Saccharomyces cerevisiae Cells Lacking the Zinc Vacuolar Transporter Zrt3 Display Improved Ethanol Productivity in Lignocellulosic Hydrolysates

Yeast-based bioethanol production from lignocellulosic hydrolysates (LH) is an attractive and sustainable alternative for biofuel production. However, the presence of acetic acid (AA) in LH is still a major problem. Indeed, above certain concentrations, AA inhibits yeast fermentation and triggers a regulated cell death (RCD) process mediated by the mitochondria and vacuole. Understanding the mechanisms involved in AA-induced RCD (AA-RCD) may thus help select robust fermentative yeast strains, providing novel insights to improve lignocellulosic ethanol (LE) production. Herein, we hypothesized that zinc vacuolar transporters are involved in vacuole-mediated AA-RCD, since zinc enhances ethanol production and zinc-dependent catalase and superoxide dismutase protect from AA-RCD. In this work, zinc limitation sensitized wild-type cells to AA-RCD, while zinc supplementation resulted in a small protective effect. Cells lacking the vacuolar zinc transporter Zrt3 were highly resistant to AA-RCD, exhibiting reduced vacuolar dysfunction. Moreover, zrt3Δ cells displayed higher ethanol productivity than their wild-type counterparts, both when cultivated in rich medium with AA (0.29 g L−1 h−1 versus 0.11 g L−1 h−1) and in an LH (0.73 g L−1 h−1 versus 0.55 g L−1 h−1). Overall, the deletion of ZRT3 emerges as a promising strategy to increase strain robustness in LE industrial production.


Introduction
Lignocellulosic ethanol (LE) has gained increasing public attention as a secondgeneration biofuel. Indeed, it produces less greenhouse gases (GHGs) and can be obtained from inexpensive and abundant agricultural and forestry residues, and thus its production does not require the food resources that first-generation biofuels do [1,2]. In addition to the positive impact from an environmental perspective, the use of biomass over fossil fuels presents unquestionable advantages from economic and health standpoints [3]. Yeasts, in particular engineered Saccharomyces cerevisiae strains with improved tolerance to lignocellulosic hydrolysate inhibitors, have been the microorganisms most commonly explored for bioethanol production from LH (Table 1) [4]. However, lignocellulosic hydrolysis-derived inhibitors, such as weak acids (e.g., acetic, formic and levulinic acids), furan aldehydes (e.g., furfural and 5-hydroxymethylfurfural (HMF)), and phenolic compounds, together with increasing concentrations of ethanol along fermentation, cause severe stress that limits ethanol productivity [5,6]. Improving the resistance of yeast cells to such different stresses will thus boost fermentative efficiency [7] and determine LE production sustainability. Table 1. Recent studies on the application of engineered Saccharomyces cerevisiae strains towards improved tolerance to inhibitors in lignocellulosic hydrolysates. Adapted from [6], with permission of Springer Nature, 2022.
It has been suggested that superoxide dismutase (SOD) activity can increase transiently in cells undergoing AA-RCD, as an adaptation mechanism to cope with superoxide anion accumulation, and that catalase (CAT) activity is important to protect cells from AA-RCD [27]. Cytosolic antioxidant enzymes Sod1 and catalase T have zinc as a cofactor, and thus zinc plays an important role in yeast stress resistance, including ethanol and acid stress [28]. Indeed, among other protein-protective effects against oxidative stress, zinc increases the activity of catalases and superoxide dismutases, consequently reducing the deleterious effects of oxidant-promoting enzymes [29]. Accordingly, zinc sulphate addition has been shown to increase cell viability and ethanol production during high gravity ethanol fermentation [30]. Improved growth and ethanol productivity under AA stress by zinc supplementation was also observed [28,31,32]. The availability of zinc in the cytosol is regulated by specific transporters, both at the plasma (Zrt1 and Zrt2) and vacuolar membrane (Zrt3, Cot1 and Zrc1). Zrt3 mediates the efflux of stored zinc from the vacuole to the cytosol and its expression is upregulated under low zinc conditions, like that of Zrt1 and Zrt2 [33][34][35]. On the other hand, Cot1 and Zrc1 mediate vacuolar zinc accumulation under zinc-replete conditions. Under these conditions, the zrt3∆ strain accumulates zinc at higher levels than the wild-type strain by upregulating zinc uptake, even though it is unable to mobilize vacuolar zinc stores [36].
Given the bottlenecks and challenges still faced by the biotechnological industry of yeast-based bioethanol production from lignocellulosic biomass [4], a deeper understanding of AA-RCD may provide new clues towards the improvement of fermentative strategies. Because previous studies have shown that both vacuolar proteins and zinc play a role in yeast AA-RCD, the current study sought to address the involvement of the zinc vacuolar transporter Zrt3 in this process. It was observed that the Zrt3 transporter is involved in the execution of cell demise in response to AA, and that its absence improves cell growth and ethanol productivity. The zrt3∆ mutant therefore emerges as a more advantageous and cost-effective alternative for sustainable industrial LE production than LH supplementation with zinc.

