Tolerance of Kluyveromyces marxianus Under Acetic Acid-, Isoamyl Alcohol-, Hydrogen Peroxide-, and Ethanol-Induced Stress

Abstract

Kluyveromyces marxianus is a yeast that can be used as a microbial factory. However, little is known about its response to stress conditions. This work evaluated the response of this yeast against ethanol, acetic acid, isoamyl alcohol, and hydrogen peroxide as stress agents. Cytotoxicity assays were performed to assess the residual viability using a direct method (CFU counting) and an indirect method based on the reduction in MTT. Then, fermentation kinetics were performed at IC30 and IC50 for each stress factor to evaluate the effect of moderate and intense stress. This work is the first report presenting IC50 values for ethanol (21.82 g/L), acetic acid (1.19 g/L), isoamyl alcohol (2.74 g/L), and hydrogen peroxide (0.09 g/L) in K. marxianus. The IC50 values for the indirect method are between 3.7 and 68% higher than those for the direct method. Hydrogen peroxide and ethanol were the stress agents showing the highest overestimations. The results presented here demonstrated the overestimation of cell viability by the indirect method. Direct CFU counting is an adequate method to determine yeast viability during toxicity studies of chemical compounds. It was also established that ethanol and hydrogen peroxide have the highest toxicity against K. marxianus ITD-01005 during fermentation at concentrations equivalent to IC30 and IC50 of each stress agent.

1. Introduction

Kluyveromyces marxianus is an unconventional yeast listed as QPS and GRAS by the European Union and the United States, respectively. It is a thermotolerant microorganism [1], exhibiting rapid growth and capable of metabolizing hexoses, pentoses, disaccharides, and organic acids such as lactic and acetic acids [2,3]. Unlike Saccharomyces cerevisiae, K. marxianus is Crabtree-negative and exhibits attractive aerobic respiration characteristics [2]. Recent studies have evaluated the potential of this yeast as a cellular factory for the production of ethanol [4], xylitol [5], esters [6], and enzymes such as β-glucosidase [7], as well as its probiotic potential in co-cultures [8].

On the other hand, microbial stress is the set of physical, chemical, or biological conditions that compromise cell growth, viability, or metabolic activity, preventing optimal microbial performance. Physical stress is caused by temperature, osmotic pressure, radiation, and hydrostatic pressure. In contrast, chemical stress is caused by extreme pH, reactive oxygen species (ROS) [9], ethanol, antibiotics, heavy metals, or toxic compounds [10]. Biological stress is due to interaction with other microorganisms that compete for nutrients and space or produce antimicrobial agents [11]. In nature, microorganisms frequently must deal with fluctuations in more than one growth parameter simultaneously or consecutively [12].

The study of stress in yeast has gained relevance, with nutritional stress and osmotic stress being the most frequently investigated [13,14,15]. It is due to their impact on yeast physiology, affecting microorganisms’ survival, growth, and metabolism. Microorganisms respond to stress by activating signal transduction pathways, synthesizing protective molecules to repair the damage caused, and regulating protein activity [16]. Betlej et al. [17] reported that Saccharomyces cerevisiae increased its glycogen and glycerol production in response to osmotic stress, observing that strains with higher amounts of glycogen showed greater vitality, better performance, and a more remarkable ability to tolerate stress. It resulted in a better ability to produce resveratrol-enriched wines.

Nutritional stress by limitation or excess of essential nutrients profoundly affects cellular metabolism. Nitrogen deficiency, for example, affects amino acid biosynthesis and secondary metabolite production. At the same time, glucose limitation under osmotic stress activates autophagy pathways as a strategy to recycle nutrients and maintain cell viability. In Saccharomyces cerevisiae, increased osmolarity activates a signal transduction pathway, which regulates the intracellular accumulation of glycerol as a compatible osmolyte. Recent studies have identified new regulators of this pathway, such as the MAPK protein kinase Hog1, which controls the expression of genes involved in glycerol synthesis and modulates processes such as the cell cycle and DNA repair under severe stress conditions [18].

