The Production of an Economical Culture Medium from Apple Pomace for the Propagation of Non-Conventional Cidermaking Yeast Starters

Abstract

The aim of the present study was to evaluate the use of different chemical treatments of apple pomace in order to produce an economical culture medium for the propagation of two non-conventional yeast strains. An experimental design approach was used for the optimization of the hydrolysis conditions of apple pomace. Both acid and alkaline treatment conditions were tested. The optimal hydrolysis conditions to disrupt the pomace lignocellulosic structure were 1% w/v of H₃PO₄, 121 °C, and 40 min for acid treatment, while 6% w/v of NH₄OH, 20 °C, and 2 h were optimal for the alkaline condition. Saccharomyces uvarum NPCC 1420 and Saccharomyces eubayanus NPCC 1292 yeasts were able to grow in the liquid fraction obtained from both acid and alkaline treatments. However, the medium with the acid treatment was found to be more suitable for yeast growth, showing, for both strains, higher µmax and biomass production and lower td and λ than that observed for the medium with the alkaline treatment. According to the growth parameter analysis for both strains, the acid treatment was selected for further studies. By taking advantage of this agroindustrial by-product, a circular economy approach is promoted, reducing environmental impact and fostering sustainable development.

1. Introduction

According to the Argentine Chamber of Integrated Fruit Growers [1], Argentina is the leading exporter of pears within the Southern Hemisphere and the fifth largest exporter of apples in the world. The 1.8 million tonnes production of these fruits, mainly located in the Upper Valley of Río Negro and Neuquén provinces, represents one of the most important economic activities in the Argentinian North Patagonia. Annually, 33% and 37% of pear and apple production, respectively, are destined for industries [1], including concentrated juice, cider, and dehydrated products production, among others. In recent years, cidermaking has become an industry of great importance in Norlitersgonia, with an annual production of around 80 million litters [2], representing approximately 54,000 tonnes of apples [3].

Both the concentrated juice and cider production industries generate a considerable amount of residues, mainly apple pomace (pulp, peels, seeds and stalks), representing 20–30% of the fresh weight [4]. This residue represents a serious disposal issue for this sector of the economy due to the high moisture content of organic matter; in fact, the humidity content of apple pomace after pressing is around 70–85%, which makes it prone to contamination [5]. The largest cidermaking company in the region generates about 6,500,000 kg of apple pomace per year [6]. Although possible biotechnological applications have been proposed for this kind of residue around the world, a large amount of it is still discarded as waste [7,8].

Apple pomace is a great source of soluble carbohydrates (e.g., glucose, fructose, and sucrose, among others), proteins, amino acids, fatty acids, phenolic compounds, vitamins, and other components considered nutritious [9]. For this reason, in recent years, there has been research into the generation of value-added products from this substrate. The most explored area is bioethanol production [10,11,12]; however, apple pomace has also been proposed for use in animal feed [13,14,15] and microbial enzyme production [16,17]. Also, the acquisition of microbial oils, polyphenols, organic acids, and several other compounds of biotechnological interest [8,18,19,20,21] and their application in the food industry [4,8] are being widely studied. Very little research has been conducted on the application of this type of waste itself as a base for culture media for microbial biomass production [22,23].

Thinking about a circular economy, cider residue could be used to produce yeast biomass that could then be used during fermentation. This has two interesting advantages; one is the use of the residue and the other is the saving of the purchase of commercial starters, which can be even more advantageous if yeasts are produced by the cider industry. In our group, we have studied regional cryotolerant yeasts of S. uvarum and S. eubayanus for the production of ciders at low temperatures. The strains NPCC 1420 and 1292 were selected from a great set of regional yeasts due to their remarkable oenological and kinetic features in fermentation conditions [24,25]. Developing an economical culture medium for the production of these yeasts is, without a doubt, an interesting biotechnological development for our region.

Apple pomace is also composed of insoluble carbohydrate polymers that constitute a chemical matrix mainly formed by cellulose (7–45%), hemicellulose (4–24%), and lignin (15–24%). In particular, the lignin presence in this residue acts as a shield around the cellulose and hemicellulose, preventing microbial and enzymatic attacks [26]. For this reason, the utilization of lignocellulosic agroindustrial wastes requires previous treatment to achieve the solubilization and separation of all structural components of the lignocellulosic network [7,27,28,29].

Many hydrolysis methods of lignocellulosic biomass have been explored, including physical, biological, chemical, or even a combination of these methods. The first includes fragmentation, grinding, shearing, or milling of the biomass lignocellulosic material [30,31], widely used for further biological bioconversion, involving the incorporation of hydrolytic enzymes [12,20]. However, enzymatic processes are not as efficient as chemical processes and are expensive, which make them less suitable for applications [32].

Regarding chemical treatment methods, the use of alkaline and acid compounds improves the hydrolysis of the lignocellulosic matrix. The alkaline compounds most evaluated for their use in applications are sodium, potassium, ammonia hydroxide, and hydrogen peroxide, among others [19,33,34,35,36], which were described as advantageous since they keep cellulose intact and enable delignification and hemicellulose solubilization, avoiding conversion to inhibitory compounds [7,27,37,38]. The advantage of using ammonium hydroxide is that the ammonium residues can be used as a source of nitrogen for culture media. Nevertheless, acidic lignocellulosic treatments have shown more effective results in breaking the glycosidic bonds between lignin and hemicelluloses, starting with the hydrolysis of hemicelluloses into its monomeric sugars, with the later depolymerization of the cellulose into oligosaccharides and further release of monomeric sugars [32]. Among them, both organic (acetic, citric, maleic or oxalic acid) or inorganic (phosphoric, nitric, hydrochloric or sulfuric acid) acids can be employed, with sulfuric acid being the most commonly applied [12,39,40,41,42,43]. Several advantages have been associated with the use of phosphoric acid, even in diluted concentrations, such as less harmfulness compared to sulfuric acid [29,44] as well as operational benefits, since the sodium phosphate formed during the neutralization with sodium hydroxide can be utilized by microorganisms as nutrients, with no need for it to be removed [29,45]. Additionally, acid recovery and even corrosion in equipment are solved with the use of this compound in its diluted form [45].