Media and Growth Conditions
S. cerevisiae strains were cultivated on YEPD (Yeast Extract Peptone Dextrose) medium plates (1% (w/v) yeast extract, 2% (w/v) bactopeptone, 2% (w/v) glucose and 2% (w/v) agar) at 30 • C for 2 d. Cells were then transferred to liquid YEPD and cell density was adjusted to an OD 640 nm of 0.0005-0.005 (depending on the strain), and incubated at 30 • C with 200 rpm agitation until the mid-exponential phase (OD 640 nm of 0.5) before initiating the different assays performed throughout this work.

Cell Viability Assays
Yeast cells were cultivated as stated above (Section 2.2). At the exponential growth phase, cells were collected and resuspended in the appropriate medium adjusted to pH 3.0. The treatment was carried out by adding AA to the resuspended cells, followed by incubation at 30 • C under agitation at 200 rpm. Cells were collected 120 min after addition of AA and 50 µL of culture at OD 640 nm = 1 was diluted in a total of 1 mL of sterile deionized water, followed by 10 −4 serial dilutions. Five 40 µL drops were withdrawn from the 10 −4 dilution and plated in YEPD plates, grown for 2 d at 30 • C, and cell viability was assessed by counting the Colony Forming Units (c.f.u.). The percentage of cell survival was calculated from the ratio between the mean number of counted colonies over the mean number of colonies at time zero of the respective strain, multiplied by 100.

Assessment of Reactive Oxygen Species (ROS) Accumulation in Response to Acetic Acid Treatment by Flow Cytometry
Total ROS production was assessed with the superoxide anion indicator dihydroethidium (DHE) (Molecular Probes, Eugene, OR, USA). In the absence of superoxide anion, this probe exhibits blue fluorescence in the cytosol, but, when oxidized by the anion, it intercalates into nucleic acids displaying red fluorescence in the nucleus. The fluorescence probe Mitosox (Life Technologies, Carlsbad, CA, USA) was used to detect mitochondrial ROS production. It is a live-cell permeant dye that rapidly and selectively targets the mitochondria, where it is exhibits red fluorescence after oxidation by superoxide anion [37].
Cells were treated with AA as described above (Section 2.2), collected by centrifugation at 3000× g for 3 min, washed with deionized water and resuspended in 500 µL of PBS. Cells were then stained with DHE and Mitosox at a final concentration of 5 µg/mL and 1.5 µg/mL, for 20 and 45 min at room temperature, respectively. After staining, the samples were analyzed by flow cytometry.

Preparation of Protein Extracts
To prepare total protein extracts, cells were cultivated and treated as described above. After treatment, cells were collected by centrifugation at 3000× g for 3 min and washed with 20 mL of H 2 O. Cell lysis was carried out in a FASTprep equipment (5-10 × 30 s cycles) in lysis buffer (0.6 M Mannitol, 2 mM EGTA, 10 mM Tris, 1 protease inhibitor cocktail tablet without EDTA/10 mL, pH 6.8) containing glass beads (1/3 volume). Cell disruption was checked under the optical microscope. Cell debris was removed by centrifugation at 13,000× g for 10 min at 4 • C, and the supernatant was collected. Protein concentration was determined by the Lowry method [38].

Assessment of Mitochondrial Superoxide Dismutase Activity by Native Gel Electrophoresis
Protein extracts (10, 20 or 60 µg) were separated by 15% non-denaturing polyacrylamide gel electrophoresis (native-PAGE) to assess Sod1 activity. Electrophoresis was run for approximately 3 h at 20 mA per gel at 4 • C. In each assay, samples were run in duplicate to perform parallel staining with Coomassie blue (Fisher Scientific, Hampton, NH, USA) and Nitro Blue Tetrazolium (NBT) (Roche Molecular Systems Inc., Basel, CH). To this end, one of the gels was stained with a solution containing 50 mM potassium phosphate (pH 7.8), 0.1 µL/mL TEMED, 0.13 mg/mL NBT, and 0.1 mg/mL riboflavin for 45 min in the dark. After 45 min of incubation under slow agitation, the gel was exposed to light for 2 min. Under light exposure, riboflavin is reduced, which leads to the production of superoxide radicals. Then, O 2− is reduced by NBT to create insoluble blue formazan. Sod1 activity is visualized as a colourless band in the blue background because Sod1 scavenges the superoxide, thus inhibiting the blue colour formation [39]. The second gel was carefully transferred to a Coomassie staining solution (0.2% Coomassie blue, 7.5% of AA, 40% ethanol) for 20 min under stirring. Subsequently, the staining solution was removed, and the gel was washed with de-staining solution (10% of AA, 20% ethanol) several times until the blue bands were observed.
Photos of both gels were taken with a transilluminator, quantification performed using ImageJ Fiji software (ImageJ, Bethesda, MD, USA) and NBT normalized to Coomassie blue.

