All the microbial strains used in this study are listed in Supplementary Table S1. Yeast transformants were screened on yeast peptone dextrose (YPD) agar medium (10 g/L yeast exact, 20 g/L peptone, and 20 g/L glucose) containing 300 μg/ml G418. YPD plates were prepared by adding Bacto agar to a final concentration of 20 g/L. Before ethanol fermentation, yeast cells were cultured in YPD medium and then transferred to fermentation medium (100 g/L glucose, 4 g/L yeast exact, and 3 g/L peptone) with or without acetic acid or mixed inhibitors. The mixed inhibitors mimicking detoxified corn-stover hydrolysate are composed of 4.33 g/L acetic acid, 0.34 g/L formic acid, 0.53 g/L furfural, and 0.36 g/L HMF (Zhang et al., 2015). The composition of the corncob hydrolysate is 0.22 g/L formic acid, 1.45 g/L acetic acid, 10.21 g/L xylose, 100 g/L glucose, and other unknown components. When using corncob hydrolysate, 4 g/L peptone was added into the medium as a nitrogen source.

For proteomic analysis, the industrial yeast S. cerevisiae SPSC01 was inoculated into a 250 ml Erlenmeyer flask containing 100 ml seed medium and cultivated at 30°C, 150 rpm overnight. Next, the strains were deflocculated in 0.1 M sodium citrate buffer (pH 4.5), after which the cells were distributed in several flasks for inoculation. Fermentation was performed in 3 L bioreactors with 1.0 L of the fermentation medium containing 10 g/L acetic acid supplemented with or without 0.03 g/L zinc sulfate, and the initial optical density at 600 nm (OD₆₀₀) was adjusted to around 1.0. The fermentation was carried out at 30°C, 200 rpm, 0.04 vvm, and pH 4.5 and was stopped when the residual sugar of the zinc sulfate addition group was less than 2 g/L. Samples were taken every 12 h, and the corresponding cell growth (dry cell weight), residual sugar, and ethanol were recorded. Yeast cells were collected at the middle log phase in 60 h, then washed three times with cold dH₂O, and were then used for proteome analysis. Two biological replicates were analyzed, and proteome analysis was performed by The Beijing Genomics Institute (BGI, Beijing, China). The sequencing data were compared with the sequence of S. cerevisiae model strain S288C. Enrichment of functional categories among differentially expressed genes was examined using the MIPS Function Catalog (http://mips.gsf.de). Specific gene functions and biological pathways were analyzed based on the information from the Saccharomyces Genome Database (SGD) (http://www.yeastgenome.org) and Kyoto Encyclopedia of Genes and Genomes (KEGG) database (https://www.kegg.jp/).

The PGK1 promoter (PGK1p) was amplified with a pair of primers PPGK1-F/PPGK1-CDC42-R or PPGK1-F/PPGK1-KIC1-R according to the previously described method (Chen et al., 2022). The KanMX4 cassettes were amplified by PCR with a pair of primers K + CDC42-F/K + P_PGK1-R or K + KIC1-F/K + P_PGK1-R, respectively. Then PGK1p and the KanMX4 cassette were used as templates to amplify homologous recombination fragments by overlapping extension PCR (Lu et al., 2018) to replace the native promoter region of CDC42 or KIC1 with the pair primers of K + CDC42-F/P_PGK1-CDC42-R or K + KIC1-F/P_PGK1-KIC1-R, respectively. Subsequently, the constructed DNA fragments were transformed into the S. cerevisiae S288C strain using the chemical method (Bergkessel and Guthrie, 2013). The transformants were screened on YPD plates with 300 μg/ml G418 and were further verified by diagnostic PCR. All primers used are listed in Supplementary Table S2. The selected overexpression transformants of S288C were named S-CDC42 and S-KIC1, respectively.

Spot assays were performed according to the reference (Zhang et al., 2015) to test the stress tolerance of the mutants and the wild-type strains toward inhibitory conditions including 5 g/L acetic acid (pH value is about 3.7), 5 mM H₂O₂, 10% ethanol (v/v), 40°C, 4% lactic acid, and 1.5 g/L furfural, respectively. The inhibitory compounds were added to the medium after the YPD plates were autoclaved. Plates were photographed after 2 days of incubation at 30°C (except for 40°C thermal stress).