Fu et al. [19] studied the response of the non-conventional yeast K. marxianus to high-temperature stress. The high temperature caused (a) an increase in reactive oxygen species generation, (b) a diminishment in the NADH/NAD ratio due to the stimulation of mitochondrial respiration but repressing the tricarboxylic acid cycle, and (c) a reduction in cell membrane resistance to ethanol. The yeast increased glycerol and ergosterol production as protective molecules. Glycerol production re-oxidizes NADH, while ergosterol protects against ethanol stress. Reactive oxygen species are neutralized by NADPH, which is obtained by decreasing the biosynthesis of branched-chain amino acids and increasing acetic acid production.

Then again, the assay based on the reduction of 3-(4,5-dimethylthiazol-2-yl)-2,5,-diphenyl tetrazolium bromide (MTT) is widely used and accepted for assessing proliferation, viability, toxicity, and chemosensitivity of cultured cells [20]. This assay was developed to study tissue cells and has also been used to estimate microbial cell viability. However, the mechanism by which bacteria and yeast reduce MTT is not fully understood, and results may be misinterpreted [21]. Despite these limitations, some studies have reported using the MTT-reduction test to assess microbial strains’ viability and respiratory activity [22]. The MTT assay is also used in various protocols, such as the detection of multidrug-resistant bacteria [23,24], the evaluation of biofilms [25], or the indirect quantification of compounds with antimicrobial activity [26].

The K. marxianus characteristics and the growing global demand for biomolecules have increased interest in this yeast for applications in food, environment, and biotechnology [27]. Considering the stress microorganisms face in production processes, it is relevant to expand the knowledge of the stress response of K. marxianus. Then, this work aimed to evaluate the response of K. marxianus under acetic acid-, isoamyl alcohol-, hydrogen peroxide-, and ethanol-induced stress.

2. Materials and Methods

2.1. Yeast Strain

Kluyveromyces marxianus ITD-01005 was isolated during population dynamics monitoring of a spontaneous fermentation of Agave durangensis for mezcal production in Durango [28] and belongs to the strain collection of the Microbial Biotechnology Laboratory of TecNM-Technological Institute of Durango.

2.2. Cytoxicity Assays

Kuyveromyces marxianus ITD-01005 grew in YDP medium (yeast extract [Sigma-Aldrich, Burlington, MA, USA], 10 g/L; casein peptone [Sigma-Aldrich, Burlington, MA, USA], 20 g/L; dextrose [Sigma-Aldrich, Burlington, MA, USA], 20 g/L) at 28 °C, with shaking (150 rpm) for 12 h to produce biomass for toxicity assays, performed in duplicate. Cells were recovered by centrifugation (9660× g for 3 min) and resuspended in a fresh 2× YDP medium. A 96-well microplate was used to apply the treatments, placing 1 × 10⁵ cells per well and adding a 2× stress factor solution via a previously sterilized 0.22 μm filter (100 μL total volume per well). Treatments consisted of 10 initial concentrations of ethanol [Sigma-Aldrich, Burlington, MA, USA] (10, 25, 30, 35, 40, 45, 50, 55, 60, and 65 g/L), isoamyl alcohol [Sigma-Aldrich, Burlington, MA, USA] (1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 5.5, and 6 g/L), acetic acid [Sigma-Aldrich, Burlington, MA, USA] (0.3, 0.6, 0.9, 1.2, 1.5, 1.8, 2.1, 2.4, 2.7, and 3 g/L), and hydrogen peroxide [Sigma-Aldrich, Burlington, MA, USA] (68, 85, 102, 119, 136, 154, 170, 187, 204, and 221 mg/L). The positive control consisted of cells with full metabolic activity (no treatment). In contrast, the negative control consisted of the reaction mixture without cells. The microplate was incubated for 12 h at 28 °C, and residual viability was determined using two methods. The indirect method was an MTT (Sigma-Aldrich, Burlington, MA, USA) assay, while the direct method counted colony-forming units (CFUs) in Petri dishes.

2.2.1. MTT Assay

After incubation with the treatments, we added MTT (5 mg/mL) dissolved in phosphate-buffered solution, and the microplate was incubated for 4 h. Subsequently, acidic isopropanol [Sigma-Aldrich, Burlington, MA, USA] (HCl 0.04 N) was added to dissolve the obtained formazan crystals. The absorbance was read at 570 nm using a Biotek Synergy HT microplate hybrid reader (Winooski, VT, USA).

2.2.2. Colony-Forming Unit Counting

Serial dilutions up to 10⁵ were made from the content of each microplate well. The dilutions were seeded on Petri plates with YPD agar and incubated at 28 °C for 24 to 36 h until 25 to 250 colonies were obtained per plate for an accurate count.