During most chemical treatments previously mentioned, several by-products are released, classified into three main groups: furaldehydes, weak acids, and phenolic compounds, which are known to affect subsequent fermentation and microbial growth when present in high concentrations and are, therefore, known as inhibitory compounds. Furaldehydes, such as furfural and 5-hydroxymethylfurfural (HMF), are derived from the dehydration of pentose and hexoses, respectively [32,39]. Furfural and HMF are derived from the dehydration of pentoses and hexoses, respectively [46,47,48], and their inhibitory effects on microorganisms are well-documented [49,50,51,52]. Several studies have shown that furfural and HMF reduce yeast enzymatic and biological activities, they break down DNA, and inhibit protein and RNA synthesis [20,53]. Organic acids including acetic, formic and levulinic acids are derived from sugars; in fact, acetic acid is formed by the acid hydrolysis of hemicellulose acetyl groups, while levulinic acid is the terminal product of hexose oxidation. However, formic acid has two pathways of formation, by the oxidation of pentose and hexoses [54,55,56]. In high concentrations, these weak acids can cause cytosol acidification, affecting metabolic activities, leading to cell growth inhibition and death [23,46,57,58]. Phenolic compounds are considered monomeric subunits of lignin, and they appear as phenolic acids, aldehydes, ketones, and alcohols. These compounds are generated during the hydrolysis of lignocellulosic materials, specifically from the breakdown of lignin and during the carbohydrate degradation. Phenolic compounds exert a considerable inhibitory effect on the fermentation processes of lignocellulosic hydrolysates, due to their low molecular weight which makes them highly toxic [47].

The aim of the present study was to evaluate the use of the different chemical treatments on apple pomace, the main organic waste produced in the regional cider industry, in order to produce a nutritive and inexpensive culture medium for the propagation of the two non-conventional yeast strains named Saccharomyces uvarum and Saccharomyces eubayanus. The culture medium was optimized by statistical experimental analysis, and both the growth parameters and the biomass yield of the two yeast strains were evaluated in order to select the best growing conditions. This study aims to make a contribution to knowledge but also to solve a problem that affects the environment and human health, such as waste generation, through the revaluation and the possible application of lignocellulosic wastes.

2. Materials and Methods

2.1. Apple Pomace

The apple pomace used in the present study was provided by a local cider factory located in Allen, Río Negro province, Argentina. For its utilization, the residue was dried at 60 °C for 48 h, milled to obtain a homogeneous sample, and stored at room temperature in closed plastic bags.

2.2. Experimental Design for the Acid and Alkaline Treatment

Two different chemical treatment methods were evaluated with the aim of removing lignin and degrading both the cellulose and the hemicellulose from apple pomace samples.

A screening test was carried out using the response surface methodology (RSM) via fractional factorial design (FFD) for the acid treatment optimization. Four factors with three levels: temperature (80, 103, 126 °C), solid/liquid ratio (1:35, 1:45, 1:55 g dry biomass/mL phosphoric acid solution), time (5, 23, 40 min), and phosphoric acid concentrations (1, 2.5, 4% v/v) were run to evaluate their influence on the total reducing sugar (TRS) and the glucose (GLU) concentration (response variables). The experimental design suggested 18 runs with different combinations, including the central point by duplicate.

Based on the screening results, further optimization was carried out using a central composite face-centered design (CCFD) (α = 1). The selected variables were temperature (121, 126, 131 °C), phosphoric acid concentration (0.2, 0.6, 1% v/v), and solid/liquid ratio (1:20, 1:25, 1:30 g/mL). A full 23 factorial design with 16 runs (including 2 central points) were employed, with TRS and GLU as response variables. The coded and uncoded values of the variables at various levels are given in

Table 1. CCFD of acid treatment matrix and observed values of the response variables.

STD	Temperature	Phosphoric Acid Concentration	Solid/Liquid Ratio	TRS	GLU
	(°C)	(%v/v)	(g/mL)	(g/100 g Dry Apple Pomace)
1	−1 (121)	−1 (0.2)	−1 (1:20)	39.43	5.64
2	+1 (131)	−1 (0.2)	−1 (1:20)	51.29	7.85
3	−1 (121)	+1 (1)	−1 (1:20)	48.72	12.57
4	+1 (131)	+1 (1)	−1 (1:20)	49.56	12.27
5	−1 (121)	−1 (0.2)	+1 (1:30)	40.47	6.27
6	+1 (131)	−1 (0.2)	+1 (1:30)	78.82 *	11.98 *
7	−1 (121)	+1 (1)	+1 (1:30)	54.66	14.02
8	+1 (131)	+1 (1)	+1 (1:30)	47.63	13.06
9	−1 (121)	0 (0.6)	0 (1:25)	52.72	11.65
10	+1 (131)	0 (0.6)	0 (1:25)	44.74	10.02
11	0 (126)	−1 (0.2)	0 (1:25)	46.73 *	8.95
12	0 (126)	+1 (1)	0 (1:25)	66.06	12.54
13	0 (126)	0 (0.6)	−1 (1:20)	66.86	11.42
14	0 (126)	0 (0.6)	+1 (1:30)	57.73	10.86
15(C)	0 (126)	0 (0.6)	0 (1:25)	55.99	11.06
16(C)	0 (126)	0 (0.6)	0 (1:25)	57.14	11.59

* Data ignored for ANOVA analysis, according to the DFFITS test; (C) central points.