For all the seeds for spot assay, yeast cells were activated twice in the seed culture medium. Single colonies were picked up and cultivated for the first 12 h, and then the culture was transferred into fresh medium and cultivated for another 12 h. After the OD₆₀₀ of all yeast strains were adjusted to be the same, the seed cultures were inoculated into the fermentation medium with the inoculation size of 10% (v/v). The fermentation was performed at 30°C, 150 rpm. Samples were taken every 12 h, and the corresponding cell growth, ethanol, residual sugar, and glycerol were recorded.

The concentrations of ethanol, glucose, glycerol, and acetic acid in fermentation samples were determined by high-performance liquid chromatography (HPLC) system with a Bio-Rad Aminex^® HPX-87H column, and the elution was carried out at 50°C with 4 mM sulfuric acid and at a flow rate of 0.5 ml/min. The aforementioned compounds were detected and calculated according to previously described methods (Wang et al., 2013).

The cells of the engineered strains of S. cerevisiae and the control strain S288C cultivated in fermentation medium with or without stress were harvested at the log phase. The cell pellets of each sample were immediately stored at −80°C after being washed with sterilized dH₂O. Total RNA was extracted using a TransZol Plant Kit (TransGen Biotech, Beijing, China) and then reversely transcribed into cDNA using a PrimeScript RT Reagent Kit (TaKaRa, Dalian, China). Subsequently, real-time quantitative PCR (RT-qPCR) was carried out using the SYBR Green qPCR Master Mix (TaKaRa, Dalian, China) with a Real-Time PCR System (Bio-Rad, United States). The ALG9 gene was used as the reference gene, while relative expression levels were calculated by the 2^−ΔΔCt method (Teste et al., 2009). All primers used in this study are listed in Supplementary Table S2.

All experiments were independently repeated three times, and reproducible results were obtained. The results of the RT-qPCR and fermentation test were expressed as means and standard deviations (SD). Statistical analysis was performed using the Student's t-test, and a p-value less than 0.05 was regarded as statistically significant.

The addition of 0.03 g/L zinc sulfate enhanced ethanol fermentation in S. cerevisiae SPSC01 (Figure 1). Analysis of the differentially expressed proteins revealed a total of 107 proteins with significant changes (fold change >1.2 or <0.8) (Supplementary Data Sheet S1). The enriched gene ontology (GO) terms of the upregulated and downregulated proteins are shown in Supplementary Figure S1. For the upregulated proteins, the functions in the alcohol biosynthetic process, metabolism of amino acids and derivatives, and ergosterol metabolic process were enriched. Interestingly, most of the downregulated proteins function in cellular respiration, tricarboxylic acid (TCA) cycle pathway, oxidation phosphorylation, and mitochondrial proton-transporting ATP synthase complex assembly. When the protein interaction network was analyzed, it was found that proteins involved in ATP metabolic process, oxidative phosphorylation, and alcohol metabolic process were clustered (Figure 2A).

Acetic acid can cause intracellular acidification and change the intracellular proton gradient, which leads to the inversion of intracellular ATP synthase proton pump and insufficient intracellular energy supply (Guo and Olsson, 2014). We found that nine proteins related to ATP synthase were downregulated by zinc sulfate addition, including Atp20p, Atp4p, Atp19p, Atp7p, Atp15p, Atp16p, Atp5p, Atp3p, and Atp2p, although not very dramatically (Supplementary Table S4). In addition, Cox7p, Cox6p, and Cox12p, which are components of mitochondrial inner membrane electron transport chain, were also downregulated. Considering that energy production is closely related to carbon metabolism, we then further analyzed proteins related to central carbon metabolism.

