**Bioethanol Production from Food Waste Applying the Multienzyme System Produced On-Site by** *Fusarium oxysporum* **F3 and Mixed Microbial Cultures**

**George Prasoulas 1, Aggelos Gentikis 1, Aikaterini Konti 2, Styliani Kalantzi 1, Dimitris Kekos <sup>1</sup> and Diomi Mamma 1,\***


Received: 27 February 2020; Accepted: 24 March 2020; Published: 26 March 2020

**Abstract:** Waste management and production of clean and affordable energy are two main challenges that our societies face. Food waste (FW), in particular, can be used as a feedstock for the production of ethanol because of its composition which is rich in cellulose, hemicellulose and starch. However, the cost of the necessary enzymes used to convert FW to ethanol remains an obstacle. The on-site production of the necessary enzymes could be a possible solution. In the present study, the multienzyme production by the fungus *Fusarium oxysporum* F3 under solid state cultivation using different agroindustrial residues was explored. Maximum amylase, glucoamylase, endoglucanase, b-glucosidase, cellobiohydrolase, xylanase, b-xylosidase and total cellulase titers on wheat bran (WB) were 17.8, 0.1, 65.2, 27.4, 3.5, 221.5, 0.7, 0.052 and 1.5 U/g WB respectively. *F. oxysporum* was used for the hydrolysis of FW and the subsequent ethanol production. To boost ethanol production, mixed *F. oxysporum* and *S. cerevisiae* cultures were also used. Bioethanol production by *F. oxysporum* monoculture reached 16.3 g/L (productivity 0.17 g/L/h), while that of the mixed culture was 20.6 g/L (productivity 1.0 g/L/h). Supplementation of the mixed culture with glucoamylase resulted in 30.3 g/L ethanol with a volumetric productivity of 1.4 g/L/h.

**Keywords:** bioethanol; food waste; on-site enzyme production; *Fusarium oxysporum* F3; *Saccharomyces cerevisiae*; mixed culture

#### **1. Introduction**

Food waste has been identified as a big economic, social and environmental problem nowadays. According to the official statistics published by Eurostat, each year, more than 240,000 t of waste is produced in the EU [1]. Bio-waste, the organic fraction of municipal solid waste, i.e., garden, kitchen and food waste, accounts for one third of the total waste and is considered to be a valuable resource that could be used as raw material for the production of high value-added products. This fact is also reflected in the updated Bioeconomy Strategy of the EU. A sustainable bioeconomy can turn bio-waste, residues and discards into valuable resources and can create innovations and incentives to help retailers and consumers cut food waste by 50% by 2030 [2]. However, the use of food by-products and the conversion of food waste is still limited. This is due to current limitations in its quantification along the food supply chain, limited data on its quality and level of homogeneity, and differences in national implementations of the waste legislation [2].

The composition of food waste, as already said, is not stable. It presents significant variations related to the season, the area, and the dietary habits of the population. Despite the inevitable variation in the composition of food waste, it can indisputably be said that it is rich in carbohydrates, proteins, lipids and minerals which make it an ideal raw material for the production of biofuels through microbial conversion [3,4]. The exploitation of food waste for the production of biofuels is also in line with the 2030 Agenda for Sustainable Development set by the UN in 2015 [5]. More precisely, it is directly related to the Sustainable Development Goals: 7. Affordable and Clean Energy, 12. Responsible Consumption and Production and 13. Climate Action. At the EU level, the importance of producing biofuels from bio-waste is reflected in the Renewable Energy Directive 2009/28/EC [6] and the recently adopted recast of the Renewable Energy Directive a.k.a. RED II [7]. The aforementioned legislation defines as advanced biofuels the 'biofuels that are produced from the feedstocks listed in Part A of Annex IX' that includes among others, the biomass fraction of municipal solid waste, the biomass fraction of industrial waste as well as bio-waste from households. The directive sets a sub-target of 3.5% for advanced biofuels within the 14% target for renewable energy in transport in 2030. Moreover, those biofuels will continue to count double towards the targets.

From a technical point of view, bioethanol production from lignocellulosic materials is a wellstudied process and has been recently reviewed [8]. It includes the following processes: pretreatment, enzymatic hydrolysis, fermentation and ethanol recovery. The pretreatment phase aims at modifying the structural characteristics of the raw material facilitating the enzymes' access and maximizing sugar monomers production. The enzymatic hydrolysis targets the structural carbohydrates starch, cellulose and hemicellulose. During this step, pentoses and hexoses that can be further used in the fermentation step are liberated. In the subsequent fermentation step, microorganisms metabolize those readily available sugars, producing ethanol, which is subsequently recovered through distillation.

In terms of cost, the most demanding step, which significantly increases the total cost of the production of bioethanol and is identified as a barrier in the further deployment of ethanol production, is enzymatic hydrolysis [9]. A possible solution to this problem is the on-site production of the relevant enzymes instead of using commercially available enzymes [10]. Few organisms are able to produce the necessary enzymes; most of them belong to the species *Aspergillus sp., Penicillium sp., Trichoderma sp. Neurospora sp*. [11].

The filamentous fungus *Fusarium oxysporum* could be used both for cellulolytic enzyme production and for ethanol production because of its ability to ferment both hexoses and pentoses [12,13]. The main aim of the present work is, on the one hand, the induction of the metabolic system of *F. oxysporum* to produce the relevant enzymes by using different raw materials (namely wheat straw, wheat bran, corn cob), and on the other hand, to exploit those enzymes for the hydrolysis of food waste (FW) in order to produce ethanol. Focusing on increasing ethanol production, the addition of a glucoamylase was studied, as well as the use of mixed *F. oxysporum* and *Saccharomyces cerevisiae* cultures.

#### **2. Materials and Methods**

#### *2.1. Raw Materials, Reagents, Microorganism*

The food waste (FW) used in the present study was provided by Professor Gerasimos Lyberatos (Organic Chemical Technology Laboratory, School of Chemical Engineering, National Technical University of Athens), in dry form and with an average particle size of approximately 3 mm. FW were household food wastes from the Municipality of Halandri, Greece. The concept of drying focuses on the dehydration of the material resulting in a significant reduction of its mass and volume, thus facilitating its storage and protecting the readily fermentable sugars by inhibiting microbial activity due to low moisture content.

Wheat straw (WS), wheat bran (WB), and corn cobs (CC) were chopped into particles of less than 3 mm in diameter.

All chemicals were of analytical grade. Commercial glucoamylase (Spirizyme® Fuel) was kindly provided by Novozymes Corporation (Denmark).

The laboratory strain F3 of F. oxysporum, isolated from cumin [14], was used in the present study. The stock culture was maintained on potato-dextrose agar (PDA).

#### *2.2. Chemical Analysis of FW*

Moisture, ash, crude fat, crude protein (Kjeldhal method) and total starch content were determined according to standard methods [15]. Pectic polysaccharides were determined according to Phatak et al. [16], while water-soluble materials, cellulose, hemicellulose and acid insoluble lignin content as described by Sluiter et al. [17]. Analysis was carried out in triplicate.

#### *2.3. Pretreatment of FW*

Dry FW were pretreated as described elsewhere [18]. Briefly, FW at concentration of 30% w/v was pretreated at 100 ◦C for 1 h in the presence of 1 g sulfuric acid/100 g of dry FW. Following pH adjustment to 6.0, the pretreated material (the whole slurry) was used in fermentation experiments.

#### *2.4. Media and Growth Conditions for the in Situ Production of Enzymes*

Solid-state cultivation (SSC) was carried out in 100-mL Erlenmeyer flasks containing 3.0 g of dry carbon source (WB, WS or CC) moistened with Toyama's mineral medium (in g·L<sup>−</sup>1: (NH4)2SO4, 10; KH2PO4, 3; MgSO4·7H2O, 0.5; CaCl2, 0.5) [19]. The initial culture pH was adjusted to 6.0 and the moisture level at 75%. Following heat sterilization (121◦C) for 20 min, each flask was inoculated with 1 mL spore suspension (approximately 107 conidia) and incubated at 30 ◦C under static conditions. Experiments were carried out in duplicate.