. The model was explained by the following quadratic equation [59]:

$$Y={\beta }_0+{\displaystyle \sum }_{i=1}^k{\beta }_i{x}_i+{\displaystyle \sum }_{i=1}^k{\beta }_{ii}{x^2}_{ii}+{\displaystyle \sum }_{i=1<}^k{\displaystyle \sum }_{j=1}^k{\beta }_{ij}{x}_i{x}_j+\epsilon$$

where Y represents TRS and GLU, βo is the intercept coefficient, βi is the coefficient of the linear effect, βii is the coefficient of quadratic effect, and βij is the coefficient of interaction effect.

An alkaline treatment was performed using ammonium hydroxide and optimized by CCFD. A small 23 factorial design was performed, which suggested fifteen runs including five central points. The factors studied were temperature (4–12–20 °C), time (2, 10, 18 h), and ammonium hydroxide concentration (1, 3.5, 6% v/v). Again, TRS and GLU were evaluated as response variables (

Table 2. CCD of alkaline treatment matrix and observed values of response variables.

STD	Temperature	Time	Ammonium Hydroxide Concentration	TRS	GLU
	(°C)	(h)	(%v/v)	g/100 g Dry Pomace
1	+1 (20)	+1 (18)	−1 (1)	43.71	16.57
2	+1 (20)	−1 (2)	+1 (6)	52.03	15.81
3	−1 (4)	+1 (18)	+1 (6)	45.92	14.49
4	−1 (4)	−1 (2)	−1 (1)	43.71	15.11
5	−1 (4)	0 (10)	0 (3.5)	47.67	15.70
6	+1 (20)	0 (10)	0 (3.5)	45.95	15.50
7	0 (12)	−1 (2)	0 (3.5)	49.73	15.92
8	0 (12)	+1 (18)	0 (3.5)	46.52	14.16
9	0 (12)	0 (10)	−1 (1)	45.64	13.46
10	0 (12)	0 (10)	+1 (6)	47.81	11.75
11(C)	0 (12)	0 (10)	0 (3.5)	48.02	12.13
12(C)	0 (12)	0 (10)	0 (3.5)	47.24	13.21

(C) Central points.

). Design Expert Version 7.0 (Free Trial) was used for multiple regression analysis and plots of the obtained data. The desirability function [60] allowed the selection of the optimum working conditions from among several predicted ranges displayed by the software.

The validation of the model was performed by a new series of experiments under the corresponding optimum conditions selected. Assays were carried out by triplicate. The fit of the experimental data to the model was confirmed when the response variable values obtained were within the confidence interval.

2.3. Physicochemical Characterization

The characterization of the untreated apple pomace as well as both the acid and the alkaline treatment was carried out in the liquid and solid fractions by duplicate (maintaining the same S:L ratio for the three conditions). TRS, easily assimilable nitrogen (includes alpha amino acids and ammoniacal nitrogen), pH, sugars (glucose, fructose, sucrose, xylose, galactose, and arabinose), potential growth inhibitors (formic, acetic, and levulinic acids as well as 5-hydroxymethylfurfural (HMF) and furfural), minerals (sodium, phosphorus, magnesium, potassium, calcium, manganese, iron, copper, and zinc), total polyphenols, and extractive components were evaluated in the liquid fraction. Ash, moisture, and lignocellulosic components were determined in the solid fraction.

pH was determined potentiometrically using an ADWA AD 1030 pH meter (Instrumental Pasteur SRL, Buenos Aires, Argentina). The total polyphenol compounds were quantified spectrophotometrically using the Folin–Ciocalteu method [61]. Moisture was determined gravimetrically by subjecting the sample to a hot air oven at 105 °C to constant weight, and for the ash content estimation the sample was calcined in a muffle furnace at 580 °C for 4 h [62]. The extractive contents of the raw material were quantified using polar solvents (ethanol–water) attending to the procedure reported by the NREL/TP-510-42619 method [63]. The structural carbohydrates (holocellulose, cellulose, and hemicellulose) and the lignin quantification was performed following Standard Test Methods [64].

The easily assimilable nitrogen and the TRS were determined by the Sørensen method [65] and dinitro-salicylic method (DNS), respectively [66]. Sugars were determined by HPLC using a refractive index detector and a Sugar SP810 column (SHODEX, Tokyo, Japan) operating at 85 °C, and distilled water as a mobile phase (0.6 mL/min of flow rate). Potential growth inhibitor quantification was performed by HPLC using an Aminex-HPX87H column (BIO-RAD) operating at 45 °C, with 5 mM H₂SO₄ as an eluent (0.6 mL/min), quantified by UV spectrophotometry using a diode array detector. In all cases, the samples were previously filtered with a nylon filter with 0.22 μm porosity. The hydrolysates and the untreated samples were neutralized at pH 5, centrifuged, and filtered through a 0.45 μm filter.