2.2.3. Estimating the Inhibitory Concentration 30 (IC30)

The inhibitory concentrations, IC50 and IC30, were calculated following a procedure similar to that described by Bhatnagar and collaborators [29]. In brief, each concentration was converted to its log₁₀ and plotted against the residual yeast viability. Next, the tool Solver of Excel served to fit the four-parameter logistic model (Equation (1)).

$$X=d-{\displaystyle \frac{a-d}{1+{\left(\frac{{Log}_{10}{C}_i}{c}\right)}^b}}$$

(1)

where $X$ is the residual viability (%), $a$ is the minimum value of $X$ that can be obtained (%), $b$ is the Hill’s slope of the curve, $C_i$ is the concentration of each stress factor or inhibitor (g or mg/L), $c$ is the inflection point, and $d$ is the maximum value of $X$ that can be obtained (%). From Equation (1), the IC30 was calculated as the inhibitor concentration that allows a 70% residual viability.

2.3. Fermentative Capacity of the Yeast Under Stress Conditions

For each stress-inducing factor tested, the fermentative response of K. marxianus was evaluated under moderate and intense stress conditions. Moderate stress was assessed at IC30 concentrations, while IC50 concentrations of each stress factor were applied for intense cell stress. Fermentations started at a cell density of 1 × 10⁶ cells/mL in YDP medium, shaking at 150 rpm, at 28 °C, and lasted 12 h. The control consisted of one fermentation under the same conditions without stress factor. Kinetics were performed in triplicate, and samples were taken every two hours to quantify glucose consumption, ethanol production, and isoamyl acetate production.

The growth rate constant was calculated by fitting each growth kinetics to Equation (2), as reported previously [30]. It was performed through a nonlinear regression using Excel’s Solver tool.

$$X=X_0{e}^{\mu t}$$

(2)

where $X$ is the cell density (CFU/mL) at $t=t$, $X_0$ is the cell density at $t=0$, $\mu$ is the growth rate constant (h⁻¹), and $t$ is the time (h).

The procedure described previously [31] allowed calculating the first-order constant for glucose consumption by integrating Equation (3):

$$-{\displaystyle \frac{d{C}_G}{dt}}=k{C}_G$$

(3)

to obtain the following equation:

$$-Ln{\displaystyle \frac{C_G}{C_{G0}}}=kt$$

(4)

where $C_G$ is the glucose concentration (g/L), $C_{G0}$ is the glucose concentration (g/L) at $t=0$, $k$ is the first-order constant for glucose consumption (h⁻¹), and $t$ is the time (h).

The biomass concerning glucose yield ($Y_{X/S}$, cells/g glucose) was estimated as the slope of linear regression on the data plot of biomass versus glucose, including only the points for the exponential growth phase, as reported previously [32].

2.4. Analytical Techniques

An Agilent 1200 HPLC system (Agilent, Santa Clara, CA, USA) equipped with a refractive index detector was used to quantify glucose and ethanol concentrations by constructing a calibration curve with external standards. A volume of 1 µL was injected into the system to perform an isocratic separation at 70 °C on a Rezex ROA-Organic Acid H+ (8%) ion exclusion column (7.8 mm × 300 mm, Phenomenex, Torrance, CA, USA). The mobile phase was a 5 mM H₂SO₄ (Sigma-Aldrich, Burlington, MA, USA) solution at a 0.6 mL/min flow rate.

Isoamyl acetate quantification started with headspace solid-phase microextraction (HS-SPME) in 20 mL flasks. The filtered sample (5 mL) was mixed with 1.5 g of NaCl and incubated for 5 min at 35 °C. Then, SPME fiber (DVB/CAR/PDMS 50/30 µm, Supelco, St. Louis, MO, USA) was exposed in the headspace for 1 h at the same temperature. Volatile compounds were desorbed by exposing the fiber to the injection port of an Agilent 7890A gas chromatograph for 10 min at 250 °C. The injector was operated in splitless mode using an SPME liner (0.75 mm i.d., Supelco), and separation was performed on an FFAP column (30 m × 0.32 mm × 0.25 µm). Ultrahigh-purity helium (AOC Gases, Saltillo, Coah., Mexico) was used as the carrier gas at a constant 1 mL/min flow rate. The oven temperature was set at 40 °C for 3 min and then increased to 52 °C by 3 °C/min. The temperature of 52 °C was held for 1 min and then increased by 10 °C/min to 200 °C and held for 15 min. The mass spectrometry detector (Agilent 5975C) was operated at 230 °C, an ionization voltage of −70 eV, and in SCAN mode (1.6 scans per second). Isoamyl acetate was quantified using a calibration curve with external standards treated equally to the samples.