Assessment of Catalase Activity
Catalase activity was measured with a Clark electrode (Kipp & Zonen, Delft, NL, USA) to monitor the rate of O 2 production. The final protein concentration of each sample was set to 0.03 µg/µL in 2 mL of 50 mM phosphate buffer (pH 7.0). After obtaining a baseline, the reaction was started with the addition of 0.75 mM H 2 O 2 . This process was repeated 3 times for each sample/condition.

Assessment of Pep4 Localization and Vacuolar Membrane Permeabilization by Fluorescence Microscopy
The wild-type and zrt3∆ strains expressing Pep4-mCherry were treated as described above (Section 2.2). To assess vacuolar membrane permeabilization (VMP), cells were collected and stained for 20 min in the dark with Celltracker™ Blue CMAC (7-amino-4chloromethylcoumarin) (Molecular Probes, Eugene, OR, USA) at a final concentration of 2 µM, using a stock solution of 100 µM in Dimethyl sulfoxide (DMSO). Under normal conditions, this probe is sequestered in the vacuole exhibiting a bright blue fluorescence. However, when the vacuolar membrane is compromised, it displays a diffuse pattern [40]. This procedure was used to assess Pep4 localization and VMP during AA treatment.
Fluorescence microscopy experiments were performed with a Leica Microsystems DM-5000B microscope (Leica Microsystems, Wetzlar, DE, Germany), with the appropriate filter settings (DIC (differential interference contrast), red and blue) with a 100× objective. Images were obtained with a Leica DFC350 FX Digital Camera (Leica, Wetzlar, DE, Germany) and processed with LAS AF Microsystems software (Leica, Wetzlar, DE, Germany). For phenotype quantification, at least 300 cells were counted in three independent experiments.

Evaluation of Vacuolar pH by Flow Cytometry
The fluorescent probe 5-(and -6)-carboxy-2 , 7 -dichlorofluorescein diacetate (CD-CFDA) (Sigma-Aldrich, St. Louis, MO, USA) was used to analyze vacuolar pH variations in AA-treated cells. This probe accumulates in acidic organelles and undergoes pH-dependent changes, producing a more intense fluorescence as vacuoles become more alkaline [41,42]. A stock of 1 mM of CDCFDA was prepared in DMSO and kept at −20 • C.
Cells were treated with AA as described above (Section 2.2), collected (1 mL at OD 640 nm of 0.1), washed with deionized water and resuspended in 1 mL of CF buffer (50 mM glycine, 10 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 40 mM Tris, 100 mM MES (2-(N-morpholino) ethanesulfonic acid), 2% (v/w) glucose, pH 4.5). The probe was added to a final concentration of 1.6 µM and the mixture was incubated at 30 • C for 20 min, at 200 rpm. Cells were collected, washed with CF buffer without glucose supplementation, resuspended in 700 µL of the same buffer, and analyzed by flow cytometry [43].
Data acquisition was performed in the same equipment as in Section 2.4 and fluorescence was collected using FL-1 (505-545 nm).

Zinc Supplementation and Limitation Assays
For zinc supplementation assays, cells cultivated and resuspended as described above were treated with 10 mM ZnSO 4 , 100 mM AA, and with 10 mM ZnSO 4 plus 100 mM AA, and incubated at 30 • C, 200 rpm. Cells were harvested before (0 min) and 60 min after the addition of AA and/or ZnSO 4 to assess cell viability and cellular accumulation of superoxide anion. As a control, cells were processed under the same conditions but in the absence of both ZnSO 4 and AA.
For zinc limitation assays, cells were pre-treated with and without 1 mM EDTA (stock solution of 100 mM), for 30 min at 30 • C and 200 rpm. Then, the cells cultivated and resuspended as stated above were treated with 1 mM EDTA plus 75 mM AA, and incubated at 30 • C, 200 rpm. Cells were harvested before (0 min) and 90 min after the addition of AA and EDTA to assess cell viability and cellular accumulation of superoxide anion. Controls were prepared in the absence of EDTA and AA.