#### *2.5. Enzyme Extraction*

After suitable periods of time, enzymes were extracted from the SSC with 10-fold (v/w) 50 mM citrate-phosphate buffer pH 6.0 by shaking (250 rpm) at 25 ◦C for 60 min. The suspended materials and fungal biomass were separated by centrifugation (12,000 x g at 4 ◦C for 15 min) and the clarified supernatant was used for enzyme activity measurements.

#### *2.6. Enzymatic Hydrolysis of FW*

SSC at maximum enzyme production was supplemented with the pretreated FW (SSC to FW ratio, 1/10 w/w) and the appropriate amount of the commercial glucoamylase (Spirizyme® Fuel) in order to achieve 20 and 40 Units of glucoamylase per g starch. Microbial contamination was prevented by the addition of sodium azide (0.02% w/v). Hydrolysis was performed at 50 ± 1 ◦C in a rotary shaker (250 rpm). Samples were withdrawn periodically, centrifuged (10,000 x g for 10 min), and analyzed for glucose.

#### *2.7. Conversion of FW into Bioethanol*

Food waste was converted into ethanol applying of a two-phases process where enzymes production under SSC was combined with simultaneous saccharification and fermentation (SSF) of FW by the mesophilic fungus F. oxysporum F3 or by a mixed culture of the latter with the yeast S. cerevisiae. Initially F. oxysporum F3 was grown under SSC, as described above for the production of the cellulolytic, hemicellulolytic and amylolytic enzymes. Whole SSC (fungal mycelia and the in situ produced enzymes) was transferred to the pretreated FW (SSC to FW ratio, 1/10 w/w). Fermentation was carried out in a rotary shaker operating at 30 ± 1 ◦C and 80 rpm in Erlenmeyer flasks provided with special rubber stoppers, which ensured anaerobic conditions and allowed release of produced carbon dioxide.

In the case of mixed microbial culture compressed baker's yeast (Yiotis, Athens, Greece) corresponding to 15 mg yeast per gram of initial dry FW was added.

#### *2.8. Analytical Methods*

Endoglucanase (EG), exoglucanase (EXG), xylanase (XYL), total cellulase (FPU) and amylase (AMYL) activities were assayed on carboxymethyl cellulose, Avicel, birchwood xylan filter paper and starch respectively, as described [18,20]. The activities of β-glucosidase (β-GLU) β-xylosidase (β-XYL) and glucoamylase (GLAMYL) were determined spectrophotometrically using the respective p-nitrophenyl glycosides as substrates [18,20]. All assays were carried out at 50 ◦C and pH 5.0. Blanks with inactivated enzyme (after boiling for 15 min) were used as a reference.

One unit (U) of enzyme activity was defined as the amount of enzyme liberating 1 μmole of product per min.

Glucose concentration was determined using a commercially available kit (Biosis S.A., Athens, Greece) that employed the Glucose Oxidase–Peroxidase (GOX–PER) method.

Ethanol analysis was conducted in an Aminex HPX-87H (Bio-Rad, 300 x 7.8 mm, particle size 9 μm) chromatography column. Mobile phase was 5 mM H2SO4 in degassed HPLC- water at a flow rate of 0.6 mL/min and column temperature was 40 ◦C [21].

#### **3. Results and Discussion**

#### *3.1. Food Waste Composition*

In general, the composition of food waste is complex and includes oil and water, as well as spoiled and leftover foods from kitchen wastes and markets. These substances are mainly composed of soluble sugars, carbohydrate polymers (pectin, starch, cellulose and hemicelluloses), lignin, proteins, lipids and a remaining, smaller inorganic part. High moisture content leads to rapid decomposition of the organic wastes and the production of unpleasant odors, which can attract flies and bugs, which are vectors for various diseases [22]. The feedstock used in the present study was provided in dry form as mentioned above. Compositional analysis of FW is presented in Table 1.


**Table 1.** Chemical analysis of dried FW.

<sup>a</sup> Total reducing sugars: 4.96 ± 0.94 % (w/w, dry basis), sucrose: 0.51 ± 0.04 % (w/w, dry basis), protein: 0.33 ± 0.02 (w/w, dry basis) and soluble starch: 1.15 ± 0.09 (w/w, dry basis).

The composition of FW, carbohydrates, protein, makes it an excellent feedstock for the production of biofuels and bio-based chemicals through microbial conversion. The polysaccharide content of FW is difficult to be used by ethanol producing microorganisms such as Saccharomyces cerevisiae. Different commercial enzymes (amylase, glucoamylase, carbohydrase, cellulase) have been used to improve the saccharification of FW [23].

#### *3.2. Multienzyme Production under Solid-State Cultivation*

To make the enzymatic hydrolysis of FW more cost-effective, the enzymes should be produced on-site from a cheap feedstock [11]. SSC has several biotechnological advantages such as higher fermentation capacity, higher end-product stability, lower catabolic repression and cost-effective technology [24,25]. SSC is an attractive process for filamentous fungus cultivation because the solid substrates have characteristics similar to the natural habitat of the fungi, resulting in improved growth and secretion of a wide range of enzymes. Selecting the appropriate substrate is an extremely important aspect of SSC, as the solid material will act as a physical support and nutrient source [26]. In the present study three different agroindustrial residues, wheat straw (WS), wheat bran (WB) and corn cobs (CC) were evaluated for the production of cellulolytic, hemicellulolytic and amylolytic enzymes by the mesophilic fungus *F. oxysporum* F3. Wheat bran (WB), a nutrient-richer intermediate of the wheat processing industry, was the most effective carbon source for multi-enzyme production by *F. oxysporum* F3 (Figure 1). Implementation of corn cobs (CC) and wheat straw (WS) particles as substrates was associated with lower enzyme titres. Maximum enzyme activities were recorded in the 5th day of fermentation. Amylase activity was found 5- fold higher, than that produced on WS and CC, while traces of glucoamylase was produced on WS and CC. Maximum endoglucanase, b-glucosidase cellobiohydrolase, xylanase and b-xylosidase titers on WB were 65.2, 27.4, 3.5, 221.5, 0.7, 0.052 U/g WB, respectively, while total cellulase activity was 1.5 FPU/g WB.

**Figure 1.** Production of cellulolytic, hemicellulolytic and amylolytic enzymes by *F. oxysporum* F3 grown under solid state cultivation on wheat straw, wheat bran and corn cobs as carbon sources.

For comparison, *F. oxysporum* F3 grown under SSC on corn stover as carbon source resulted in final endoglucanase, b-glucosidase, cellobiohydrolase, xylanase, b-xylosidase titers of 211, 0.088, 3.9, 1216, 0.052 U/g, respectively [27]. Futhermore, *Trichoderma virens* grown on alkali-treated WS under SSC produced 123.26 and 348 U/g endoglucase and xylanase, respectively [28].

It is well documented that the type and composition of the carbohydrates present in WB are suitable for induction of cellulases, hemicellulases and amylases from filamentous fungi under SSC [24,25]. Since the target of the present study is the production of a multi-enzyme system capable of hydrolysing the main polysaccharides present in FW, WB was chosen for further experiments.

#### *3.3. Hydrolysis of FW by the On-Site Produced Multienzyme System of F. Oxysporum F3*

Multienzyme system of F. oxyporum F3 produced under SSC on WB as carbon source was evaluated in FW hydrolysis. Furthermore, due to low glucoamylase titer produced the hydrolysis mixture was supplemented with 20 and 40 Units/g FW. The concentration of glucose increased gradually with time and reached a constant value at 69 h (Figure 2a). Glucose release applying F. oxysporum enzyme system was found 33.7 g/L corresponding to a hydrolysis yield of 47.4% (based on cellulose and starch content of FW). It should be mentioned that glucose release due to pretreatment of FW accounted for 11.2 g/L. Supplementation of hydrolysis mixture with glucoamylase (Spirizyme® Fuel) increased glucose release by approximately 25%.

**Figure 2.** (**a**) Time course of glucose release during FW hydrolysis by *F. oxysporum* F3 multienzyme system (•), supplemented with 20 (-), 40 (-) Units/g starch Spirizyme® Fuel. (**b**) Initial rate of glucose release.