2.5. Statistical Analysis

One-way analysis of variance (p < 0.05) with Tukey’s test was used to compare means using the static Minitab software, version 20.3 (Minitab Inc., State College, PA, USA).

3. Results and Discussion

Yeasts are eukaryotic microorganisms with a versatile metabolism and a high capacity for producing biomolecules. Then, they are considered cellular factories in industrial biotechnology, capable of synthesizing high-commercial-value products from accessible substrates. However, during production processes, they are subjected to conditions and factors that cause stress, so it is helpful to evaluate their tolerance to the most frequent stress factors in fermentations [33]. For this work, four chemical species reported as inhibitors of yeast growth were chosen, since they represent relevant conditions in industrial and biological processes, providing a comprehensive approach to understanding the metabolic and physiological responses of this yeast [30,34]. Yeasts produce ethanol, but it is also highly toxic at certain concentrations. Mo and collaborators [35] report that the ethanol production capacity of a yeast depends on its tolerance to it. Acetic acid is a common byproduct in fermentation. In the processing of lignocellulosic biomass, acetic acid causes stress, affecting the proton gradient in the cell membrane. Its presence can decrease the intracellular pH and affect cell viability. The evaluation of the response of K. marxianus to this agent is relevant for applications where acid hydrolysates are used as a carbon source [10,36]. Isoamyl alcohol is a compound of industrial interest produced by yeasts and a precursor to isoamyl acetate, an aromatic compound of great interest in the cosmetic, pharmaceutical, and food industries [37]. Oxidative stress is one of the most common types of stress in any environment because ROS are derived from metabolic processes, so yeast survival depends on its ability to eliminate ROS [38].

3.1. Cytotoxicity of Stress Agents

Figure 1 shows the cytotoxic effect of the four selected stress compounds on Kluyveromyces marxianus ITD-01005. The data are presented as the residual viability percentage relative to a positive control (no stress) and a negative control (reaction mixture without cells).

This figure shows discrepancies between the direct and indirect methods in the residual viability measurements. The most significant discrepancies were observed in the cases of ethanol (Figure 1A) and most pronounced for hydrogen peroxide (Figure 1D). In the case of the isoamyl alcohol test, the discrepancy is intermediate (Figure 1C). The slightest discrepancy was observed in the test for acetic acid (Figure 1B). In general, an overestimation of viability was observed with the indirect method at low concentrations of the stress agents, which would lead to an overestimation of the half-maximal inhibitory concentration (IC50) value.

A similar discrepancy was previously reported by Fai and Grant [39], who determined the sensitivity of several yeasts to different toxic substances using a growth inhibition bioassay and an indirect method, both in 96-well microplates. The bioassay determined the decrease in cell density, measured as absorbance at 600 nm. These researchers used an indirect method based on the intracellular reduction of resazurin to the fluorescent compound resorufin. The IC50 values determined by the indirect method were usually lower than those obtained by the growth inhibition bioassay. The differences between yeasts could be associated with the Crabtree effect since reductases and the electron transport system mediate the reduction of resazurin. However, they did not find such a relationship since, for example, S. cerevisiae (Crabtree+) and K. marxianus (Crabtree-) both reduced resazurin slowly.

Residual viability progressively decreases as the concentration of stress agents increases. Viability is less than 10% for ethanol concentrations greater than 50 g/L (Figure 1A), while acetic acid concentrations greater than 2.4 g/L show viability close to 0, indicating a significant cytotoxic effect. Figure 1C shows that isoamyl alcohol concentrations higher than 5 g/L produce viability values of less than 10%. For hydrogen peroxide (Figure 1D), viability drops dramatically at concentrations higher than 102 mg/L. Similarly, in this case, viability values higher than 100% were observed with the indirect method for the first three concentrations tested, confirming the overestimation with the MTT assay.