Assessment of zrt3∆ Cell Growth and Fermentative Performance in the Presence of Acetic Acid
To assess cell growth of wild-type and zrt3∆ strains, cells were pre-inoculated on YEPD, at 30 • C and 200 rpm, overnight. The culture was then set at an OD 640 nm of 0.5 in YEPD medium containing 40 g/L glucose, pH 4.5, before addition of 60 mM AA. Controls were performed in the absence of AA. Cell growth was evaluated by measuring the OD 640 nm along 32 h. Lag phase time was determined as the time required to double the initial optical density [44]. Glucose and ethanol in the medium were measured by HPLC (Beckman Coulter, Brea, CA, USA) [45].

Assessment of zrt3∆ Cell Growth and Fermentative Performance in a Lignocellulosic Hydrolysate
Wild-type and zrt3∆ cells were cultivated as described in Section 2.9, but in a different medium. Vine pruning residue (Vpr) was submitted to hydrothermal treatment (autohydrolysis) under non-isothermal conditions (Tmax of 220 • C) in a 2 L stainless steel reactor (Parr Instruments Company, Moline, IL, USA) equipped with Parr PDI temperature controller (model 4848) at liquid to solid ratio of 6 g distilled water/1 g of Vpr. After autohydrolysis, solid and liquid phases were separated by filtration and autohydrolysis liquors (liquid phase) were supplemented with 1% (w/v) yeast extract, 2% (w/v) bactopeptone, 4% (w/v) glucose, and adjusted to pH 4.5 with NaOH and used as culture medium for growth assays. An aliquot of autohydrolysis liquors (liquid phase) was filtered through 0.45 µm membranes and used for direct HPLC determination of AA, formic acid, hydroxymethylfurfural (HMF) and furfural as described above. Cell growth was evaluated by measuring the OD 640 nm for 30 h. Lag phase time was determined as described in Section 2.9.

The zrt3∆ Strain Is Resistant to Acetic Acid-Induced Cell Death
Cell survival assays showed that ZRT3 deletion increases resistance of cells to AA treatment (100 mM, pH 3.0) in comparison with the wild-type strain. Indeed, cell survival of wild-type cells after 120 min of treatment was close to zero, while most zrt3∆ cells remained alive (ca. 70% survival) ( Figure 1). bactopeptone, 4% (w/v) glucose, and adjusted to pH 4.5 with NaOH and used as c medium for growth assays. An aliquot of autohydrolysis liquors (liquid phase) w tered through 0.45 μm membranes and used for direct HPLC determination of AA, acid, hydroxymethylfurfural (HMF) and furfural as described above. Cell grow evaluated by measuring the OD640 nm for 30 h. Lag phase time was determined as des in Section 2.9.

The zrt3Δ Strain is Resistant to Acetic Acid-Induced Cell Death
Cell survival assays showed that ZRT3 deletion increases resistance of cells treatment (100 mM, pH 3.0) in comparison with the wild-type strain. Indeed, cell su of wild-type cells after 120 min of treatment was close to zero, while most zrt3Δ c mained alive (ca. 70% survival) (Figure 1).

The Acetic Acid Resistance Phenotype of zrt3Δ Is Associated with Lower Accumulatio Superoxide Anion and Increased Activity of Catalases
To address whether the AA resistance phenotype of zrt3Δ was associated with mitochondrial ROS accumulation, cellular and mitochondrial superoxide anion ac lation was measured by flow cytometry after staining with DHE and Mitosox, r tively. Both cellular and mitochondrial superoxide anion accumulation in response decreased significantly in the zrt3Δ mutant, as only less than 20% and 5% of the cell stained with DHE and Mitosox, respectively. In contrast, the wild-type strain exh around 80% and 15% of DHE-and Mitosox-positive cells, respectively, (Figure 2a). fore, the next step was to evaluate how the activities of superoxide dismutases a alases are affected by ZRT3 deletion, which is associated with increased cellular zinc [36]. zrt3Δ cells displayed a significantly higher basal activity of catalases than wil cells (Figure 2d), while Sod1 activity was not significantly different (Figure 2b,c). M ver, AA did not affect the activity of catalases in the wild-type strain, but signif decreased the enhanced basal activity of catalases in zrt3Δ cells (Figure 2d). On th hand, treatment with AA increased the Sod1 activity of the wild-type strain, thou was not statistically significant, while it had no effect in zrt3Δ cells (Figure 2c).