The amount of added glucoamylase affects the initial rate of glucose release (Figure 2b) during hydrolysis. Addition of 20 U/g starch glucoamylase resulted in a 66% improvement in the rate of glucose release while further increase in glucoamylase load did not improve the result.

#### *3.4. Bioethanol Production by Mixed Microbial Culture of F. Oxysporum F3 with the Yeast S. Cerevisiae*

The processes generally used in bioethanol production are simultaneous saccharification and fermentation (SSF) and separate hydrolysis and fermentation (SHF). Performing hydrolysis and fermentation in a single step, the SSF process, has several advantages over SHF. In SSF, end-product inhibition of b-glucosidase is avoided, and the need for separate reactors is eliminated [29]. In the present study FW were converted to bioethanol applying SSF process in batch mode. Ethanol production from FW by mono-culture of *F. oxysporum* F3 reached 16.3 g/L after 94 h of fermentation (Figure 3a) corresponding to 31.8% of the maximum theoretical based on the soluble fraction and the carbohydrate content (cellulose, starch and hemicellulose) of FW (Figure 4).

**Figure 3.** Bioethanol production from FW by (**a**) mono-culture of F. oxysporum F3 (•), supplemented with 20 (-) and 40 (-) Units/g starch Spirizyme® Fuel; and (**b**) mixed microbial culture of F. oxysporum F3 with the yeast S. cerevisiae (•), supplemented with 20 (-) and 40 (-) Units/g starch Spirizyme® Fuel.

**Figure 4.** Maximum values of bioethanol production (g/L) (), bioethanol yield (g/100 g FW) (), theoretical yield (%) ().

When the fermentation medium was supplemented with 20 U/g starch glucoamylase, ethanol titer was 23.9 g/L (corresponding to 46.8.0% of the maximum theoretical). Increase in glucoamylase supplementation did not improve bioethanol production (Figure 3a). Maximum ethanol production was recorded at 139 h of fermentation, and no further increase in ethanol concentration was found when fermentation was extended to 163 h. Bioethanol volumetric productivities were in the range of 0.17 to 0.23 g/L/h (Figure 5).

**Figure 5.** Bioethanol volumetric productivities.

To improve the sugar assimilation rate, mixed culture of *F. oxysporum* F3 with the yeast *S. cerevisiae* were applied with the same bioconversion setup. The results are presented in Figure 3b. As can be seen, ethanol concentration at 18 h of process was about 18.3 g/L in the mixed culture of *F. oxysporum* F3 with *S. cerevisiae* approximately 7.5 times higher than the corresponding value achieved with the mono-culture of *F. oxysporum* F3. Fermentation was completed after 42 h using mixed microbial culture and ethanol production reached 20.6 g/L (corresponding to 40.1% of the maximum theoretical) (Figure 4). Ethanol concentrations of about 26.8 and 25.1 g/L were reached at 18 h when the mixed cultures were supplemented with 20 and 40 U/g starch of glucoamylase, while the corresponding values in mono-cultures were in the range of 1 g/L. Maximum ethanol production of 29.9 and 30.8 g/L was achieved at 69 h (Figure 3b). At the end of all fermentations, glucose concentration in the fermentation

broth was lower than 2.0 g/L. Volumetric productivities ranged from 1.0 g/L/h (mixed microbial culture) to 1.5 g/L/h (mixed microbial culture supplemented with 20 U/g glucoamylase) (Figure 5).

It is evident that mixed microbial culture decreases significantly the time needed for the fermentation to be completed therefore increasing volumetric productivity. Mixed cultures of mesophilic fungi, such as *F. oxysporum* F3 and *N. crassa*, with the yeast *S. cerevisiae*, have been successfully implemented for the bioconversion of sweet sorghum and sweet sorghum baggasse to bioethanol [30,31]. Table 2 summarizes results obtained from the bioconversion of different kinds of FW.


**Table 2.** Production of bioethanol from FW applying different processes.

<sup>1</sup> Cellulolytic enzymes produced from the thermophillic fungus Myceliophthora thermophila, <sup>2</sup> cellulolytic enzyme system <sup>3</sup> amylolytic, cellulolytic and hemicellulolytic enzymes produced by Aspergillus awamori.

Bioethanol production of the present sudy was 30.8 g/L which is higher than that reported by Matsakas and Christakopoulos [33] using the on-site produced cellulolytic enzymes from the thermophillic fungus Myceliophthora thermophila. On the other hand, Kiran and Liu [37] achieved much higher ethanol production (58.0 g/L) using the multienzyme system produced on-site by the fugus Aspergillus awamori. Wang et al. [36] used response surface methodology to optimize the conditions of SSF for ethanol production from kitchen garbage. Maximum ethanol concentration of 33.0 g/L was reported with the optimum conditions of time of 67.60 h, pH= 4.18 and T = 35 ◦C using glucoamylase in the saccharification step. Kitchen wastes were treated with a mixture of amylolytic and cellulolytic enzymes, the hydrolyzate was converted to bioethanol using commercial dry baker's yeast, and resulted in 23.3 g/L bioethanol production [35].

Bioethanol volumetric productivities ranged from 0.33 to 1.8 g/L/h and that of the present study is among the higher (Table 2). It is evident that the variability in FW content from region to region, the substrate concentration, the type and dosage of used enzymes and processes applied affect bioethanol production.

Enzyme cost still remains high, and this is identified as a major challenge in the deployment of lignocellulosic ethanol [9]. If the necessary enzymes could be efficiently produced on-site, the cost could be significantly reduced. A recent study has estimated that this cellulase cost can be reduced, from 0.78 to 0.58 \$/gallon, by shifting from the off-site to the on-site approach of cellulase production [38].

#### **4. Conclusions**

In the present study, the feasibility of producing bioethanol from FW applying the on-site produced multienzyme system was demonstrated. The mesophilic fungus *F. oxysporum* F3 grown under SSC in wheat bran as carbon source produced a mixture of hydrolytic enzymes capable of hydrolyzing the polysaccharides in FW. Mixed-microbial cultures of *F. oxysporum* with the yeast *S. cerevisiae* increased bioethanol volumetric productivity compared to mono-culture of the fungus. Bioethanol production increased by approximately 23% when the mixed microbial culture was supplemented with commercial glucoamylase. The results of the study demonstrated that non-commercial enzyme products obtained from fungi could be an efficient alternative to commercial preparations in technologies which use elevated substrate loadings or where an accurate loading is impossible due to practical limitations.

**Author Contributions:** Conceptualization, A.K., D.K. and D.M.; Methodology, A.K. and D.M.; Investigation, G.P., A.G., S.K.; Writing—Original Draft Preparation, D.M.; Writing—Review & Editing, all authors. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding

**Acknowledgments:** The authors would like to thank Gerasimos Lyberatos of Organic Chemical Technology Laboratory, School of Chemical Engineering, National Technical University of Athens for providing the dry FW. The authors would also like to thank Novozymes Corporation for generously providing the cellulose enzyme samples.

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

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

### **Biodegradation of Residues from the Palo Santo (***Bursera graveolens***) Essential Oil Extraction and Their Potential for Enzyme Production Using Native** *Xylaria* **Fungi from Southern Ecuador**

#### **Vinicio Carrión-Paladines 1,\*, Andreas Fries 2, Rosa Elena Caballero 3, Pablo Pérez Daniëls <sup>4</sup> and Roberto García-Ruiz <sup>5</sup>**


Received: 9 July 2019; Accepted: 17 August 2019; Published: 23 August 2019

**Abstract:** The degradation dynamics of lignin and cellulose were analyzed by means of a solid state biodegradation experiment, using residues from the essential oil extraction of the Palo Santo tree (*Bursera graveolens*). As such, two native *Xylaria* spp. and an exotic mushroom *Trametes versicolor* were incubated on the spent substrate (Residues of *B. Graveolens*, BGR's). The relatively high lignin and cellulose contents of the BGRs (9.1% and 19%, respectively) indicated the potential of this resource for the production of methane (biogas) and ethanol. However, the degradation of the lignin and cellulose content could be traced back to the relatively high activity of the enzymes laccase, cellulase, and xylanase, produced by the fungi. The results showed that laccase (30.0 U/L and 26.6 U/L), cellulase (27.3 U/L and 35.8 U/L) and xylanase (189.7U/L and 128.3 U/L) activities of *Xylaria feejeensis* and *Xylaria* cf. *microceras* were generally higher than *T*. *versicolor* (9.0 U/L, 29.5 U/L, 99.5 U/L respectively). Furthermore, the total carbon (TC: 47.3%), total nitrogen (TN: 1.5%), total phosphorus (TP: 0.2%) and total potassium (TK: 1.2%) dynamics were analyzed during the experiment and their importance for the degradation process highlighted. The results of this work might serve as guidance for future studies in dry forest areas, while furthering the understanding of the potential use of native fungi as ecologic lignocellulosic decomposers and for industrial proposes.