Mineral contents were determined using an Agilent 7500 cx ICP-MS, equipped with an Agilent autosampler (ASX-500) (BIO-RAD, Hercules, CA, USA). Sample digestion was performed in a microwave digestor (Berghof, Speedwave 4) with 7 mL of 65% nitric acid (Merk), using Teflon flasks and following a two-step program: (1) a 5° ramp, 170 °C for 10 min; (2) a 1° ramp, 200 °C for 15 min. The digested samples were diluted for further measurements.

2.4. Yeast Growth Assays

S. uvarum NPCC 1420 and S. eubayanus NPCC 1292 from the North Patagonian Culture Collection (NPCC) were assessed for their growth capability in the treated (acid and alkaline treatment) and the untreated apple pomace liquid fractions. Before the yeast inoculation, the fractions were adjusted to pH 5 (with NaOH or H₃PO₄, as appropriate), filtered, and sterilized in an autoclave (120 °C for 20 min). 100 mL Erlenmeyer flasks containing 25 mL of either acid or alkaline treatment liquid fractions were inoculated with a pre-inoculum (the respective yeast culture growth for 24 h on GPY agar) to a final concentration of 1.106 cells/mL. The flask was incubated at 150 rpm in an orbital shaker at 20 °C and performed by duplicating. The microbial growth was estimated by measuring the optical density at 640 nm, and the dry weight was determined once the stationary phase was reached (constant weight at 105 °C). The growth parameters—maximum growth (A), maximum specific growth rate (µmax), stationary stage (lag), and duplication time (td)—were calculated by the reparametrized Gompertz equation [67].

2.5. Statistical Analysis

The growth kinetic parameters and the physicochemical profiles of apple pomace were analyzed by mean comparison using ANOVA and Tukey honest significant difference tests (a = 0.05) using the Statistica 8.0 software.

The analysis of the models and RSM were performed using Design Expert 7 Trial Free software, and model performance was checked by statistical analysis with a 95% confidence interval (ANOVA) (p-values ≤ 0.05) and the lack of fit test (p-value ≥ 0.05 and radj² as well as r² coefficients).

3. Results

3.1. Optimization of Acid and Alkaline Treatment Conditions for Apple Pomace

3.1.1. Acid Treatment

The screening using an FFD was performed in order to evaluate the effect of different factors (temperature, solid/liquid ratio, time, and phosphoric acid concentration) on the two response variables, TRS and GLU, considered to be relevant for our objective. The adequacy of the models was determined using the model analysis and R² (coefficient of determination) analysis. The R² values for GLU and TRS were 0.75 and 0.96, highlighting the high significance of the model.

The F value of the model was 9.82 for GLU and 78.90 for TRS, with a p-value less than 0.0007, which implies that the model is significant. Furthermore, the F value for lack of fit obtained for TRS (p-value of 0.08 > 0.05) and GLU (p-value of 0.80 > 0.05) was 88.62 and 0.52, so this implies that the lack of adjustment is not significant in relation to the pure error.

The analysis of estimated coefficients obtained from ANOVA test for both response variables revealed that temperature affected positively, while both the phosphoric acid concentration and the interaction between the two factors evidenced a negative effect. On the other hand, the time and solid/liquid ratio variables affected positively for GLU and TRS (

Table 3. Coefficients estimated by means of the ANOVA analysis for the responses obtained by FFD using four different variables for acid treatment.

Factors	Level		Coefficients 1
	(−1)	(+1)	TRS	GLU
Temperature (°C)	80	126	4.46	0.59
Time (min)	5	40	ns	0.50
Solid/liquid ratio (g/mL)	1/7	1/11	1.91	ns
Phosphoric acid concentration (%v/v)	1	4	−9.10	−0.61
Temperature * acid phosphoric conc.			−1.87	−0.64

¹ The coefficient estimate represents the expected change in response per unit change in factor value when all remaining factors are held constant. ns: not significant. The symbol * means the interaction between two factors

).

Considering this screening, the three factors showing significant effects on TRS and GLU were selected for further optimization using a CCD adjusted to the response surface methodology (Table 1). After ANOVA test of the data, the resulting second order polynomial equations were given:

$$Y \left(TRS\right)=60.76+1.46 x_1+13.91 x_1^2$$

$$Y \left(GLU\right)=19.27+5.17 x_2$$

where X₁ and X₂ are temperature and phosphoric acid concentration, respectively (solid/liquid ratio was not significant for the model). The interaction terms between the factors were not significant, so they should not be taken into account in the model.

The goodness of fit of the data supported by: (i) the model F-values of 14.34 (TRS) and 62.68 (GLU) imply that the model is significant, (ii) the lack of fit F-values of 28.85 (TRS) and 8.47 (GLU) imply the lack of fit is not significant in relation to the pure error (non-significant lack of fit is expected), and (iii) high values of R² coefficient (0.72 for TRS and 0.82 for GLU) ensured an adequate adjustment of the proposed model. Additionally, the R²-adjusted values were 0.82 (GLU) and 0.66 (TRS). The difference between R² values was less than 0.2, indicating a high significance of the model and the adequate adjustment to the experimental data. Three experimental points considered as outliers by the DFFITS test were ignored for the optimization study (Table 1). The evaluation of the DFFITS test is useful to detect the influence of each point on the predicted values and, thus, to determine outliers. In this type of plot, it determines the difference in model fit when a response value is removed [59].