The Acetic Acid Resistance Phenotype of zrt3∆ Is Associated with Lower Accumulation of Superoxide Anion and Increased Activity of Catalases
To address whether the AA resistance phenotype of zrt3∆ was associated with lower mitochondrial ROS accumulation, cellular and mitochondrial superoxide anion accumulation was measured by flow cytometry after staining with DHE and Mitosox, respectively. Both cellular and mitochondrial superoxide anion accumulation in response to AA decreased significantly in the zrt3∆ mutant, as only less than 20% and 5% of the cells were stained with DHE and Mitosox, respectively. In contrast, the wild-type strain exhibited around 80% and 15% of DHE-and Mitosox-positive cells, respectively, (Figure 2a). Therefore, the next step was to evaluate how the activities of superoxide dismutases and catalases are affected by ZRT3 deletion, which is associated with increased cellular zinc levels [36]. zrt3∆ cells displayed a significantly higher basal activity of catalases than wild-type cells (Figure 2d), while Sod1 activity was not significantly different (Figure 2b,c). Moreover, AA did not affect the activity of catalases in the wild-type strain, but significantly decreased the enhanced basal activity of catalases in zrt3∆ cells (Figure 2d). On the other hand, treatment with AA increased the Sod1 activity of the wild-type strain, though this was not statistically significant, while it had no effect in zrt3∆ cells (Figure 2c).

The zrt3Δ Acetic Acid Resistance Phenotype Is Associated with Delayed Vacuolar Alterations and Maintenance of Vacuolar pH
To further explore the role of Zrt3 in vacuolar dysfunctions occurring in the cell dea process under study, the zrt3Δ strain was transformed with pRS413-Pep4-mCherry monitor Pep4 localization in response to AA by fluorescence microscopy. About 80% the wild-type cells exhibited Pep4-mCherry translocation from the vacuole to the cyto

The zrt3∆ Acetic Acid Resistance Phenotype Is Associated with Delayed Vacuolar Alterations and Maintenance of Vacuolar pH
To further explore the role of Zrt3 in vacuolar dysfunctions occurring in the cell death process under study, the zrt3∆ strain was transformed with pRS413-Pep4-mCherry to monitor Pep4 localization in response to AA by fluorescence microscopy. About 80% of the wild-type cells exhibited Pep4-mCherry translocation from the vacuole to the cytosol 60 min after AA treatment, whereas Pep4-mCherry translocation was observed only in ca. 15% of zrt3∆ cells. This difference was equally apparent after 120 min (Figure 3a,b). To evaluate vacuolar membrane permeabilization (VMP), cells transformed with pRS413-Pep4-mCherry were stained with the vacuolar dye CMAC, as reported in Materials and Methods. As shown in Figure 3a, the majority of the wild-type cells displayed concomitant Pep4 release and VMP both 60 and 120 min after AA treatment. In contrast, only ca. 15% of zrt3∆ mutant cells exhibited red and blue fluorescence diffused into the cytosol 60 min after AA treatment, indicating a significant delay/decrease in VMP and Pep4 release. In parallel, the results show that AA induced a 2.0-fold increase in CDCF fluorescence, indicating an increase in vacuolar pH in the wild-type cells, whereas it did not affect the vacuolar pH of zrt3∆ cells (Figure 3c). These data are in accordance with the delayed VMP observed in AA-treated zrt3∆ cells.

Loss of Cell Survival and Superoxide Anion Accumulation Induced by Acetic Acid Are Affected by Zinc Availability
ZnSO 4 increased the cell survival in response to AA from 10 to 50% (Figure 4a), indicating that supplementation with ZnSO 4 protects cells from AA-RCD. In agreement, limiting zinc (in the presence of EDTA) decreased cell survival (Figure 4b). In experiments with zinc limitation, 75 mM of AA was used instead of 100 mM (used in zinc supplementation), since co-treatment with EDTA and 100 mM AA was extremely toxic.
To uncover whether ROS accumulation is involved in the effect of zinc availability on AA-RCD, cellular accumulation of superoxide anion was evaluated under zinc supplementation and limitation conditions by flow cytometry using the DHE probe. Zinc supplementation did not alter AA-induced ROS accumulation (Figure 4c), while zinc limitation (in the presence of EDTA) promoted an apparent increase in DHE-positive cells, but that was not statistically significant.

zrt3∆ Displays Improved Fermentative Performance in the Presence of Acetic Acid
Fermentation experiments in YEPD medium in the presence of AA showed a large difference between the performance of the zrt3∆ mutant and the wild-type strain regarding both cell growth and ethanol production. Remarkably, while the exponential growth phase of the zrt3∆ mutant was initiated 8 h after inoculation, the lag phase of the wild-type strain was much longer (31 h, Figure 5a). This translated into a much higher ethanol production 32 h after inoculation in the mutant strain when compared with the wild-type strain (9.4 g L −1 versus 3.4 g L −1 , respectively) ( Figure 5b). parallel, the results show that AA induced a 2.0-fold increase in CDCF fluorescence, indicating an increase in vacuolar pH in the wild-type cells, whereas it did not affect the vacuolar pH of zrt3Δ cells (Figure 3c). These data are in accordance with the delayed VMP observed in AA-treated zrt3Δ cells.

