the S. cerevisiae mutants ssq1∆, isa1∆, and iba57∆ were impaired in their growth (IC50EtOHc. 15–20% lower) and survival (less than 50% survival) as compared to control WT yeast. However, this phenomenon was observed in lesser proportion in the isu1∆ or grx5∆ mutants (Fig. 1). These results suggest that the SSQ1, ISA1, and IBA57 gene products may be correlated with ethanol tolerance level in S. cerevisiae. The minimal effects in isu1∆ and grx5∆ mutants suggest that complementation of their function may take place by their paralog genes, being not essential for the ISC assembly system.
Based on these results, and using other known ROS generators such as acetaldehyde (data not shown), it was concluded that the growth of S. cerevisiae ISC mutants is clearly impaired (Fig. 2). The mechanism of ethanol toxicity may be related to the ROS generation promoted by its metabolite acetaldehyde, which has been described as a strong ROS generator (Novitskiy et al., 2006). This finding suggests that the ISC gene cluster indirectly contributes to both ROS-generating compounds and ethanol toxicity. Thus, the main mechanism of ethanol toxicity may be mediated by ROS generation as proposed previously (Kitagaki et al., 2007; Farrugia & Balzan, 2012).
Using three different and selective fluorescent probes for intracellular ROS determination, we showed that ROS were increased with or without ethanol treatment in the S. cerevisiae ISC mutant strains. Interestingly, higher fluorescence values were observed in the ssq1∆, isa1∆, grx5∆, and iba57∆ mutants (Figs 3 and S1). These results confirm that ROS generation by the mitochondrial ECT is dependent on the functionality of the ISC assembly system, because ROS levels were exacerbated under ethanol treatment. In addition, the findings indicate that the main ROS species generated during ethanol treatment were hydrogen peroxide and superoxide (although hydroxyl ions, not specifically measured in the study, may play an important role as well). The results also indicate that ROS are generated and accumulated at the mitochondrial level and are increased in an additive manner in the SSQ1, ISA1, GRX5, and IBA57 mutants by the presence of ethanol. In addition, it has been reported that when the ETC is dysfunctional, ROS generation increases (Kitagaki et al., 2007). In this context, the mutants sdh2∆ and cox2∆, deficient in complex II and IV of the ETC, respectively, their ROS levels were lower than those observed in the ISC mutants. Additionally, the complex III of the ETC has been described as an important cellular component in superoxide generation by electron leakage. In the mutant cox2∆, blocked in the complex IV, an increase in superoxide generation would be expected. Interestingly, in the ISC mutants, the superoxide generation was superior to that observed in the cox2∆ mutant, contrary to what was anticipated. Our results demonstrated that when the ETC is blocked (petite mutations), an increment in ROS levels occurred, which was more exacerbated in the ssq1∆, isa1∆, and iba57∆ mutants, than in the sdh2∆ and cox2∆ strains. These results indicate that in the ISC mutants, dysfunction in the ETC produces an increase in ROS which is exacerbated by ethanol accumulation, and this phenomena may be caused by defective incorporation of [2Fe-2S] or [4Fe-4S] clusters in the mutants, rather than by ETC blocking. The failure of the cellular antioxidant responses to efficiently neutralize the ROS generated may be lethal for the yeast cell.