The optimization conditions were selected using the desirability function based on the following criteria: (i) to keep the significant factors within the evaluated ranges, (ii) to set the solid/liquid ratio at 1:20, to obtain more concentrated sugars, and (iii) to maximize the GLU concentration. Based on the maximum desirability value (0.911), the working conditions selected were 121 °C and 1% (v/v) of phosphoric acid, represented by response surface method (Figure 1). The program yielded the predicted values of response variables of 13.27 g GLU/100 g of dry apple pomace and 54.01 g TRS/100 g of dry apple pomace. The validation analysis showed response values of 14.02 g GLU /100 g of dry apple pomace and 55.6 g TRS/100 g of dry apple pomace, evidencing an adequate adjustment of experimental data to the model.

3.1.2. Alkaline Treatment

An alternative strategy to disrupt lignocellulosic wastes was an alkaline treatment using ammonium hydroxide. A CCD was used to determine the optimum levels of the three factors (at three levels): temperature, time, and ammonium hydroxide concentration. In this case, the solid/liquid ratio factor was set at 1:20 and was not included in the design. The ratio was kept constant as it allowed an optimum experimental dilution to obtain a minimum volume of liquid fraction with concentrated sugars for the subsequent trials. Twelve experimental runs carried out in the CCD as well as the observed results for the responses corresponding to TRS and GLU are shown in Table 2. After the ANOVA test of the data, the resulting second order polynomial equations were the following:

$$Y \left(TRS\right)=46.99-0.85 x_1-1.55 x_2+1.09 x_3-1.55 x_1{x}_2-2.38 x_2{x}_3$$

$$Y \left(GLU\right)=13.44+0.43 x_1+2.09 x_1^2$$

where X₁, X_2, and X₃ parameters were temperature, time, and ammonium hydroxide concentration, respectively.

As it was observed for the acid treatment, experimental data were fitted to the model (significant p-values of TRS 0.0045 and of GLU 0.0335), being R² values 0.91 for TRS and 0.53 for GLU and not achieving a significant lack of fit (p-values ˃ 0.05).

To find the optimal conditions, the selection criteria were: (i) to keep the significant

Factors within the evaluated ranges and (ii) to maximize the GLU and TRS content. The most desirable condition, given by the desirability function (0.93), for the alkaline treatment was: 6% of NH₄OH, 20 °C at 2 h (Figure 2), with predicted values of 15.96 g for GLU/100 g of dry apple pomace and 52.71 g for TRS/100 g of dry apple pomace. The experimental values of the responses obtained were close to the predicted ones: 13.5 g for GLU/100 g of dry apple pomace and 46.97 g for TRS/100 g of dry apple pomace.

3.1.3. Physicochemical Characterization of Apple Pomace

The composition of both untreated and treated (acid and alkaline) apple pomace was analyzed. Among general constituents, the acid-treated apple pomace showed the highest contents of total polyphenols (6.22 ± 0.11 g/100 g of DAP) and easily assimilable nitrogen (771.46 g/100 g of DAP), which meant an almost twofold increase compared with the untreated samples (

Table 4. Physicochemical characterization of untreated, acid and alkaline treated apple pomace.

Constituents	Apple Pomace
	Untreated	Acid Treatment	Alkaline Treatment
General Composition
Total polyphenols	2.75 ± 0.06 b	6.22 ± 0.11 c	2.1 ± 0.08 a
pH	3.86 ± 0.15 b	2.26 ± 0.33 a	11.36 ± 1.90 c
Moisture *	79.30 ± 0.03	nd	nd
Easily assimilable nitrogen	339.81 a	771.46 b	nd
Ash *	2.38 ± 0.07 c	0.40 ± 0.14 a	1.34 ± 0.08 b
Extractives	65.34 ± 3.11 b	5 ± 0.76 a	5.25 ± 0.62 a
Alcohol-insoluble fraction of carbohydrate
Lignin	9 ± 1.23 a	8 ± 0.93 a	7.5 ± 0.74 a
Cellulose	17 ± 0.60 b	11 ± 0.83 a	13.4 ± 0.43 ab
Hemicellulose	7 ± 1.06 a	5 ± 0.70 a	6.4 ± 0.52 a
Alcohol-soluble fraction of carbohydrate
Glucose	7.91 ± 0.33 a	14.02 ± 0.63 b	10.00 ± 0.02 b
Fructose	17.08 + 0.07 a	26.02 ± 0.08 b	7.38 ± 0.05 a
Sucrose	ND	ND	ND
Xylose	0.17 ± 0.01 a	0.99 ± 0.30 b	ND
Galactose	ND	0.87 ± 0.22	ND
Arabinose Potential growth inhibitors	ND	2.95 ± 0.06	ND
Formic acid	ND	ND	0.71 ± 0.082
Acetic acid	0.13 ± 0.003 a	1.73 ± 0.100 c	0.59 ± 0.030 b
Levulinic Acid	1.04 ± 0.026	ND	ND
HMF	0.14 ± 0.001 a	1.09 ± 0.027 b	0.09 ± 0.002 a
Furfural	0.004 ± 0.000	ND	ND
Minerals
Sodium	<LDM	<LDM	<LDM
Phosphorus	64 ± 5 a	11800 ± 72 b	52 ± 5 a
Magnesium	14.40 ± 0.89 a	60 ± 1 b	<LDM
Potassium	340 ± 1 a	480 ± 2 c	360 ± 3 b
Calcium	<LDM	104 ± 6	<LDM
Manganese	0.07 ± 0.02 a	0.48 ± 0.02 b	<LDM
Iron	<LDM	4.8 ± 0.01	<LDM
Copper	0.38 ± 0.04 b	0.82 ± 0.01 c	0.14 ± 0.03 a
Zinc	<LDM	2.6 ± 0.01	<LDM

Data are expressed in g/100 g dry apple pomace, except in case indicated with asterisk (*) where data are in %. ND: not detected; nd: not determined; <LDM: limit detected measurements (LDM: Na 0.2 ug/g, Ca 2 ug/g, 28 ug/g, Mn 0.02 ug/g, Fe 0.7 ug/g, 1 ug/g). Superscript letters indicate significantly different (p < 0.05) between treatments.