Differentially expressed proteins involved in the glycolytic pathway, PP pathway (pentose phosphate pathway), and TCA pathway are shown in Supplementary Figure S2. Tdh1p is a glyceraldehyde-3-phosphate dehydrogenase (NAD+) and catalyzes the conversion of glyceraldehyde-3-phosphate to 3-phospho-d-glycerol-phosphate during glycolysis and the reverse reaction during gluconeogenesis (Tkach et al., 2012). Expression of Tdh1p was upregulated (1.32-fold) by zinc sulfate addition, together with Gpm2p (1.23-fold), which is a phosphoglycerate mutase family protein. On the other hand, downregulation of six proteins (Cit1p, Dal7p, Aco1p, Idh2p, Idh1p, and Fum1p) that are involved in the TCA pathway was observed. These downregulated proteins may increase the flux of pyruvate into ethanol production, which is consistent with the increased ethanol production by zinc sulfate supplement.

The transcription factor Zap1p is responsible for regulating gene expression under zinc deficiency conditions (Wilson and Bird, 2016). Based on the analysis using SGD and studies in the literature (Wu et al., 2008), many carbon metabolism enzyme-encoding genes, such as ADH1, ADH4, ENO1, ENO2, TKL2, and TDH1, are proved or hypothesized to be Zap1p target genes. Among these genes, we only found that TDH1 encoding enzyme was elevated by zinc sulfate addition (1.32-fold relative to the non-addition control) in this study, suggesting that there may be new zinc responsive proteins that were not identified as known Zap1p targets or indirect regulation by zinc sulfate plays an important role.

In our previous studies, transcriptome analysis was performed under the same condition used in this study. However, many differentially transcribed genes in the transcriptome data (Chen et al., 2017) were not found to be changed in our proteome results. This indicated that protein biosynthesis and transcription were differentially regulated by zinc sulfate, which is consistent with the previous reports that transcription changes do not necessarily lead to variation in protein biosynthesis (Lee et al., 2011). Nevertheless, some proteins are changed in both the proteome and the transcriptome results, including the upregulated ones, Ino1p, Pet18p, Pho3p, Thi20p, Sam2p, and Cho2p, and the downregulated ones, Gdh1p, Tmt1p, and Pdc5p. Interestingly, the expression of stress responsive protein Hsp12p was downregulated, but its gene transcription was upregulated by zinc sulfate. Among the proteins, expression of Ino1p and Thi20p was significantly upregulated (6.21-fold and 5.74-fold, respectively) by zinc sulfate under acetic acid stress. Thi20p is a trifunctional enzyme required for thiamine biosynthesis, degradation, and salvage (Onozuka et al., 2008). The C-terminal domains of Thi20p and the whole region of Pet18p of S. cerevisiae are homologous to bacterial thiaminase II. We tested the effect of overexpressing PET18 and SAM2 in S. cerevisiae S288C but did not find any difference of the engineered strains in yeast stress tolerance (data not shown). Ino1p is an inositol-3-phosphate synthase that is responsible for the synthesis of inositol phosphates and inositol-containing phospholipids. It was reported that overexpression of INO1 considerably improved tolerance of S. cerevisiae to lignocellulose-derived inhibitors that include acetic acid, furfural, and phenol (Wang et al., 2015). Our results implied that the increased expression of Ino1p by zinc sulfate may contribute to improved growth in the presence of acetic acid.

Some metabolites protect cells from stress damage, such as glycerol and ergosterol (Wang et al., 2014). In this study, we found that ergosterol biosynthesis enzymes were significantly increased in proteome by zinc sulfate addition, such as Erg20p, Erg1p, Erg11p, Erg3p, Erg5p, and Erg4p (Supplementary Figure S3). Our previous study found that the content of ergosterol was 14.6% higher than that of the control with the zinc sulfate addition, and we also demonstrated that increased ergosterol content may be correlated with improved membrane integrity (Zhang et al., 2017).