Stepanenko and collaborators [40] reported that some inhibitors can directly interact with the MTT reagent or influence the metabolic pathways responsible for its reduction, leading to an overestimation or underestimation of cell viability. They also describe that, in response to stress under the treatment conditions, there can be metabolic and mitochondrial reprogramming, affecting the rate of MTT reduction and, therefore, the interpretation of the results. They suggest complementing the MTT assay with other methods not based on metabolic activity, such as trypan blue exclusion or lactate dehydrogenase release assays. For hydrogen peroxide, they explain that this compound can react directly with MTT, altering the assay result independently of cell viability. It can lead to overestimating cell viability due to oxidation–reduction reactions unrelated to normal cellular metabolism. Finally, these authors mention that in response to stress, yeasts activate diverse metabolic pathways depending on the type of stress, which could overexpress genes for proteins involved in MTT reduction and, therefore, overestimate viability.

The response mechanisms to different types of stress can overlap. For example, ethanol also induces oxidative stress responses in yeast [41]. These stress responses include the modulation of numerous cellular processes, gene regulation pathways, and changes in metabolite concentrations. Bleoanca et al. [42] evaluated the contribution of different defense mechanisms to the adaptation of several S. cerevisiae strains exposed to hydrogen peroxide- and ethanol-induced stress. They used qPCR to assess mRNA levels from the SOD1, TPS1, and GPH1 genes. SOD1 encodes a superoxide dismutase, TPS1 expression is linked to trehalose accumulation and protective activity under various stress conditions, and GPH1 expression is essential for glycogen metabolism. They found that SOD1, TPS1, and GPH1 mRNA levels increased 2.5-, 3.4-, and 2.4-fold, respectively, in cells exposed to H₂O₂. However, the increase in the relative expression of these genes was shown to be much greater (6-, 20-, and 19.3-fold, respectively) in cells subjected to ethanol stress. These results demonstrated that ethanol exposure can trigger oxidative stress, which could partially explain the overestimation of viability with MTT in ethanol assays.

Supplementary Figures S1 and S2 show the fits of the four-parameter logistic model to the experimental viability data obtained with the direct and indirect methods, respectively.

Table 1. Concentrations of ethanol, acetic acid, isoamyl alcohol, and hydrogen peroxide causing 50 (IC50) and 30% (IC30) loss in Kluyveromyces marxianus ITD-01005 cell viability. The R² values correspond to the fits of the four-parameter logistic model.

Stress Agent	Viability by Direct Method (CFU Counting)			Viability by Indirect Method (MTT)
	IC50 (g/L)	IC30 (g/L)	R2	IC50 (g/L)	IC30 (g/L)	R2
Ethanol	21.82	17.12	0.9854	36.70	32.79	0.9961
Acetic acid	1.19	0.97	0.9914	1.29	0.93	0.9805
Isoamyl alcohol	2.74	1.07	0.9630	2.84	2.40	0.9964
Hydrogen peroxide	0.09	0.06	0.9949	0.128	0.123	0.9907

shows the IC50 and IC30 values obtained for both methods, along with their corresponding R² values, ranging from 0.9630 to 0.9964, indicating that the model is suitable for estimating inhibitory concentrations.

The indirect method’s overestimation of viability also leads to overestimating the IC50 values. The IC50 values for the indirect method are between 3.7 and 68% higher than those for the direct method, with hydrogen peroxide and, significantly, ethanol being the stress agents that present the highest overestimations. Hydrogen peroxide is the most toxic agent evaluated, with an IC50 of 0.09 g/L. This high toxicity is due to generating ROS that cause oxidative damage to lipids, proteins, and nucleic acids, compromising cell viability.

After careful searching, no reports were found for IC50 in yeasts for the stress agents evaluated here. Therefore, to our knowledge, this is the first report presenting IC50 values for ethanol (21.82 g/L), acetic acid (1.19 g/L), isoamyl alcohol (2.74 g/L), and hydrogen peroxide (0.09 g/L) in yeast. It should be highlighted that these values were obtained by CFU counting, so this is viability in the strict sense of the microbiological term. It is also relevant to note that the indirect method based on MTT reduction is unreliable for this yeast determination.