zrt3Δ Displays Improved Fermentative Performance in the Presence of Acetic Acid
Fermentation experiments in YEPD medium in the presence of AA showed a large difference between the performance of the zrt3Δ mutant and the wild-type strain regarding both cell growth and ethanol production. Remarkably, while the exponential growth phase of the zrt3Δ mutant was initiated 8 h after inoculation, the lag phase of the wildtype strain was much longer (31 h, Figure 5a). This translated into a much higher ethanol

zrt3Δ Shows Improved Fermentative Performance in a Lignocellulosic Hydrolysate
As reported in Materials and Methods, the lignocellulose biomass used was vine prune residue subjected to autohydrolysis. This pre-treatment resulted in the accumulation of AA (21 mM), formic acid (19 mM), furfural (2 mM) and 5-HMF (2.4 mM). To avoid nutrient limitation, the LH was supplemented with YEPD medium components. Although the OD at stationary phase was identical for both strains 30 h after the inoculation (Figure 6a), the fermentative performance of zrt3Δ was better than that of the wild-type strain. Indeed, the lag phase was 5 and 7 h; µ max, 0.24 and 0.19 h −1 for the zrt3Δ strain and wild-type strain, respectively. Importantly, at the end of the exponential phase (12 h), the zrt3Δ mutant produced a higher ethanol concentration (Figure 6b).

zrt3∆ Shows Improved Fermentative Performance in a Lignocellulosic Hydrolysate
As reported in Materials and Methods, the lignocellulose biomass used was vine prune residue subjected to autohydrolysis. This pre-treatment resulted in the accumulation of AA (21 mM), formic acid (19 mM), furfural (2 mM) and 5-HMF (2.4 mM). To avoid nutrient limitation, the LH was supplemented with YEPD medium components. Although the OD at stationary phase was identical for both strains 30 h after the inoculation (Figure 6a), the fermentative performance of zrt3∆ was better than that of the wild-type strain. Indeed, the lag phase was 5 and 7 h; µ max , 0.24 and 0.19 h −1 for the zrt3∆ strain and wild-type strain, respectively. Importantly, at the end of the exponential phase (12 h), the zrt3∆ mutant produced a higher ethanol concentration (Figure 6b).

Discussion
Among the microorganisms used for industrial bioethanol production, yeasts are the most commonly explored [4], as they can produce high concentrations of ethanol, are able to ferment a wide range of sugars [46] and, importantly, allow low process costs. Cultures are also less prone to contaminations, as they ferment at an acidic pH. S. cerevisiae is considered the favorite species for ethanol production because of its high ethanol yield in media with glucose and ethanol tolerance, and is expected to play a central role in multiwaste valorization approaches [47]. However, its main drawback is its incapacity to ferment the xylose that is released during hydrolysis pretreatment of lignocellulosic biomass [2]. Scheffersomyces stipitis, previously known as Pichia stipitis, is one of the most efficient naturally occurring xylose-fermenting yeasts, showing a high ethanol yield in media with xylose, but it is highly sensitive to ethanol [48,49]. Thus, the construction of xylose-fermenting recombinant S. cerevisiae strains has been exploited as a strategy to develop economically viable processes for the production of LE [4,50]. Of note, a synergistic effect of simultaneous expression of the two xylose pathways (xylose reductase/xylitol