**Keywords:** Decomposition dynamics; *Bursera graveolens* waste; Xylariaceae; laccase; cellulase; xylanase

#### **1. Introduction**

The Palo Santo (*Bursera graveolens* [Kunth] Triana & Planchon; Burseraceae) is a native, deciduous, dioecious, non-timber tree species of dry forest areas and is distributed from western Mexico to northwestern Peru. In Ecuador, *B. graveolens* is native in the western coastal plains and also on the Galapagos Islands [1]. Particularly, in the dry forest areas of southern Ecuador, the Palo Santo tree is the most dominant native tree species [2].

The Palo Santo is considered a vital resource for the local communities of the dry forest, as different parts of the tree are used in traditional medicine, as well as for the extraction of essential oil [3]. In addition, the wood and stalks are applied to prevent mosquito bites and to treat aches and pains of differing origins, such as fibrosarcoma, atherosclerosis, and arthritis [4]. The production of essential oil from different parts of the Palo Santo tree has grown during the last few decades due to the increase of the global demand within the cosmetic and pharmacological industry. Generally, the woody material of the tree is used for essential oil extraction. However, currently it is also extracted from the fruits, as is practiced in Ecuador.

The essential oil extraction process generates abundant organic waste, which is commonly discarded directly into the natural ecosystems or burned. The organic waste can cause environmental problems such as air pollution and/or water and soil contamination because of the low natural degradation capacity of these residues. The essential oil content of the different parts of *B. graveolens* is relatively high. The wood in the form of kindling contains up to 5.2% of essential oil and the shavings up to 3.4% [3]. Using the fruits, the distillation process is less efficient because the fresh fruits only contain up to 3% of essential oil. Therefore, a considerable amount of organic waste is concurrently produced (at least 95% of the fresh biomass), for which reason waste management is necessary.

One possible reuse of this waste is the production of vermicomposts for farming purposes. However, the waste must be mixed with other organic residues, like kitchen waste or animal manures, to make the product suitable for agricultural proposes [5]. Another reuse potential is enzyme production for industrial purposes by means of biodegradation of the wastes with specific fungi. Indeed, microbial enzymes are known to play a crucial role as metabolic catalysts, which is why they are frequently used in various industries among other applications. The use of enzymes is extremely wide-spread, especially in industries [6] where over 500 products are made of enzymes and about 150 industrial processes need enzymes as microbial cell catalysts [7]. Therefore, the demand for industrial enzymes is continuously rising, which leads to a growing need for sustainable solutions. Microbes are one of the largest and most useful sources of enzymes in nature [8], but research is still needed due to the immense biodiversity, particularly in tropical countries. Analyzing the diversity of potential substrates and microorganisms in different ecosystems, such as the dry forest of Ecuador, may expand the knowledge about enzyme production sources.

Some of the enzymes which are frequently used for industrial application are cellullase, laccase, and xylanase. These enzymes detoxify industrial effluents from the paper, pulp, textile, and petrochemical industries. Furthermore, they are used in medical diagnostic tools, as catalysts in drug manufacturing, and as cleaning agents for water purification systems. Additionally, these enzymes are needed as ingredients in the cosmetic industry, as well as for the bioremediation of herbicides and pesticides [9].

Cellulase is especially important as a detergent additive because it catalyzes the breakdown of chemical bonds and is used in the textile industry for cleaning processes and to reduce waste production. Besides this, cellulase contributes to the sustainable production of second generation biofuels and other chemical derivatives [10]. Laccase is used for decoloration of textile effluents and textile bleaching [11], as well as to oxidize phenolic and non-phenolic lignin-related compounds and other environmental pollutants [12]. Moreover, laccases are also used in the formulation of biofuels, biosensors and the synthesis of new hybrid molecules [13]. Xylanase is mainly needed in the pulp, paper, food, and beverage industries, as well as for the saccharification of pre-treated lignocellulosic biomass for the production of second generation biofuels [14].

In the tropics, a prominent fungus genus used for biodegradation and enzyme production is *Xylaria* Hill ex Schrank (Xylariaceae, Xylariales, Sordariomycetes, Ascomycota), which comprises more than 300 species [15]. *Xylaria* spp. are considered to be significant producers of different ligninolytic enzymes, including laccase, xylanase and cellulase [16], besides their ability to degrade lignin [17], hemicellulose [18], and cellulose from trees and agricultural residues [19–22]. Within the ecosystems, this type of fungus participates in the carbon and nitrogen cycles [23] and plays an important role in the biodegradation processes of wood and leaf litter due to its complex and diverse enzymatic system.

In other biotechnological applications, the spent substrates are analyzed, especially the amounts of lignin, hemicellulose and cellulose. Lignin and hemicellulose are used in bioconversion processes for the production of bioethanol, biogas, and other biofuel products, while cellulose is used in the manufacture of cosmetics as well as in the development of new renewable energy sources [24]. Besides these parameters, the C/N ratio of the spent substrate is important, because a high C/N ratio may cause the immobilization of nutrients, which limits some biological processes such as respiration rates and the development of microbial biomass [25]. Furthermore, inorganic nutrients (phosphorus and potassium) of spent substrates are essential during the biodegradation process, because these elements can restrict the degradation of the spent substrate and the development of the microbial cells [26].

As far as we know, residues of *B. graveolens* obtained during the essential oil extraction process have not been evaluated for their potential as a spent substrate for enzyme production, despite other tropical resources such as rice cane (*Oryza sativa*), banana stems (*Musa paradisiaca*), wheat straw (*Triticum* spp.), sugar cane (*Saccharum o*ffi*cinarum*), and olives (*Olea europea*) [27] being utilized. Furthermore, in Ecuador, only two fungi, namely *Xylaria guianensis*(Mont.) Fr. and *Lentinula edodes*(Berk.) Pegler were studied for enzyme production, in which the former didn't show any enzymatic activity [28,29], which underlines the need for further biodegradation experiments with other *Xylaria* spp. to be carried out.

The overall objective of this study was to assess the possible biotechnological use of the *Xylaria* spp. using residues obtained from the Palo Santo essential oil extraction (BGR's). For that, the degradation dynamics of lignin and cellulose of the BGR's were analyzed by two native fungi (*Xylaria* spp.) of the dry forest areas of southern Ecuador, and compared to the behavior of *Trametes versicolor* (L.) Lloyd, due to its well-known degradation and enzyme production capacities. Furthermore, changes in the carbon, nitrogen, phosphorus, and potassium contents of the BGR's during the degradation process were evaluated.

#### **2. Materials and Methods**

#### *2.1. Identification of Fungi*

The fruiting bodies of *Xylaria* spp. were collected on stumps of dead wood of *B. graveolens* trees in the tropical dry forest of southern Ecuador (Figure 1). The samples were analyzed and deposited in the Herbarium of the Technical University of Loja (HUTPL), Fungarium section, using taxonomic criteria [30,31]. The fungus used as control was *Trametes versicolor* (isolated on MEA: Malt Extract Agar), which was donated by the Spanish National Research Council (Consejo Superior de Investigaciones Científicas, CSIC). This fungus was selected as a model due to its high enzymatic activity, especially in laccase, xylanase, and cellulose, during the degradation process [27].