Successful conversion of glucose to ethanol during fermentation depends mainly on the yeast's ability to counteract the stress factors during the process. Ethanol accumulation during fermentation has been described as one of the main stressors, affecting the viability and fermentative capabilities of yeast, provoking a halt of the fermentation process and a reduction in ethanol yield. Although S. cerevisiae has inherent ethanol tolerance, the selection of yeast strains with superior ethanol tolerance levels is an important goal of the ethanol-production industry (Endo et al., 2008; Marks et al., 2008; Teixeira et al., 2009; Lewis et al., 2010; Stanley et al., 2010). The findings of this study indicate that the addition or accumulation of a toxic ethanol concentration (10%) provokes a ROS burst mainly at the mitochondria level and this increase in ROS is exacerbated by dysfunction of the ISC assembly system. Therefore, the generation of ROS during the fermentation process and the effects of the mutations in the ISC genes on the generation of ROS were evaluated. We carried out real-time ROS determination in yeast cultures during fermentation assays, using a concentration of sugar sufficient to generate a toxic level of ethanol (over 10% v/v). We have previously determined that sugar concentrations in the range 240–280 g L−1 cause an arrest in ethanol production at c. 10–11% in the S. cerevisiae strain BY4741 and thus hypothesized that for the S. cerevisiae BY4741 strain, the limit of ethanol yield may be limited by the level of ethanol tolerance of the strain. Our findings indicate that all the yeast strains tested were able to grow efficiently in YPD supplemented with 200 g L−1 of glucose, and that the rate of glucose consumption and ethanol yield was similar in the mutants. Interestingly, at 168 h of fermentation, the glucose was totally consumed in all strains and the maximal ethanol yield was reached, corresponding to c. 10–11% (v/v) (Fig. 4a–b), suggesting that the fermentative capabilities of the ISC mutants were not significantly affected. The ROS levels determined during the fermentation assay, using the general ROS-probe H2DCFDA, indicate that in the first 24–72 h of fermentation, when the ethanol concentration was c. 2–7%, the ROS levels in the ISC mutants were lower or similar to WT, suggesting that the ROS generated are counteracted by the antioxidant cellular responses (Fig. 4c). Interestingly, at 96–144 h of fermentation, when ethanol was c. 9–10%, ROS generation was significantly increased in the ssq1∆, isa1∆, and iba57∆ mutants. At 144 h of fermentation (ethanol yield 10–11%), the maximal ROS generation was reached in the yeast cultures, the ROS levels were significantly higher in the ssq1∆, isa1∆, and iba57∆ mutants, and as expected, the control mutants aco1Δ, sdh2Δ, and cox2Δ had ROS levels similar to WT. We observed that between 24–144 h of fermentation, the WT yeast and control mutants maintain low levels of ROS (1000–4000 RU of fluorescence), but in the ssq1∆ and iba57∆ mutants, the ROS levels were around 5-fold higher than those of the WT (Fig. 4c). These results support the hypothesis that ethanol accumulation during the fermentation process provokes a significant increase in ROS levels, which is more pronounced in the ISC gene mutants. The ROS increase in the fermentation medium, when ethanol accumulated to 9–10%, suggests that it is responsible for the fermentation halt, confirming that the maximal ethanol yield obtained in the fermentation process is dependent on the level of ethanol tolerance of the yeast strain. The findings also suggest that although the ISC assembly system in S. cerevisiae is indirectly related with ROS generation, it also indirectly contributes with ethanol yield and susceptibility.
Glutathione is considered one of the most important cell antioxidants and exists in reduced (GSH) or oxidized (GSSG) states (Farrugia & Balzan, 2012). Glutathione has an important role in detoxifying reactions such as scavenging cellular hydrogen peroxide. Glutathione is predominantly reduced under normal physiologic conditions due to the action of NADPH-dependent GSSG reductase. However, under some conditions, a moderate level of oxidative stress occurs, causing a decrease in GSH and an increase in GSSG, modifying the GSH/GSSG ratio (Drakulic et al., 2005). Under severe oxidative stress, the ability of the cell to reduce GSSG to GSH may be over-loaded, leading to an accumulation of GSSG. Therefore, mitochondrial GSH depletion leads to increased levels of ROS and mitochondrial dysfunction (Farrugia & Balzan, 2012). The results obtained from our glutathione content measurements indicated that the GSSG content was significantly increased, while that GSH decreased, and that the GSH/GSSG ratio decreased at both the cellular and mitochondrial levels in the ISC mutants (ssq1∆, isa1∆, grx5∆, and iba57∆), which is consistent with the oxidative stress observed under ethanol treatment (Fig. 5).