Yeasts are subjected to ethanol toxicity, which acts on cell membranes and proteins and induces the generation of ROS [34,43]. A decrease in the growth rate has been reported when yeasts are exposed to different ethanol concentrations [44]. Song and collaborators [30] performed fermentations at 0.5 g/L and 2.5 g/L of isoamyl alcohol. They found that isoamyl alcohol causes stress in Saccharomyces cerevisiae from the concentration of 0.5 g/L, inhibiting yeast growth. Acetic acid is also reported as a strong inhibitor of yeast growth [10,45]. Fernandes and collaborators [10] reported that at concentrations of 3–4 g/L, cell viability and growth of Rhodotorula toruloides were drastically reduced due to osmotic damage, cytoplasmic acidification, and oxidative stress caused by acetic acid. Although no specific references were found on the toxicity of hydrogen peroxide in K. marxianus, it is widely documented that hydrogen peroxide negatively affects Saccharomyces cerevisiae [46].

3.2. Fermentative Capacity of the Yeast Under Stress Conditions

The growth and glucose consumption kinetics of K. marxianus ITD-01005 under moderate (IC30) and intense (IC50) stress conditions are shown in Figure 2.

Cell growth shows a progressive increase in all conditions during the 12 h of incubation, with the kinetics at the control condition showing sustained growth, reaching the highest level of cell density at the end of the period (Figure 2A,C). Moderate stress conditions (Figure 2A) affect yeast growth, with acetic acid and isoamyl alcohol being the compounds with less intense effects than hydrogen peroxide and ethanol, the most inhibitory agents for cell growth. Instead, Figure 2C shows that increasing the IC30 to IC50 induced slower growth, most notably in the kinetics under isoamyl alcohol and hydrogen peroxide stress. The effect on the kinetics under acetic acid stress was modest. In contrast, the kinetics with ethanol-induced stress showed practically no change when stress changed from moderate to intense.

On the other hand, glucose concentration shows a gradual decrease in all conditions, indicating continuous glucose catabolism during the 12 h of incubation (Figure 2B). Control and acetic acid-stress fermentations showed rapid glucose consumption, reaching concentrations close to zero around 10–12 h. In the other stress conditions, a slower decrease in glucose was observed, especially during the first 6 h, suggesting a slower metabolism due to chemical inhibition. Hydrogen peroxide and ethanol treatments affect glucose consumption more pronouncedly. At the same time, acetic acid and isoamyl alcohol show less marked effects, which coincide with cell growth. It suggests that chemical treatments affect yeast’s viability and metabolic activity, with hydrogen peroxide and ethanol being the most damaging agents. This phenomenon can be attributed to inhibiting key enzymes in the glycolytic pathway and cell damage induced by stressors. In the case of ethanol, it has been reported that its presence can alter cell membrane permeability, affecting nutrient absorption and, therefore, glucose metabolism [34].

Conversely, Figure 2D shows that glucose consumption under intense acetic acid stress was similar to that observed under moderate stress. In the case of the kinetics under hydrogen peroxide-induced stress, it was observed that increasing the stress level diminishes glucose consumption strongly, indicating that oxidative stress is particularly harmful to K. marxianus. In contrast, glucose consumption increased with increasing stress levels due to isoamyl alcohol and ethanol, particularly in the case of the last one. These observations indicate that K. marxianus adapts differently to the type and level of stress to which it is subjected. Although all stress tests affect growth and glucose consumption in K. marxianus, it tolerates acetic acid stress better. These observations indicate that K. marxianus adapts differently to the type and level of stress to which it is subjected. Although all stress tests affect growth and glucose consumption in K. marxianus, it tolerates acetic acid stress better. Notably, the response to ethanol-induced stress can prevent a significant increase in the applied stress level. It should be noted that, under IC50 stress conditions, yeast consumes all the glucose under all stress conditions. Higher glucose consumption at IC50 and reduced growth indicate that yeast is expending much energy operating defense mechanisms against the adverse environment. These results are relevant to define the tolerance of K. marxianus to stress conditions in industrial applications, such as fermentation in the presence of inhibitors derived from lignocellulosic substrates or fermentation products.

Table 2. Fermentation parameters (μ: growth rate constant—equal to Ln2 times the inverse of the duplication time; k: first-order constant for glucose consumption; Y_X/S: biomass concerning substrate yield; and isoamyl acetate production) at control conditions against the concentrations of acetic acid, isoamyl alcohol, hydrogen peroxide, and ethanol inhibiting 30% (IC30) and 50% (IC50) of cell viability. Values are expressed as mean ± standard deviation. Different superscript letters indicate statistically significant differences between values in the same row (p < 0.05) based on Tukey’s test.