#### *2.2. Isolation, DNA Extraction, PCR and Fungal Sequencing*

After the taxonomic identification of the *Xylaria* spp. the samples were disinfected with 5% sodium hypochlorite solution for three minutes. To flush out the solution and to clean the samples, distilled water was used for three minutes. Then, a second disinfection was applied with a 70% ethanol solution for one minute. Finally, the samples were cleaned with distilled water for one minute again [20].

The disinfected ascomes of each fungus were placed separately into Petri dishes on MEA (malt extract Agar), where the cell cultures were incubated at 25 ◦C until the mycelia growth was completed (after seven days). After this, the individual fungi were extracted and placed into Petri dishes on MEA again to guarantee the purity of the fungal isolation.

**Figure 1.** Digital Elevation Model (DEM) of continental Ecuador (left) and natural ecosystems of the province of Loja (right). The map was adapted from the Ecuadorian Ministry of the Environment [32]. The red symbols indicate the sampling point of the *Xylaria* spp.

To be sure that the individual fungal isolation only contained the required fungus species, a DNA extraction (PCR and sequencing test) was executed, following the protocol described by Iotti and Zambonelli [33] and Tamura et al. [34]. The phylogenetic location of the isolates was established by morphological observations and the DNA sequence of the ITS-5.8S region. As universal primers ITS1/LR5 or ITS1/NL4: ITS1 5 TCC GTA GGT GAA CCTGGG 3 [35], LR5 5 TCC TGAGGG AAA CTT CG 3 [36], NL4 5 GGT CCG TGT TTC AAG ACGG 3 [35] were used to amplify the ITS-5.8S region. The DNA sequences of the *Xylaria* spp. were classified by means of the registered species in the GenBank database using BLAST searching (https://www.ncbi.nlm.nih.gov; see also (supplementary material).

#### *2.3. Preparation and Chemical Analysis of the Spent Substrate*

The spent substrate used for the degradation experiment were the fruit residues from the essential oil extraction of *B. graveolens* (BGR's). The BGR's mainly consists of the skin and seeds of the fruits, which contain fiber, water, and fatty acids [5]. The residues were obtained from the UTPL Natural Products Institute (Loja, Ecuador), where the majority of the country's essential oil is produced.

The raw BGR's were dried at 60 ◦C for 48 h to avoid contamination and biodeterioration [37]. Then, the substrate was sieved through a 2-mm mesh and the pH determined. The contents of acid detergent fiber (ADF), acid detergent lignin (ADL) and cellulose in the BGR's were analyzed according to the method of Van Soest [38]. Briefly, the samples were digested with cetyl trimethyl ammonium bromide (CTAB) for 1 h at 150 ◦C and afterwards the residues washed with distilled water and then filtered (Whatman GF/C) using a vacuum pump. The retained residues in the filter were oven dried and weighed to calculate the ADF and the weight of the residues after the digestion were compared to the original weight and the difference determined. Then, a second digestion of the residues was applied using H2SO4 over 3 h at 25 ◦C. Afterwards, the residues were washed several times with distilled water to remove the excess acid, and then oven dried for one day at 105 ◦C and weighed again. The ADL was calculated by means of the percentage of the residues after the second digestion compared to the weight of the material after the first digestion. Finally, residues of the second digestion were incinerated in an oven for a period of 5 h at 500 ◦C to estimate the ash content. The percentage of cellulose was estimated by the difference between ADF and ADL [39].

The total carbon (TC) and total nitrogen (TN) contents of the BGR's were measured using an auto-analyzer CHNS (Elemental Thermo Finnigan Flash EA1112 CHNS-O). The total phosphorus (TP) and total potassium (TK) contents were determined after the acid digestion following the methodology proposed by Sommer and Nelson [40]. Briefly, 200 mg of the crushed BGR's samples were mixed with 5 mL of a perchloric acid solution (60%) and nitric acid (60%) in relation 3:5 (v/v), which was executed in a BD-40 digester block divided in two phases: 90 min at 130 ◦C and 75 min

at 204 ◦C. Finally, TP and TK were measured by means of acid extractions in the Agilent 750 Series ICP-MS kit.

#### *2.4. Solid State Fermentation*

For the solid state fermentation experiment, 20 g of the BGR's was added into 250 mL Erlenmeyer flasks and mixed with 80 mL of distilled water. To evaluate the biodegradation and enzymatic activities over time by means of the fermentation of the BGR's, different test series of the two selected native fungi (*Xylaria* spp.) and the control fungus *T. versicolor* were prepared (20 Erlenmeyer flasks for each fungus type).

Before inoculating the fungi in their specific flask, the mycelium grew for 15 days on MEA, and all flasks containing the BGR's were sterilized in a Gemmy SA-300VF autoclave. After cooling the flasks, the individual fungi were inoculated using 1 cm2 of each cell culture. Then, the flasks and their contents were incubated at a temperature of 25 ◦C, following the method proposed for *Xylaria* spp. by Liers et al. [41] and Rodrigues Negrão et al. [22].

The solid-state fermentation experiment lasted 60 days and samples were analyzed on day 7, 15, 30 and 60 after incubation. At each sampling day five recipients of each fungus, randomly chosen, were selected and examined, applying the quartering method [42]. Then, the five selected samples of each fungus were transferred to other recipients for lyophilization. Lyophilization was executed at −4 ◦C and 0.1 mm Hg using a LABCONCO equipment [43].

#### *2.5. Chemical Analysis and Quantification of BGR's Degradation*

After lyophilization, the BGR's were analyzed to quantify the degradation kinetics at each sampling day (7, 15, 30 and 60), applying the same methods related to the preparation and chemical analysis of the spent substrate. In the process, the degradation ratio of lignin and cellulose was calculated using the following equation [44]:

$$Ri\left(\%\right) = 100 \* \left(m\_{\odot} - m\_{\mathrm{i}}\right) / m\_{\odot} \tag{1}$$

where Ri is the percentage of degradation for the sampling of the week; mo is the initial content of lignin and mi is the content of lignin sampling for each week of degradation.

Furthermore, the reduction in total carbon (TC), total nitrogen (TN), total phosphorus (TP), and total potassium (TK) of the BGR's was determined.

#### *2.6. Enzymatic Assay*

The lyophilized samples were mixed with distilled-deionized water, filtered and centrifuged for 30 min at 7 ◦C and 8500 rpm, before the enzymatic analysis [45]. The laccase activity (E.C.1.10.3.2; p-diphenol: dioxygen oxidoreductase) was determined by the oxidation of syringaldazine (4-hidroxi-3.5-dimetoxibenzaldehidacine) in a 0.22 mM methanolic solution [45,46]. The reaction was carried out at pH 6.5 to measure the corresponding quinone at a wave length of 530 nm (ε= 65,000 M−<sup>1</sup> cm<sup>−</sup>1). The enzymatic activity of laccase was defined by the amount of enzyme that catalyzed the transformation of 1 μmol substrate per minute.

The xylanase activity was determined by measuring the reduction of glucose. Therefore, a 1% xylene solution in an acetate buffer (50 mM) was used [47,48]. The reaction was executed at a pH of 5.0 and measured at a wave length of 575 nm. To quantify the enzymatic activity, the amount of enzyme that reduced 1 μmol of xylose per minute was determined.

The cellulase activity (E.C.3.2.1.4; β-1,4-endoglucanase) was determined by measuring the formation of reduced sugar (glucose) [48,49] using carboxymethylcellulose (1%) in an acetate buffer (50 mM) as substrate. The reaction was executed at a pH of 5.0 and the product measured at a wave length 550 nm. The enzymatic activity was defined by the amount of enzyme that produced 1 μmol of glucose per minute.

In this study, the enzyme activity is expressed in terms of the volumetric activity in Units per Liter (U/L). To obtain the activity value for each enzyme (laccase, xylanase and cellulase) and for each fungus on the respective sampling day, the arithmetic mean of the five analyzed samples was calculated.