In yeast cells, the excess of ROS is removed to restore the redox balance. ROS induce both mitochondrial and cytosolic superoxide dismutase, which scavenges superoxide to convert it to H2O2, which can be detoxified by cytosolic and peroxisomal catalases (Drakulic et al., 2005; Farrugia & Balzan, 2012). Catalase activity is up-regulated during oxidative stress as an antioxidant response to counteract the hydrogen peroxide increase. The activity of the antioxidant enzyme catalase in yeast was measured to confirm the cellular oxidative stress status. The activity assays showed that catalase activity was increased in WT yeast when treated with ethanol (Fig. 6a). Interestingly, ISC mutants showed increased catalase activity both with and without ethanol treatment. This result suggests that deletion of the ISC genes promotes a sustained increase in ROS generation, which in turn induces catalase expression, confirming that H2O2 is produced by the toxicity mechanism of ethanol. Importantly, according to our data, after prolonged exposure to ethanol (6 h at 10% ethanol, data not shown), the increase in oxidative stress and depletion of antioxidants may cause the inactivation of catalase (indicating that this enzyme is susceptible to ROS) and affects the yeast mitochondrial target proteins, α-ketoglutarate dehydrogenase, pyruvate dehydrogenase, and the iron–sulfur [4Fe-4S] enzymes aconitase and succinate dehydrogenase (Farrugia & Balzan, 2012). Additionally, the toxicity in yeast cultures exposed to ethanol was counteracted by the addition of ROS scavengers such as ascorbic acid and GSH, confirming that the toxicity of ethanol was due to ROS generation and that the ROS level is dependent on the ISC system integrity and the antioxidant response.
The depletion of GSH in the mitochondria may lead to mitochondrial dysfunction and this organelle has also been described as an important trigger of the apoptotic pathway (Drakulic et al., 2005; Farrugia & Balzan, 2012). During conditions of mitochondrial dysfunction, the fragmentation and release of mitochondrial factors are widely accepted as the initiating events in the apoptotic pathway. Factors activating caspase-dependent pathways include the cytochrome c of the electron transport chain, which plays an essential role in mitochondria-dependent apoptotic death. The release of cytochrome c has been described in S. cerevisiae as an indication that the ROS produced by ethanol treatment induce mitochondrial-dependent apoptosis; however, those experiments were assayed under extremely high ethanol concentrations (21% and 23%) (Kitagaki et al., 2007). In agreement with those results, in the present study, the cytochrome c release occurred at 10% ethanol after 3–6 h of treatment (conditions that commonly occur during fermentation processes). The release of cytochrome c was exacerbated in the ISC mutants, indicating that an apoptotic response occurs following ethanol and ROS accumulation, affecting the viability and survival of the yeast.
In conclusion, our results indicate that the mechanism of ethanol toxicity includes a generalized increase in oxidative stress, in agreement with ethanol accumulation during batch fermentation, and that hydrogen peroxide and superoxide are the main ROS generated. Consequently, this oxidative burst disrupts the antioxidant defense of glutathione at the cellular and mitochondrial levels and promotes an increase in catalase activity. Due to the overwhelmed antioxidant system of the cell, the oxidative stress causes inactivation/denaturation of proteins such as cytochrome c, triggering a mitochondrial apoptotic response. Interestingly, the ROS increase in mitochondria was dependent on the functionality of the ISC system, rather than on the ETC blocking, which in this instance could be associated with a defective incorporation or recycling of Fe–S clusters in recipient apoproteins, which in consequence affects the ethanol susceptibility in S. cerevisiae. These findings presented in this study contribute to the understanding of the mechanisms of ethanol toxicity and demonstrate the important role of proteins involved in Fe–S prosthetic group assembly in the mechanisms of ROS generation and consequently in ethanol tolerance. Future studies will focus on the contribution of Fe released from Fe–S-containing proteins and Fe recycling dependent on the ISC system on ROS generation via the Fenton reaction and the effects on mitochondrial integrity and functionality.
We thank Drs G. Del Rio and L. Ongay from Instituto de Fisiología Celular/UNAM for yeast strains donation. This research was funded by CONACYT (106567), FOMIX-C01-117130, and C.I.C. 2.14/UMSNH grants.