			Stress Agent
		Control	Ethanol	Acetic Acid	Isoamyl Alcohol	Hydrogen Peroxide
$\mu$ (h−1)	IC30	0.549 ± 0.009 a	0.422 ± 0.007 c	0.550 ± 0.010 a	0.459 ± 0.010 b	0.437 ± 0.008 bc
	IC50	0.549 ± 0.009 a	0.413 ± 0.009 bc	0.442 ± 0.018 b	0.398 ± 0.016 c	0.342 ± 0.02 d
$k$ (h−1)	IC30	0.0547 ± 0.0026 a	0.0286 ± 0.0088 b	0.0207 ± 0.0018 b	0.0100 ± 0.0061 b	0.0308 ± 0.0129 ab
	IC50	0.0547 ± 0.0026 a	0.0260 ± 0.0032 b	0.0283 ± 0.0037 b	0.0236 ± 0.0032 b	0.0054 ± 0.0019 c
$Y_{X/S}$ × 10−7 (cells/g S)	IC30	1.400 ± 0.081 a	1.133 ± 0.085 b	1.197 ± 0.091 ab	1.305 ± 0.046 ab	1.141 ± 0.095 b
	IC50	1.400 ± 0.081 a	0.668 ± 00.018 b	0.707 ± 0.027 b	0.626 ± 0.022 b	0.371 ± 0.016 c
Isoamyl Acetate (mg/L)	IC30	40.90 ± 6.08 b	7.77 ± 0.27 b	26.52 ± 0.54 b	596.57 ± 41.60 a	71.40 ± 7.90 b
	IC50	40.90 ± 6.08 b	6.77 ± 0.56 c	23.94 ± 1.40 bc	238.61 ± 10.00 a	30.60 ± 0.12 b

shows the fermentation parameters, which quantitatively evaluate the effect of the kind and level of stress. The adjustments to calculate the different parameters had R² values ranging from 0.9377 to 0.9996. The growth rate constant value for the control conditions (0.549 h⁻¹) is similar to that reported previously (0.56 h⁻¹) (32) for K. marxianus grown in a mineral medium supplemented with vitamins. Ethanol, isoamyl alcohol, and hydrogen peroxide caused significantly lower $\mu$ values (p < 0.05), confirming that they are the agents with the highest toxicity. Regarding IC50 fermentations, all $\mu$ values were diminished considerably compared to the control (p < 0.05). Song et al. [30] reported a diminishment of 25.2% in $\mu$ value when exposed S. cerevisiae to isoamyl alcohol at 2.5 g/L. It is close to the decrease observed in the present study, 27.5% at IC50 (2.74 g/L). Concerning the first-order constant for glucose consumption ($k$), which is a measure of glucose consumption in the initial stage of fermentation (first 6 to 8 h), the control showed the significantly highest value (p < 0.05) compared to the treatments with stress agents at both inhibitory concentrations, IC30 and IC50. It is remarkable that the $k$ value strongly dropped (≈90%) when hydrogen peroxide stress increased from moderate to intense.

Biomass yield ($Y_{X/S}$) under stress conditions at IC30 is slightly lower than the control value (1.4 × 107 cells/g of substrate consumed). However, the difference is significant (p > 0.05) in the cases of ethanol and hydrogen peroxide stress. In contrast, at IC50, $Y_{X/S}$ decreases significantly (p < 0.05) under all conditions, again markedly with hydrogen peroxide (73.5%). It indicates that increasing stress levels affect the efficiency of metabolism and the conversion of substrate into biomass.

The control fermentation showed an isoamyl acetate production of 40.9 mg/L, the de novo production of K. marxianus ITD-01005. At inhibitory concentrations of IC30, ethanol and acetic acid decreased isoamyl acetate production, while hydrogen peroxide stress slightly increased it. Isoamyl alcohol at IC30 significantly increased isoamyl acetate production (p < 0.05). It should be noted that isoamyl alcohol is the most suitable precursor to produce isoamyl acetate (banana aroma) [47,48,49], which explains the high production of the aroma. However, it has also been reported that isoamyl alcohol negatively affects the growth and fermentative capacity of yeasts [47,50,51]. On the other hand, isoamyl acetate production decreased in fermentations with ethanol, acetic acid, and hydrogen peroxide when the stress increased from moderate to intense. In the case of isoamyl alcohol stress at IC50, a significantly higher production was observed compared to the control, but considerably lower than at IC30. It indicates that the increase in isoamyl alcohol concentration highlights its inhibitory effect over and above its role as the best precursor for this aroma production. A systematic study of the toxicity of this alcohol has not been reported. Therefore, the results reported here may improve the production of isoamyl acetate, helping to design an appropriate strategy for incorporating the precursor, isoamyl alcohol.