). The extractives content decreased significantly (92%) in both acid and alkaline treatments, compared to the untreated apple pomace, while the ash content was significantly lower only after the acid treatment (6-fold) (Table 4). The content of each constituent of the alcohol-insoluble fraction was determined individually in order to evaluate the degree of removal of the lignocellulose network. The reduction in cellulose concentration was observed in both treatments while it was only significant for the acid treatment, showing 35% less cellulose compared to the untreated apple pomace. However, no significant differences were obtained in the content of lignin and hemicellulose. (Table 4). For the total sugars quantified in the alcohol-soluble fraction, an increase of 44% was observed in the acid treatment with respect to the untreated apple pomace. The most abundant sugars detected were glucose and fructose, followed by arabinose, xylose, and galactose. On the other hand, in the alkaline treatment, only glucose and fructose were quantified, and their values did not differ significantly from those observed in the untreated residue (Table 4).

Some acids and furfurans were also analyzed, among these toxic compounds only acetic acid and hydroxymethylfurfural (HMF) were detected in both treatments. However, the highest concentrations of acetic acid and HMF were detected in the acid treatment: 94% and 88% more than the untreated apple pomace, respectively. On the other hand, amounts of formic acid were detected only in the alkaline treatment (Table 4).

With respect to the nine minerals analyzed, only five elements (K, Cu, Mg, Mn, and P) were detected in the untreated apple pomace, while after the acid treatment three additional minerals were detected, including calcium, iron and zinc (Table 4). All five minerals detected in the untreated apple pomace increased their concentration significantly after the acid treatment (the increases ranged between 1.6- and 6.8-fold). In the case of phosphorus, its increase could not be quantified precisely since the acid treatment incorporates large amounts of phosphoric acid.

On the other hand, in the alkaline treatment only three minerals were detected including phosphorus, copper, and potassium, with the last being the only one that showed a significant increase compared to the untreated residue, while the other two compounds presented lower values than those observed in the untreated apple pomace (Table 4).

3.2. Yeasts Growth Assays Using Treated Apple Pomace as Substrate

In order to evaluate the possible biotechnological application of agricultural waste, cultivation studies in the liquid fraction obtained from apple pomace treatments were carried out to examine a potential use of this residue in yeast biomass production. Selected S. uvarum NPCC 1420 and S. eubayanus NPCC 1292 cryotolerant yeasts were able to grow in the liquid fraction obtained from both acid and alkaline treatments after the pH correction (Figure 3). However, the acid treatment medium was found to be more suitable for yeast growth, showing higher µmax and A values in both strains and lower td and λ values than those observed in the alkaline treatment medium (

Table 5. Kinetic parameters of S. uvarum NPCC 1420 and S. eubayanus NPCC 1292 obtained from their growth in acid and alkaline treatment of apple pomace liquid fraction.

Growth Parameters	NPCC 1292		NPCC 1420
	Acid Treatment	Alkaline Treatment	Acid Treatment	Alkaline Treatment
A	6.32(0.08) bB	4.79(0.08) aB	4.63(0.07) bA	4.01(0.03) aA
λ (h)	5.73(0.10) aA	6.74(0.07) bA	5.26(0.89) aA	7.28(1.34) bA
µmax (h−1)	0.20(0.003) bA	0.12(0.003) aA	0.19(0.002) bA	0.14(0.001) aB
Td (h)	3.46(0.10) aA	5.78(0.12) bB	3.65(0.03) aA	5.13(0.00) bA

Different lowercase letters at the top of the values indicate significant differences between treatments for each yeast by one-way ANOVA (p > 0.05). Different uppercase letters indicate significant differences between yeasts for each treatment.

). S. eubayanus NPCC 1292 showed greater growth than its sister yeast, showing an A value 35% higher than that observed for Su NPCC 1420 in the acid treatment medium. According to the growth parameter analysis for both strains, the acid treatment was selected for further studies.

To evaluate the effect of different sugar concentrations on yeasts growth and possible inhibitory compounds present in the liquid fraction obtained from the AT treatment, three culture medium conditions were studied: (A) the untreated apple pomace (2 g/L of TRS), (B) the diluted liquid fraction (2 g/L of TRS), and (C) the undiluted liquid fraction (3 g/L of TRS). The strains were able to grow in all the tested conditions (Figure 4). S. eubayanus 1292 reached the highest biomass development in the condition (C), with a significant difference in the A and ΔX values compared to the other two conditions as well as a high biomass yield (

Table 6. Growth parameters of Saccharomyces eubayanus NPCC 1292 and Saccharomyces uvarum NPCC 1420 obtained by growth in untreated medium and pretreated with phosphoric acid with 2% and 3% total reducing sugar.