Discussion
Among the microorganisms used for industrial bioethanol production, yeasts are the most commonly explored [4], as they can produce high concentrations of ethanol, are able to ferment a wide range of sugars [46] and, importantly, allow low process costs. Cultures are also less prone to contaminations, as they ferment at an acidic pH. S. cerevisiae is considered the favorite species for ethanol production because of its high ethanol yield in media with glucose and ethanol tolerance, and is expected to play a central role in multi-waste valorization approaches [47]. However, its main drawback is its incapacity to ferment the xylose that is released during hydrolysis pretreatment of lignocellulosic biomass [2]. Scheffersomyces stipitis, previously known as Pichia stipitis, is one of the most efficient naturally occurring xylose-fermenting yeasts, showing a high ethanol yield in media with xylose, but it is highly sensitive to ethanol [48,49]. Thus, the construction of xylose-fermenting recombinant S. cerevisiae strains has been exploited as a strategy to develop economically viable processes for the production of LE [4,50]. Of note, a synergistic effect of simultane-ous expression of the two xylose pathways (xylose reductase/xylitol dehydrogenase and xylose isomerase) with furan detoxification was observed [51]. However, the use of xylosefermenting strains that are not resistant to AA is still not a competitive strategy for LE production. Indeed, AA is one of the inhibitors derived from lignocellulosic hydrolysis [4] that causes severe stress, limiting yeast ethanol yield and productivity. Therefore, a deeper understanding of the molecular mechanisms underlying resistance to AA will contribute to designing robust strains for sustainable industrial LE production.
This study uncovered that the absence of Zrt3 results in higher cell survival in response to AA, suggesting that Zrt3 is involved in AA-RCD, possibly through its role in the regulation of zinc levels. A recent study identified that the AP-3 adaptor complex is a death-promoting factor during heat-induced cell death and is also involved in vacuole permeabilization. A recent study uncovered that disruption of the AP-3 adaptor complex, which promotes stress-induced cell death by transporting and installing proteins on the vacuolar membrane, confers resistance to AA and hydrogen peroxide [52]. Since Zrt3 is a candidate cargo protein for the AP-3 adaptor complex, those results are consistent with the data obtained in this study linking absence of Zrt3 from the vacuole with a cell death resistance phenotype.
Previous studies showed that S. cerevisiae cells undergoing AA-RCD exhibit accumulation of ROS [22,53]. Cytosolic and mitochondrial superoxide dismutases (Sod1 and Sod2, respectively) are responsible for the conversion of the superoxide anion, generated by the reduction of oxygen, into hydrogen peroxide [54]. In turn, cytosolic and peroxisomal catalases (Cat T and Cat A, respectively) convert hydrogen peroxide into water and oxygen [55]. A well-established set of reactions with these and other antioxidant enzymes allows cells to maintain homeostasis [56]. Sod2, which is particularly required during stationary phase, is involved in mitochondrial oxygen radical detoxification [56]. In the current study, the higher survival of the zrt3∆ mutant in response to AA was associated with lower accumulation of both cellular and mitochondrial superoxide anion, in comparison with the wild-type strain. This could be due to higher cytosolic antioxidant defenses, in particular to higher activity of zinc-dependent antioxidant enzymes such as Sod1 and catalase T [57]. In agreement with this hypothesis, the decreased percentage of zrt3∆ cells with total cellular accumulation of superoxide anion (DHE-positive) in comparison with the wild-type strain, was higher than the correspondent decrease in the percentage of cells with superoxide anion mitochondria accumulation (Mitosox-positive). Although Sod1 basal activity was identical for both strains, basal activity of catalases was higher in the zrt3∆ mutant, which likely results in lower hydrogen peroxide accumulation. In wild-type cells, the basal activity of catalases is probably not sufficient to counteract the accumulation of hydrogen peroxide upon AA stress, which leads to the induction of Sod1, as described for its human ortholog [58]. In turn, the high basal activity of catalases in zrt3∆ cells ensures low levels of hydrogen peroxide throughout the course of exposure to AA. This would result in a lack of Sod1 induction and even decreased activity of catalases, whose stimulation depends on hydrogen peroxide levels [59]. This hypothesis is in agreement with the observation that cells pre-adapted to acid stress, where the activity of catalases is enhanced, are resistant to AA-RCD [27]. Although other studies obtained by these authors, as well as by others, are not conclusive about the effect of RCD-inducing concentrations of AA on catalase or superoxide dismutase (SOD) activities, there is a consensus that increased activity of catalases protects cells from AA-RCD [24]. Thus, the results obtained in the current study suggest that the higher activity of catalases endows the zrt3∆ mutant with greater resistance to AA, in accordance with a major role of hydrogen peroxide detoxification in the prevention of AA-RCD [21].
AA is also known to induce vacuolar membrane permeabilization (VMP), along with the release of Pep4 to the cytosol [25]. This vacuolar protease plays a role in mitochondrial degradation during AA-RCD and, depending on the strain background, can play a protective or an executioner role [25,60]. In fact, BY4741 yeast cells lacking Pep4 become more resistant to AA, suggesting an executioner role in this strain background [26]. Interestingly, both Pep4 release and VMP were delayed in the zrt3∆ mutant, consistent with its increased cell survival. Moreover, wild-type cells exhibited significant vacuolar alkalinization in response to AA, while the acidic pH is maintained in zrt3∆ cells. These data support the interpretation that the increased survival of the zrt3∆ mutant to AA is associated with a delayed perturbation in vacuolar function, including a later VMP and Pep4 release.
It was previously found that zinc protects yeast cells from oxidative damage caused by AA [61]. Another study with S. cerevisiae demonstrated that zinc modulates cellular amino acid metabolism and redox balance, particularly the biosynthesis of alanine and glutathione, to exert its antioxidant effect [28]. In agreement, supplementation with ZnSO 4 was shown here to protect wild-type cells from AA-RCD. However, this protective effect was not associated with a significant decrease in superoxide anion levels. On the other hand, the limitation of zinc rendered cells more sensitive to AA-RCD, and led to increased superoxide anion levels. While an effect of zinc-chelator EDTA on the homeostasis of other cations and cell processes cannot be ruled out, these results are in accordance with previous reports that S. cerevisiae shows increased ROS levels under zinc-limiting conditions [61]. Taken together, these data suggest that zinc attenuates the cytotoxic effects of AA, but not by minimizing the accumulation of superoxide anion.
Although the wild-type strain displays increased AA resistance when supplemented with zinc, and this supplementation is known to improve fermentation efficiency, above certain concentrations it becomes cytotoxic [28,30,31]. Indeed, besides being associated with additional costs, zinc supplementation can affect cell function and ultimately compromise ethanol productivity, if not carefully handled. Thus, the exploitation of the zrt3∆ mutant in LE production emerges as a more advantageous alternative, as it avoids supplementation of LH with zinc.
To characterize the performance of zrt3∆ cells under conditions more similar to LE production, their fermentation profile was monitored under AA stress conditions. It is well established that AA originates from acetyl-groups of hemicellulose during the pre-treatment that lignocellulosic biomass undergoes to release fermentable sugars [17,62]. Depending on the biomass and hydrolytic process used, AA can be found in concentrations between 8 and 192 mM (0.5 and 11.5 g/L) [63]. Here, the zrt3∆ mutant stood out when compared with the wild-type regarding both cell growth and fermentative performance under 60 mM (3.6 g/L) AA stress, with a 2.6-fold improvement in ethanol productivity. These results are in agreement with previous reports showing that a xylose-recombinant S. cerevisiae strain subjected to 30 mM AA stress produces more ethanol in the presence of 180 µM ZnSO 4 [31]. The improved growth of the zrt3∆ strain, and consequently its fermentative performance, may rely on the fact that this mutation causes upregulation of zinc transport across the plasma membrane to compensate for zinc scarcity, resulting in a richer zinc pool. Accordingly, under zinc-replete conditions, there is a higher zinc accumulation in zrt3∆ mutants than in the wild-type strain [36]. Although these results were encouraging, the presence of many different inhibitors in LH has been reported to have a synergistic effect on yeast growth inhibition [64,65]. Therefore, the fermentation profile of wild-type and zrt3∆ mutant cells in vine prune residue autohydrolysis liquors containing AA (21 mM), formic acid (FA, 19 mM), furfural (F, 2 mM) and 5-hydroxymethylfurfural (HMF, 2.4 mM) was compared. The zrt3∆ strain still displayed increased fermentative capacity and a 1.3-fold improvement in ethanol productivity in comparison with the wild-type strain. Despite the observed difference in productivity, ethanol yield was similar for the two strains, as for both cases 1 g of glucose was converted to 0.23 g of ethanol, which corresponds to 45% of the theoretical maximum yield. Although the concentration of each inhibitor per se was relatively low in the LH, the observed cell growth inhibition and ethanol yield reduction are likely a result of their combined inhibitory effects. Whether the absence of Zrt3 also protects against FA, F and HMF alone or in combination with AA is currently being addressed.