The strong effect of ethanol-induced stress at IC30 can be explained by considering the adverse effects of this compound. Ethanol interacts with the cell membrane, intercalating into the lipid bilayer and increasing the passive flow of protons through it. It alters the electrochemical gradient, decreases intracellular pH, and depolarizes the plasma membrane, hindering nutrient transport [34]. Interactions with ethanol also alter the structure and function of proteins. It hinders hydrogen bonding between amino acid residues, competes with water for the solvation of the protein surface, and disrupts interactions with hydrophobic residues, disrupting the hydrophobic core in globular proteins [52]. Ethanol stress also increases endogenous ROS generation and disrupts iron homeostasis in mitochondria [53]. However, it should be highlighted that the results presented here demonstrate that the response mechanisms to this type of stress showed virtually the same effectiveness when the stress level increased from IC30 to IC50.

In fermentation with hydrogen peroxide at IC30, isoamyl acetate production increased compared to the control. This result is consistent with the results obtained by Arellano et al. [54], who reported that oxidative stress increased ester production by two K. marxianus strains and two S. cerevisiae strains. Yeasts produce esters from alcohols (ethanol or higher alcohols) and acyl-CoA molecules (mainly acetyl-CoA) through the activity of enzymes specialized in ester synthesis [55]. The availability of acetyl-CoA, the composition of the medium, and the fermentation conditions (temperature, osmotic pressure, O₂, and CO₂ levels, as well as the presence of unsaturated fatty acids) influence the regulation of the synthesis of these compounds [56]. In S. cerevisiae, the presence of oxygen and unsaturated fatty acids reduces ester production because it decreases acetyl-CoA availability and inhibits the expression of the ATF1 and ATF2 genes [57]. However, under oxidative stress conditions, lipids undergo lipoperoxidation, which causes a reduction in fatty acids and, consequently, increases acetyl-CoA levels, thus favoring greater ester production. In the present study, in fermentation at IC30, oxidative stress negatively affected the $\mu$ value but increased isoamyl acetate production. It may be because yeast metabolism was induced to lipoperoxidation, increasing the concentration of acetyl-CoA and favoring the production of isoamyl acetate. When the level of oxidative stress increased to IC50, K. marxianus could not maintain an adequate response. It decreased the glucose assimilation capacity, biomass yield, and isoamyl acetate production fell below that observed for the control kinetics.

Overall, K. marxianus ITD-01005 responded better to isoamyl alcohol- and acetic acid-induced stress. The effect of higher alcohols on cell structure and function in yeast is considered to be similar to that attributed to ethanol [58,59]. A previous work [30] on isoamyl alcohol-induced stress in S. cerevisiae found that cell wall stability and cell membrane fluidity were affected by isoamyl alcohol. The authors concluded that gene expression related to ion homeostasis and energy production may be significant against isoamyl alcohol stress. On the other hand, acetic acid causes intracellular acidification but can be degraded by yeast through respiratory metabolism [36]. It could explain the relatively good fermentative capacity that K. marxianus ITD-01005 retained even at an inhibitory concentration of IC50.

4. Conclusions

The results of the present work demonstrated the overestimation of cell viability by the indirect method based on MTT reduction. Although this method is widely accepted in other areas, such as the growth of mammalian cells, it is inadequate to assess the toxicity of stress factors in yeasts. Therefore, despite the inconvenience of being time-consuming, direct CFU counting should be used to determine yeast viability during toxicity studies of chemical compounds.

This is the first report presenting accurate IC50 values for ethanol, acetic acid, isoamyl alcohol, and hydrogen peroxide in K. marxianus based on directly counting viable cells. Moreover, it was established that ethanol and hydrogen peroxide are the agents with the highest toxicity against K. marxianus ITD-01005 during fermentation at concentrations equivalent to IC30 and IC50 of each stress agent.