#### *2.7. Statistical Analysis*

The degradation capacity of the studied fungi inoculated on the BGR's was evaluated by a one-way ANOVA using the SPSS Statistical Software package (v.15.0; SPSS Inc., Chicago, USA). The correlations between the measured variables were determined by the Pearson correlation coefficient; significance was accepted at *p*-value < 0.05 in all cases. The enzymatic activity of the fungi during the solid-state fermentation experiment was assessed through a repeated measures ANOVA. The differences between the means were evaluated through the multiple range Tukey Test (HSD) and accepted at a significance level of 5% (*p*-value < 0.05).

#### **3. Results and Discussion**

#### *3.1. Identification of Fungi*

The sequences of the utilized *Xylaria* sp. were compared to the species registered in the international GenBank (https://www.ncbi.nlm.nih.gov) to realize the phylogenetic analysis (Table 1).



The results are shown in Figure 2, where two clades are presented, which indicate an accordance of over 70%, when analyzing the ITS-5.8S regions. By means of the GenBank sequences, one fungus used in this study was identified as *X. feejeensis*, whereas the other native fungus could not be classified precisely. The closest sequence was related to *X. microceras* (accession code GU300086), because the morphology was similar but the genetics different. Therefore, this fungus is probably a new species, which should be analyzed more systematically. Consequently, this fungus is named *X*. cf. *microceras* for the present study.

**Figure 2.** Phylogenetic location of *X. feejeensis* (green) and *X*. cf. *microceras* (red) based on our ITS-5.8S sequences (in bold) and the most related sequences from the GenBank. Only values ≥ 70% are shown on the nodes. The sequence from *Camarops ustulinoides* with accession number AY908991 was used as out group.

#### *3.2. Spent Substrate Characterization*

Table 2 shows the chemical characterization of the BGR's before the biodegradation by the three fungus species. The spent substrate had an average lignin and cellulose content of 9.1% and 19.5%, respectively. The mean TC and TN were 47.3% and 1.5 %, which resulted in an average C/N ratio of 30.4. The TP and TK contents were 0.2% and 1.2% respectively. Sulfur (S) was absent in the BGR's, which indicates that the substrate is suitable for biodegradation because no SO2 (reactive gas) can be emitted [61]. The average pH value of the spent substrate was 7.0.

**Table 2.** Biochemical properties of BGR's in natural form, prior to inoculation with the strains (*X. feejeensis*, *X*. cf. *microceras* and *T. versicolor*) and their standard deviation based on four replicates.


The lignin content is important for methane (biogas) production [62], where garden waste (lignin: 10.5%), rice straw (10.8%), shells of *Durio zibethinus* (11.4%) and vinegar residues (lignin: 12.4%) are generally used [63]. The lignin content of the BGR's (9.1%, Table 2) is slightly lower than these substrates, but higher compared to other wastes, which are also used for methane production, such as the leaves and seeds of *Chenopodium album* (lignin: 7.7%), its fruit and vegetable (7.9%), and seeds of *Durio zibethinus* (8.8%) [63,64], which makes the BGR's suitable for methane (biogas) production.

The cellulose content is important for ethanol production [65], where generally banana peel and skin (13.2% and 9.2%, respectively) as well as rice bran are used. The BGR's (19.5%, Table 2) had notably higher cellulose content, which indicates that ethanol can be potentially produced from the BGR's. However, for the manufacture of cosmetics, higher cellulose contents are required [24], which can be found in fiber sorghum (*Sorghum* sp. 41.8%) and in rice husk (*Oryza sativa*, 33.0%) [66].

The TC (47.3%) and TN (1.5%) content establish the C/N ratio, which is important for biodegradation experiments, because substrates with high C/N ratios usually produce the immobilization of nutrients during the process [25]. The optimal range of the C/N ratio lies between 20 and 30, as Montingelli et al. [67] stated. For the BGR's, a nearly optimal C/N ratio (30.4) was obtained, although the TC and TN contents were lower compared to other wastes used for oil extraction, like residues of olive fruits (*Olea europaea*; 58.5% TC, 1.8 % TN, C/N ratio: 31) or the seeds of the litchi (*Litchi chinensis* 56.1% TC, 1.1% TN; C/N ratio: 51) [68]. However, the C/N ratio of the BGR's is much more appropriate than these spent substrates, as well as other spent substrates (e.g., rice straw, *Oryza sativa*; TC: 57.7, TN 0.5%, C/N ratio: 115.0) used in biodegradation experiments [69], which underlines the utility of the BGR's. Furthermore, Motingelli et al. [67] found, that the maximum methane yield is produced when the C/N ratio is around 30, which additionally affirms the potential use of the BGR's for biogas production.

The contents of TP and TK (Table 2) can restrict the degradation of the spent substrate, because these nutrients have an influence on the physiology and the growth of the fungi [26,70,71]. The BGR's showed an average TP content of 0.2%, which is similar to olive residues and rice straw [68]. However, according to El-Haddad et al. [69], the optimal TP values range between 0.7% and 1.1%, which indicates a deficiency of this element in the BGR's. The optimal range of TK lies between 1% and 3% [59], which indicates that the TK content of the BGR's (1.2%, Table 2) is at the lower end of the optimal range, but still adequate for biodegradation experiments. Finally, the neutral pH of the BGR's favors

the development of the fungi, because pH values around 7 increase their growth, especially for *Xylaria* spp. [72].

#### *3.3. E*ffi*ciency of Degradation of the Three Fungi in the BGR's*

Figure 3 shows the degradation of lignin (a) and cellulose (b) of the BGR's in the presence of the fungi. The degradation increased for all fungi during the incubation period, reducing the original content of lignin between 15% and 34% and of cellulose between 28% and 56%.

**Figure 3.** Lignin (**a**) and cellulose (**b**) degradation of BGR's by native fungi of *Xylaria* spp. and *T. versicolor* in solid state fermentation. The bars show the standard error of the mean measured activity of the four samples analyzed on each sampling day. Different letters (**a**,**b**) indicate significant difference (*p* ≤ 0.05%, HSD Tukey).

However, significant differences were found between the fungus types, at which *T. versicolor* was the most efficient fungus for lignin degradation (33.8%, Figure 3a), but the poorest for cellulose degradation (28.3%, Figure 3b). The two native fungi isolated from the Palo Santo wood, *X*. cf. *microceras* and *X. feejeensis*, were less effective in the degradation of lignin (Figure 3a), but demonstrated their capacity, which is also confirmed by Osono and Takeda [17] and Koide et al. [18]. These studies illustrated that *Xylaria* spp. have the potential to degrade lignin as well as holocellulose, because they produce selective delignification and therefore have a good ligninolytic capacity. Furthermore, the high degradation capacity of *Xylaria* spp. is reported by Pointing et al. [73], Chaparro et al. [20] and Rodrigues Negrão et al. [22], who consider *Xylaria* spp. within the group of white-rot fungi, which degrade lignin effectively.

In contrast, *X*. cf. *microceras* was the most effective fungus for cellulose degradation, reducing the original cellulose content of the BGR's by about 56.6% during the incubation period (Figure 3b), followed by *X. feejeensis* (42.3%) and *T. versicolor* (28.3). The relatively high degradation rate of the two *Xylaria* spp. are consistent with findings of previous studies [73], where it was stated that all Xylariaceae taxa have high capacities to hydrolyze lignocellulosic resources.

#### *3.4. Mineralization of TC, TN and Evaluation of C*/*N Ratio*

During degradation of the organic matter, organic carbon is needed by the fungi as an energy and biomass source, converting it partially into CO2 and under certain circumstances also into methane (CH4). The organic nitrogen is mainly transformed into available N (nitrate and ammonium) by the fungi during the decomposition, and afterwards partially assimilated to build new biomass [74]. Therefore, after 60 days of incubation, the initial TC content of the BGR's was strongly reduced, whereas the original TN content of the BGR's (1.5%, Table 1) showed only small variations during the whole solid state experiment (Table 3).

The TC mineralization was highest during the first seven days of incubation for all three fungi, when almost 50% of the initial TC content was degraded. However, TC content was most effectively degraded by *T*. *versicolor* (final content: 16.1%) followed by *X. feejeensis* (final content: 17.0%) and *X*. cf. *microceras* (final content: 19.1%), which can be traced back to the high secretion of cellulolytic enzymes of all fungi during the experiment, because the spent substrate is relatively rich in organic carbon.