Conclusions and Future Perspectives
Acetic acid is the main inhibitor of fermentation of lignocellulosic hydrolysates. Given its influence in the fermentative process, and consequently in the production of ethanol, it is increasingly important to find strategies to overcome this limitation. In this study, we aimed to clarify the role of the zinc transporter Zrt3 in the mechanisms underlying acetic acid-induced RCD in yeast, towards the improvement of LE production. Taken together, the results indicate that the zrt3∆ mutant displays higher resistance to cell death induced by acetic acid and an improved fermentation profile, particularly a higher level of ethanol productivity, in comparison with the wild-type strain. These data validate the potential of exploiting the zrt3∆ mutant in the bioethanol industry as a much more practical and cost-effective alternative to medium supplementation with zinc that had been proposed to mitigate the effects of the LH-derived inhibitor AA. Moreover, the results confirm the hypothesis that the zinc vacuolar transporter Zrt3 is involved in vacuolemediated AA-RCD.
In the near future, the next step will be the validation of the applicability of our findings in the industrial production of bioethanol from LH through the deletion of ZRT3 in a S. cerevisiae xylose-utilizing strain. Furthermore, studies should be directed towards determining the sustainability of using this mutant strain. The use of methodologies such as life cycle assessment (LCA) and exergy, together with environmental and economic analyses (e.g., exergoenvironmental and exergoeconomic analyses, respectively), will allow for an adequate assessment of the degree of sustainability of the proposed solution for LE production [66,67].  Acknowledgments: The authors acknowledge Helena Pereira for constructing the pRS413-Pep4-mCherry plasmid and Sara L. Baptista for the preparation of vine prune residue autohydrolysis liquors. BIOVINO project (0688_BIOVINO_6_E) funded by INTERREG España-Portugal and European Regional Development Fund (ERDF) is acknowledged for the supply of vine prune residues.

Conflicts of Interest:
The authors declare no conflict of interest.