To further analyze the effect of zinc ions on global protein expression, the proteins containing zinc ions as cofactors or structural components were analyzed (Supplementary Table S3). Six proteins, namely, Rsf2p, Mal33p, Scs7p, Ams1p, Bet2p, and Axl1p, were identified to show significant changes. Among them, Rsf2p, Mal33p, Scs7p, and Ams1p were upregulated, whereas Bet2p and Axl1p were downregulated by zinc sulfate addition. The involvement of these proteins in stress tolerance was further analyzed by literature search. Rsf2p is a zinc-finger protein that regulates gene expression for acid pH resistance (Mira et al., 2009). On the other hand, the deletion of Mal33p and Scs7p encoding genes led to yeast cells being sensitive to visible light-induced stress resistance (Molin et al., 2020). It will be interesting to study the roles of these zinc-containing proteins in acetic acid stress response and tolerance of S. cerevisiae.

We further analyzed the potential interaction of the upregulated proteins (Figure 2B) in the comparative proteomic data. Cdc42p was reported to interact with Kic1p and Tdh1p, and Tdh1p is a key enzyme involved in ethanol production (Tkach et al., 2012). Therefore, we are interested in whether Kic1p and Cdc42p could promote ethanol fermentation efficiency of S. cerevisiae under stress conditions due to their specific functions. Upon zinc sulfate addition under acetic acid stress, the expression of Kic1p and Cdc42p were upregulated 2.13-fold and 1.21-fold, respectively, compared with the non-addition control. Kic1p localizes to the cytoplasm and is involved in yeast cell integrity, signal transduction, and budding cell apical bud growth (Sullivan et al., 1998; Vink et al., 2002; Hsu and Weiss, 2013). Cdc42p is involved in the establishment of cell polarity, and it also plays a role late in cell fusion via activation of key cell fusion regulator Fus2p (Johnson and Pringle, 1990; Ydenberg et al., 2012). So far, no related studies have shown that there is an obvious regulatory relationship between zinc ions and the two proteins, namely, Kic1p and Cdc42p. Interestingly, the relationship of Kic1p and Cdc42p with the key mitogen-activated protein kinase (MAPK) Hog1p that regulates multiple stress response was reported previously (Raitt et al., 2000; Janschitz et al., 2019). Hog1p is known to regulate acetic acid stress by affecting the aquaglyceroporin Fps1p (Mollapour and Piper, 2006; Mollapour and Piper, 2007) and the positive regulator Rtg2p, which is the RTG-dependent mitochondrial retrograde signaling that is related to Hog1p function during acid stress (Guaragnella et al., 2019). On the other hand, KIC1 and CDC42 were also reported to be the target genes for Yap1p (Salin et al., 2008). Yap1p is responsible for oxidative stress response, and it also functions in ethanol stress tolerance (Gulshan et al., 2011; Zyrina et al., 2017). Although no studies have been reported on the effect of Yap1p in acetic acid tolerance, we have found that acetic acid stress induces reactive oxygen species (ROS) accumulation and that zinc sulfate addition decreased ROS (Wan et al., 2015). Considering the connection of Kic1p and Cdc42p with Hog1p and Yap1p as well as their upregulation by zinc sulfate under acetic acid stress, we were interested whether KIC1 and CDC42 may be engineered to improve yeast acetic acid stress tolerance. The involvement of Kic1p and Cdc42p in yeast stress responses was revealed in several previous reports. For example, Kic1p was identified as a novel Hog1p target during hyperosmotic stress response (Janschitz et al., 2019), and Cdc42p is related to sensitivity to heat, cold (Davis et al., 1998), osmotic stress (Basu et al., 2020), and responses to several chemicals (Butcher et al., 2006; Breslow et al., 2008; Sukhai et al., 2013), but how these two proteins influence inhibitor tolerance remains unclear. Therefore, we further studied the effects of these two genes on stress tolerance.

Firstly, a spot assay was performed to test the effect of KIC1 and CDC42 on the tolerance of S. cerevisiae to various stress conditions. Improved resistance against different stresses in the engineered yeast strains overexpressing the two genes was observed when compared with the parental strain S288C (Figures 3A, B). In addition, the growth of S-KIC1 and S-CDC42 was similar to that of the control strain S. cerevisiae S288C when cultured under stress-free conditions. Specifically, S-CDC42 and S-KIC1 showed better growth than that of the control strain when cultured at 40°C or in the presence of 5 g/L acetic acid, and improved growth performance was also observed when the two engineered strains were challenged by 5 mM H₂O₂. Together, we revealed that KIC1 and CDC42 overexpression exerts significant promoting effects on yeast growth under various stresses in agar plates.