The TC mineralization was positively correlated to the degradation of lignin (0.58, *p* < 0.01) and higher to the degradation of cellulose (0.82, *p* < 0.01; Table 4). In addition, the correlation between the reduction of lignin or cellulose and the decrease of the TC content in the BGR's was significant, which was also found in other investigations [75].

TN content remained stable (~1.4%) during the incubation period for the three fungi (Table 3). The small variations in TN content during the mineralization process can be explained by the moderate TN content of the BGR's. The fungi mainly used the carbon content as an energy source and to build biomass. This finding is confirmed by Rigby et al. [76], who stated that the TN content of the spent substrate is only degraded if it is needed for the metabolic requirements of the fungi (microbial cells). In this case, the TC content of the BGR's is sufficient for the metabolic requirement and development of the fungi, while the TN content stayed more or less stable during the whole incubation period. Besides this, the moderate TN content of the BGR's (initially; 1.5%) makes the substrate suitable as organic fertilizer, because, according to the European eco-label, an organic fertilizer should not exceed 3% of TN.

As expected, the C/N ratio decreased notably during the inoculation period (Table 3) because of the degradation of the TC content of the BGR's [77]. The C/N ratio was positively correlated with the degradation of lignin (*r* = 0.56, *p*-value < 0.01) and cellulose (*r* = 0.84, *p*-value < 0.01), which underline the relation between TC degradation and lignin/cellulose reduction (Table 4).


**Table 3.** Changes in biochemical constituents of Palo Santo waste (*B. graveolens*) during 60 days of solid-state fermentation, with native fungi *Xylaria* spp. and


**Table 4.** Pearson correlation coefficient among BGR's properties. Significant correlation is shown at *p* < 0.05 (\*) and *p* < 0.01 (\*\*).

#### *3.5. Mineralization of Phosphorus and Potassium*

The TP content of the three test series slightly increased from 0.2% to 0.3% during the solid state experiment (Table 3). This is caused by the transformation of the organic phosphorus (Po) into its inorganic form (Pi) [78]. The soluble Pi is afterwards incorporated into the OM of the fungi to build up biomass. These results coincide with those reported by Kuehn and Suberkropp [79], who also observed an increase of TP when inoculating different fungi in decaying litter of the *Juncus e*ff*usus*. The optimal range of TP content in organic fertilizer lies between 0.15% and 1.5%, which makes the BGR's an adequate resource for soil improvers.

The initial TK content of the BGR's (1.2%, Table 2) was reduced to 50% after the first 15 days of incubation by all three fungi, and afterwards the values increased again, reaching the final value of approximately 0.8% on the last sampling day (Table 3). Potassium is a very mobile and unstable nutrient, which is needed by the fungi particularly at the beginning of the degradation process [80]. However, the increase of TK at the end of the incubation period was probably due to the mineralization of the OM and the production of CO2 by the fungi. The typical range of potassium in organic fertilizers is 0.4-1.6%, which indicates that the BGR's do not present a deficiency in this nutrient. Furthermore, the normal TK content of the BGR's induces a good C/N balance, because TK plays an important role in carbon (C) and nitrogen metabolism (N) [81].

#### *3.6. Enzyme Activities*

The degradation of lignin and cellulose is a consequence of the enzymatic activity of the fungi, which is mainly caused by the production of the enzymes laccase, xylanase and cellulase [44,82]. Figure 4 shows the variation in laccase (a), xylanase (b) and cellulase (c) activities for the individual fungus types during the incubation period.

#### 3.6.1. Laccase Activity

Laccase activity was detected for all three fungi on the first sampling day (Figure 4a), which was probably due to the absence of sulfur (S) and the moderate contents of TC and TN of the BGR's (Table 2), facilitating the immediate decomposition of lignin [83]. The temporal variability of laccase activity of *X. feejeensis* and *T. versicolor* were similar, increasing notably during the first seven days of incubation (33.5 U/L and 32.4 U/L; respectively), and afterwards remaining more or less stable until the end of the incubation period (final values: of 33.7 U/L, *X. feejeensis*; 31.5 U/L, *T. versicolor*). *X*. cf. *microseras* showed a different behavior, especially during the first seven days of incubation, when the lowest laccase activity of all three fungi was detected (15.6 U/L), and during the end of the incubation period, when the laccase activity of *X*. cf. *microceras* increased notably, reaching 41.3 U/L, which was the highest value of all test series during the complete observation period.

**Figure 4.** Laccase (**a**), xylanase (**b**) and cellulase (**c**) activities in BGR's with native species of *Xylaria* spp. and *T. versicolor* incubated at 25 ◦C. The bars show the standard error of the mean measured activity of the four samples analyzed on each sampling day. Different letters indicate significant difference (*p* ≤ 0.05%, HSD Tukey).

However, the variability in the production of the laccase enzyme was expected, because, as Dong et al. [44] showed, laccase acts synergistically with other lignin-degrading enzymes such as polyphenol oxidase (PPO) and manganese peroxidase (MnP).

The observed laccase activity of the three studied fungi seems to be low in comparison to other studies [41], but these investigations applied liquid fermentation (liquid spent substrates) and used 2.5-xylidine or veratryl alcohol as an enzymatic inductor, which increases the laccase activity notably [84]. The solid-state experiment of this study did not utilize any enzyme inductor, which explains the relatively low laccase activity of *T. versicolor* and the *Xylaria* spp. However, investigations using solid substrates obtained similar activity values as observed here [20,85]. Generally, all *Xylaria* spp. have an unusually high ability to degrade lignin compared to other Ascomycota, as Liers et al. [23] and Rodrígues Negrão et al. [22] stated, because *Xylaria* spp. mineralize lignin almost as efficiently as the aggressive fungi of white-rot (e.g., *T. versicolor*).

Besides the solid fermentation applied, the moderate laccase enzyme activity might be due to the low lignin content of the BGR's (9.1%; Table 2), which cause a reduction in laccase production. As Coronel et al. [85] indicated, the enzymatic activity is highly influenced by the chemical composition of the substrate where the fungus was incubated. Therefore, average laccase activity of all fungi was more or less similar (*X. feejeensis*: 30.0 U/L; *T. versicolor*: 29.5 U/L; *X*. cf. *microceras*: 26.6 U/L). However, the statistical analysis showed that a significant difference between *X*. cf. *microceras* and the other two fungi (*X*. *feejeensis* and *T. versicolor*) respective to laccase activity exists (*p*-value < 0.05; Figure 4a, different letters), whereas no difference was found between *X. feejeensis* and *T. versicolor* (Figure 4a, similar letters).

#### 3.6.2. Xylanase Activity

Xylanase also act synergistically with other hydrolytic enzymes, such as esterase, to modify the structural configuration of lignocelluloses [44], which is necessary to make the cellulose accessible for degradation. For example, esterase, in combination with xylanase, cleaves covalent bonds between polysaccharides or hemicelluloses, and therefore plays a key role in the degradation of the hemicellulose matrix [82]. The preliminary degradation of the lignin-hemicellulose matrix is proven by the present study because xylanase activity was detected in all three fungi species on the first sampling day, reaching values between 100.7 U/L and 196.0 U/L (Figure 4b).

During the complete observation period, the xylanase activity of *X. feejeensis* was always highest compared to the other two fungi, except on day 15 when the activity of *X. feejeensis* decreased and similar values for both *Xylaria* spp. were measured (140.8 U/L and 141.6 U/L). Highest activity of *X. feejeensis* was observed on day 30 with 284.3 U/L, but afterwards activity decreased. The xylanase activity of *X*. cf. *microceras* and *T. versicolor* started with lower values (150.3 U/L and 100.7 U/L respectively), but their activity curves showed similar behaviors (Figure 4b), decreasing slightly between day 7 and day 15 after incubation and then remained almost stable during the rest of the observation period, reaching final values of 117.9 U/L (*X.* cf. *microceras*) and 99.0 U/L (*T. versicolor*). Statistically significant differences were found between *X. feejeensis* and the other two fungi species, but differences between *X*. cf. *microceras* and *T. versicolor* were not significant (Figure 4b).