Growth Parameters	NPCC 1292			NPCC 1420
	Untreated	Acid Treatment 2%	Acid Treatment 3%	Untreated	Acid Treatment 2%	Acid Treatment 3%
A	5.87(0.08) bB	4.97(0.00) aA	6.21(0.04) cB	5.05(0.00) aA	6.28(0.03) bB	5.28(0.10) aA
µmax (h−1)	0.27(0.006) bB	0.17(0.002) aB	0.17(0.000) aB	0.23(0.000) bA	0.15(0.005) aA	0.14(0.004) aA
ΔX (g L−1)	1.90(0.01) aA	2.32(0.16) aB	3.04(0.10) bB	1.61(0.03) aA	1.79(0.01) bA	1.75(0.01) bA
ΔS (g L−1)	18.46(0.2) bA	16.16(0.24) aA	24.07(0.33) cA	18.42(0.04) bA	15.9(0.19) aA	23.91(0.12) cA
Y (ΔX/ΔS)	0.10(0.001) aB	0.14(0.009) bB	0.13(0.001) bA	0.09(0.002) aA	0.11(0.001) cA	0.10(0.001) bA

Different lowercase letters at the top of the values indicate significant differences between treatments for each yeast by one-way ANOVA (p > 0.05). Different uppercase letters indicate significant differences between yeasts for each untreated culture medium and acid treatment with 2% and 3% of TRS.

). On the other hand, strain NPCC 1420 showed the best growth in condition (B), with higher values for A, ΔX, and Y compared to the other conditions, and with significant differences for A and Y (Table 6). It is worth mentioning that the two strains showed higher µ_max values (0.27–0.23 h⁻¹) when grown in the untreated condition compared to those obtained in the chemically treated conditions; however, their yields were lower (0.10–0.09) (Table 6).

4. Discussion

Many studies have shown the exploitation of agroindustrial residues to obtain different by-products of biotechnological interest mainly by harnessing metabolic activities of certain microorganisms [52,68,69,70]. Agroindustrial residues have been proposed as a substrate to produce bioethanol, microbial enzymes, and animal feed, among other products [8,12,15,20,68,71]; however, there are no studies registered in the literature that describe the use of apple pomace (a residue obtained from cider or concentrated apple juice industries) as a substrate for microorganism propagation. The liquid extracted from a secondary pressing of apple pomace has recently been proposed as culture media for the propagation of yeasts [72]. Nevertheless, this study does not make use of the full potential of this residue as a source for organic nutrients that could be exploited, after some inexpensive chemical treatments, for microorganisms growth. Through the use of experimental statistical design methods in the present work, we evaluated different hydrolysis treatments on a regional waste (apple pomace) trying to increase the bioavailability of its constituents. The final goal of this work was to produce an optimized culture medium for the propagation of yeast biomass and to reduce the organic residues amounts generated by regional agroindustrial activities.

On the other hand, the use of agroindustrial wastes as alternative substrates for the formulation of culture media offers an economical and sustainable solution that significantly reduces operating costs, with up to a tenfold reduction in raw material expenditure compared to conventional methods [73]. The use of these sub-products reduces dependence on expensive raw materials such as purified carbon and nitrogen sources, making them an efficient option for the industrial production of enzymes, biofuels, and other value-added compounds. Moreover, from an economic sustainability perspective, the reuse of these wastes not only minimizes production costs but also generates additional income by revalorizing products previously considered of low value [15].

In this work, the experimental design approach was suitable for the optimization of the hydrolysis conditions of apple pomace, allowing the factors involved in this process to be selected quickly and reliably. The search for suitable treatments for the hydrolysis of lignocellulosic residues remains a challenge for researchers. Most reports describe the effect of different factors individually on the residue treatment [12,19,74]. Only a few studies have applied experimental statistical design to define the chemical treatment conditions for the apple residues hydrolysis [29,75]. In our case, using FFD, CCD, and RSM analyses was effective to determine the optimal hydrolysis conditions within the experimental region, to obtain an increase in simple sugars from apple pomace. Both the acid and the alkaline treatment conditions were defined, and the significance of the model, the non-significant lack of fit, and the high R² coefficient values obtained ensured the validation of the models. According to our results, the optimal hydrolysis conditions to disrupt the pomace lignocellulosic structure were defined as 1% w/v of H₃PO₄, 121 °C, and 40 min for acid treatment, while 6% w/v of NH₄OH, 20 °C, and 2 h were defined for the alkaline condition.

Yeast growth depends on the availability of essential nutrients such as easily assimilable nitrogen, sugars, and minerals. Among these, the levels of nitrogen and sugars are considered to be the limiting factors for their growth [72]. In general, the aerobic yeast growth for biomass production requires a concentration of yeast-assimilable nitrogen in the range of 50 to 350 mg/L, depending on the yeast strain and environmental conditions [76]. In this context, the results obtained suggest that cider residues could be an economical and efficient substrate for biomass production, as the nitrogen and sugar contents are adequate (Table 4). In addition, the levels of K and Cu are within the required ranges, although Mg is insufficient for the yeasts. It is important to note that the concentrations of P, Ca, Mn, Fe, and Zn were at or above the requirements of yeasts only in the liquid fraction from the acid treatment [77]. In the future, these data could allow the optimization of the culture medium by the addition of specific salts.

As described in the results, the acid treatment evaluated in this work was substantially superior to the alkaline treatment in several aspects including the total amount of sugars released, total nitrogen availability, and assimilable minerals. The extract obtained after the acid treatment evidenced not only twice the amount of sugars (free glucose and fructose) detected in the untreated apple pomace, but also, a significant increase in the amounts of assimilable nitrogen and essential minerals indispensable for yeast growth. In contrast, the alkaline treatment was not effective in increasing either the simple carbohydrate fraction or the mineral content compared to the untreated apple pomace. Moreover, when studying the growth kinetics of the two yeasts in the liquid fractions obtained from both acid and alkaline chemical treatments, a lower biomass production was observed in the yeasts growing in the media obtained after the alkaline treatment. This treatment additionally presented some technical difficulties that proved to be really complex, mainly when trying to adjust the pH to the appropriate value for the development of the yeasts. All these aspects described previously were decisive in rejecting the use of alkaline treatment for subsequent studies.