Subsequently, ethanol fermentation in the presence of 5 g/L acetic acid (initial pH at ∼3.5) was further investigated in shake flask culture. The two mutant strains showed similar performance in the presence of acetic acid (5 g/L). S-KIC1 and S-CDC42 showed better ethanol fermentation when compared with S288C under acetic acid stress (Table 1). Specifically, the fermentation time was shortened by 12 h in S-KIC1 and S-CDC42, leading to up to 52.99% higher ethanol productivity than that of the control strain, while the ethanol yield among these strains was comparable (Table 1). In addition, S-KIC1 and S-CDC42 also showed higher glycerol production than the wild-type strain.

To better understand the role of CDC42 and KIC1 in the adaptation of yeast cells to lignocellulose-derived inhibitors, we evaluated the performance of S-KIC1and S-CDC42 using the mixed inhibitors (0.34 g/L formic acid, 4.33 g/L acetic acid, 0.53 g/L furfural, and 0.36 g/L HMF) that mimicked the composition of corn stalk hydrolysates. There was a different performance of ethanol fermentation of S-KIC1 and S-CDC42 compared with the control strain S288C under the mixed inhibitors stress. All strains grew poorly due to the strong inhibition of mixed inhibitors, but both engineered strains showed 12 h shorter lag phase, and the strain S-CDC42 showed clearly improved growth than that of S-KIC1 and S288C.

We further explored changes in gene transcription in the yeast strains under stress. As shown in Figure 4, transcription levels of the two genes were more upregulated in the engineered yeast strains than that of the control strain (Figure 4A). Specifically, the transcription of CDC42 and KIC1 reached more than three and five times higher than that of the control strain under acetic acid stress, respectively. Higher expression of CDC42 and KIC1 was also found under the mixed inhibitors stress. At the same time, we also studied the expression of stress-related genes in the engineered yeast strains (Figure 4B). Transcription levels of CTT1 (Calabrese et al., 2019), GRE1 (Garay-Arroyo and Covarrubias, 1999), HAA1 (Mira et al., 2011), YAP1 (Okazaki et al., 2007), MSN2, and MSN4 (Hasan et al., 2002) were increased in the two engineered strains under mixed inhibitors stress, and GRE1 transcription was the most significantly improved. In contrast, no significant changes were observed in HSP30 (Samakkarn et al., 2021) and STB5 (Larochelle et al., 2006), respectively. The changes of most of the detected genes except GRE1 were almost the same in the two engineered strains, and the expression of GRE1 was higher in S-CDC42 than that in the S-KIC1 strain. These results implied that the mechanisms of enhancing stress tolerance may be different in the two engineered strains.

To better evaluate the effects of overexpressing CDC42 and KIC1 on the utilization of cellulosic hydrolysate, corncob hydrolysate was used to investigate the ethanol fermentation performance of the engineered yeast strains. There is no difference in the growth of S-CDC42 and S-KIC1 compared with S288C (Figure 5A). However, the engineered strains completely consumed glucose within 24 h, faster than the wild-type strain S288C (Figure 5B). The ethanol production of wild-type strain is 28.8 g/L in 24 h, which is 10 g/L lower than the strains overexpressing CDC42 and KIC1 (Figure 5C). The ethanol productivity of S-CDC42 and S-KIC1 reached about 1.60 g/L/h, increased by more than 33% compared with the wild-type S288C. Interestingly, the two engineered strains consumed acetic acid, but the parental strain S288C produced acetic acid as a by-product (Figure 5D).

The mass spectrometry proteomics data have been deposited to the PRIDE Archive (http://www.ebi.ac.uk/pride/archive/) via the PRIDE partner repository with the data set identifier PXD015459.