Regarding the acid treatment of apple pomace, the conditions defined in this study differed from those proposed by other researchers [29,78]. Ucuncu and colleagues [29], also using experimental statistical design (RSM analysis), phosphoric acid, and similar temperature and time ranges as those tested in this work for hydrolysis of apple pomace, evidenced a total sugar yield five times higher than the amount present in the untreated pomace. The yield differences between our results and those obtained by Ucuncu et al. [29] may be attributed to the heterogeneity of the residue. The chemical composition of apple pomace is affected by numerous factors such as variety, ripening stage, physical and chemical properties of the apples, and juice extraction technologies among others [52,72].

During the thermochemical processes that use acid solutions, different inhibitory compounds are generally produced. Both the nature and the final concentration of these compounds vary greatly with the treatment conditions applied (temperature, time, type, concentration of acid, and pH) as well as the residue used [47,76,79,80]. Several authors have reported the presence of inhibitory compounds such as furfural and HMF in variable concentrations (0.06 to 0.95 g/L of furfural and 0.01 to 0.24 g/L of HMF) in the liquid fraction obtained after the acid treatment of apple pomace [75,81,82,83]. In our case, only minor amounts of HMF (0.5 g/L, equivalent to 1.09 g/100 g of DAP) and no furfural were detected. Similar results were shown by Parmar and Rupasinghe [75] but, under different apple pomace hydrolysis conditions, they used sulphuric acid at a concentration of 1.5% w/v, 91 °C, and 16 min of treatment. On the other hand, in our study, the lack of furfural could also be explained by the low concentrations of arabinose and xylose in our hydrolysates [29,84].

Other inhibitory compounds, such as phenolic compounds and weak acids (acetic, formic, and levulinic acids), can also be produced during the thermochemical process due to the de-acetylation of hemicellulose or HMF breakdown. In this study, low amounts of acetic (0.9 g/L, equivalent to 1.73 g/100 g of DAP) and formic acid (0.4 g/L, equivalent to 0.71 g/100 g of DAP) were observed after hydrolysis, being the formic acid only detected after the alkaline treatment. Formic acid has been described as the most toxic acid for the yeasts, more so than acetic and levulinic acids [85,86], which becomes an extra aspect to discard alkaline hydrolysis as a putative apple pomace treatment method. The increased toxicity of formic acid seems to be associated with a smaller molecule size, which may facilitate its diffusion through the plasma membrane and enables its higher anion toxicity. Hasunuma and colleagues [87] reported that 0.46–0.93 g/L concentrations of formic acid in the pretreated hydrolysate were sufficient to hinder fermentation by Saccharomyces cerevisiae, a species that is closely related to those studied in this work. Although the amount of formic acid in our alkaline liquid fraction was close to the lowest limit of that concentration range mentioned, the presence of this compound could have affected the yeasts growth in the medium performed with the alkaline treated pomace, as previously mentioned.

Concentrations higher than 2 g/L and 5 g/L of HMF and acetic acid, respectively, have shown inhibitory effects on S. cerevisiae growth [49,58,86,88,89]. The HMF and acid acetic concentrations found in our study were significantly lower than those previously mentioned. Although some studies have shown the effects of HMF on other yeast species different from S. cerevisiae, there is currently no information about the effect of this compound on the species used in this work, including both S. uvarum and S. eubayanus. The effect of HMF and other inhibitory compounds on the growth of different microorganisms has been reviewed by many authors, and the levels of inhibition reported vary strikingly with inhibitor concentrations and microbial strains [46,90,91,92,93].

The next step in our study was to evaluate the growth of the strains in different dilutions of the liquid fraction obtained from the acid treatment and the possible effect of inhibitory substances. Our results show that each strain exhibited its maximum biomass production under different conditions; for the S. eubayanus NPCC 1292 strain, the optimum growth was observed in a medium containing 3% w/v of TRS while for S. uvarum NPCC 1420 strain, the optimum was obtained in a medium with 2% w/v of TRS. Interestingly, the growth rates were lower, but the yield (Y) was higher than the one obtained after culturing the same two strains on untreated pomace. These results coincide with those observed by other researchers [46,49,51,90,94]. Additionally, these authors evidenced that cultures supplemented with inhibitory compounds significantly affected the yeast growth rates but did not have an impact on the total yield. Most toxicity studies were performed by evaluating the individual effect of an inhibitory compound on growth, fermentation, or productivity [49,51]. Although there are not many reports evaluating the joint impact of these inhibitory compounds on yeast growth, there are some observations showing that these compounds, even if present in low quantities as could be the case in our study (HMF, acetic acid, polyphenols), could give rise to a synergistic inhibitory effect [58].

In conclusion, the results of this work demonstrate that apple pomace residues, when subjected to appropriate chemical treatment, have the potential to become a valuable and economical culture medium for the propagation of yeasts belonging to the unconventional yeast species S. uvarum and S. eubayanus. The optimized acid treatment not only allowed for a greater release of simple sugars, but also for a significant increase in essential nutrients for yeast growth, far outperforming the alkaline treatment. These findings suggest that the liquid fraction obtained after acid hydrolysis of apple pomace can serve as a basis for an economical and efficient culture medium. By taking advantage of this agroindustrial by-product, a circular economy approach is promoted, reducing environmental impact and fostering sustainable development.