In general, *X. feejeensis* and *X*. cf. *microceras* showed higher xylanase activity compared to *T. versicolor*, but also to other *Xylaria* spp. which were inoculated on solid-state experiments [23]. All *Xylaria* spp. can degrade hemicellulose effectively [18,19], but the two *Xylaria* spp. studied here showed even higher activity values than those reported in the literature and therefore might be suitable for commercial xylanase production required for industrial processes in the paper, food and wine industry. *T. versicolor* also showed high xylanase activity, although the xylanase production of this fungus species is typically low and an inducible mechanism (enzyme inductor) is needed to increase the activity. Irbe et al. [86] used glycerol alcohol as an inductor and observed a notable increase in xylanase activity and enzyme production. However, in this study no inductor was applied, but xylanase activity was high for all three fungi, which is probably due to the xylan content of the BGR's. Xylan is the

second most abundant hemicellulosic polysaccharide in nature and present in the cell walls of the plants [43]. The residues from the Palo Santo essential oil extraction consist mainly in parts of the fresh fruits [5] and therefore the xylan content was not reduced by other degradation processes, which might explain the high xylanase activity of all three fungi.

#### 3.6.3. Cellulase Activity

As expected, cellulase activity was delayed (Figure 4c), because the lignin-hemicellulose matrix had to be degraded first to make the cellulose accessible [44], which indicates that cellulase is an inducible enzyme [87]. As Arantes and Sadler [88] stated, the cellulose regions are tightly packed with lignin and hemicellulose, which is the major contributing factor to cellulose resistance to degradation. Therefore, amorphogenesis, a process consisting of the degradation of lignin and hemicellulose, is needed to liberate the cellulose, which afterwards can be degraded by cellulase.

Cellulase activity of the two *Xylaria* spp. was detected on the first sampling day, in contrast to *T. versicolor*, where significant activity was not measured until day 30 (Figure 4c). The *X. feejeensis* fungus reached its maximum activity (49.2 U/L) on day 30, whereas *X*. cf. *microceras* peaked on day 15 (59.5 U/L), which was concurrently the highest value of all three fungi species during the entire observation period. *T. versicolor* did not show any cellulase activity until day 15 of incubation and afterwards displayed an almost linear increase to its maximum and final value of 28.7 U/L on the last sampling day.

In general, the cellulase enzyme system consists of three types of enzymes; endo-1,4-β-glucanase (cellulase), cellobiohydrolase or exoglucanases (avicelase), and β-glucosidase (cellobiase), which act in conjunction to degrade the cellulose content of the spent substrate [44]. As Du et al. [89] showed, *T. versicolor* first degrades cellulose through the enzyme cellobiase (β-glucosidase), and later in combination which the enzyme cellulase, for which reason a delay in cellulase activity of *T. versicolor* was observed. In contrast, the *Xylaria* spp. produced the enzyme cellulase immediately after incubation to degrade the cellulose content.

As is also shown in Figure 4c, maximum values of cellulase activity of the individual fungi were not reached simultaneously, due to the different metabolic requirements of each fungus. The difference may be due to the fungus type used (Basidiomycota and Ascomycota), because *T. versicolor* is a fungus of white-rot, which is specifically indicated in the degradation of lignin. Furthermore, the lower cellulase activity of *T. versicolor* might be a consequence of the substrate used (solid), because *T. versicolor* can produce up to 100 U/L of cellulase when inoculated in liquid mediums [90].

The differences in the cellulase activity are also depicted by the repeated measures ANOVA, in which no significant differences between *X*. *feejeensis* and *X*. cf. *microceras* were found, but differences between the *Xylaria* spp. and *T. versicolor* were significant (*p*-value < 0.05). In general, average cellulase production of the *Xylaria* spp. was notably higher (27.3 U/L to 35.8 U/L) compared to *T. versicolor* (9.0 U/L), especially during the first 30 days of incubation.

The high cellulase activity of the two *Xylaria* spp. is consistent with results from previous investigations comparing the enzymatic production of different fungi species [19,73]. These studies demonstrated that *Xylaria* spp. are potential producers of cellulolytic enzymes, due to their high cellulolytic activity, and therefore can be used in biotechnological applications as well as for industrial purposes. This is confirmed by Gutiérrez-Soto et al. [16], who measured the cellulase activity of *Xylaria* spp. up to 199 U/L. However, these studies incubated the fungi on liquid mediums; in solid spent substrates, the activity is generally lower [23]. The variation in cellulase enzyme activity of *Xylaria* spp. apparently depends on two factors: first, the species [73] and second, the type of substrate [19].

#### **4. Conclusions**

The content of lignin and cellulose in the BGR's makes the substrate suitable for biotechnological applications, especially for the production of methane and ethanol. Furthermore, the contents of the macro nutrients were within the optimal range, and therefore the BGR's can be applied to soils as fertilizers.

The BGR's were also suitable for the production of enzymes for industrial purposes by means of fungal degradation. The native *Xylaria* spp. showed generally higher enzymatic activity than the control fungus and were especially practical for the production of xylanase and cellulase.

**Supplementary Materials:** Supplementary materials can be found at http://www.mdpi.com/2311-5637/5/3/76/s1.

**Author Contributions:** Formal analysis, V.C.-P. and A.F.; Investigation, V.C.-P.; Methodology, V.C.-P., R.E.C. and R.G.-R.; Supervision, R.G.-R.; Validation, P.P.D.; Writing—original draft, V.C.-P., R.E.C. and R.G.-R.; Writing—review & editing, V.C.-P., A.F., P.P.D. and R.G.-R.

**Funding:**This researchwas funded by the Secretary of Science and Technology of Ecuador (SENESCYT—CEREPS—2007) and the Universidad Técnica Particular de Loja (UTPL).

**Acknowledgments:** We would like to thank the Secretary of Science and Technology of Ecuador (SENESCYT), the Ecuadorian Ministry of the Environment (MAE, research permits 015-IC-UPN-DRLEOZCH-MA and 027-2013-DPL-MA) and the Department of Biological Sciences of the Universidad Técnica Particular de Loja for their support. Special gratitude to the Spanish National Research Council (Consejo Superior de Investigaciones Científicas, CSIC) for their generous contribution. Finally, we would like to thank Gregory Gedeon for text revision.

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

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Spent Yeast from Brewing Processes: A Biodiverse Starting Material for Yeast Extract Production**

#### **Friedrich Felix Jacob 1,\*, Lisa Striegel 2, Michael Rychlik 2, Mathias Hutzler <sup>3</sup> and Frank-Jürgen Methner <sup>1</sup>**


Received: 5 June 2019; Accepted: 20 June 2019; Published: 24 June 2019

**Abstract:** Spent yeast from beer manufacturing is a cost-effective and nutrient-rich starting material for the production of yeast extracts. In this study, it is shown how physiologically important ingredients in a yeast extract are influenced by the composition of the spent yeast from the brewing process. In pilot fermentations, the time of cropping (primary fermentation, lagering) of the spent yeast and the original gravity (12 ◦P, 16 ◦P, 20 ◦P) of the fermentation medium was varied, and four alternative non-*Saccharomyces* yeast strains were compared with two commercial *Saccharomyces* yeast strains. In addition, spent yeast was contaminated with the beer spoiler *Lactobacillus brevis*. The general nutrient composition (total protein, fat, ash) was investigated as well as the proteinogenic amino acid spectrum, the various folate vitamers (5-CH3-H4folate, 5-CHO-H4folate, 10-CHO-PteGlu, H4folate, PteGlu) and the biological activity (reduction, antioxidative potential) of a mechanically (ultrasonic sonotrode) and an autolytically produced yeast extract. All the investigated ingredients from the yeast extract were influenced by the composition of the spent yeast from the brewing process. The biodiversity of the spent yeast from the brewing process therefore directly affects the content of physiologically valuable ingredients of a yeast extract and should be taken into consideration in industrial manufacturing processes.

**Keywords:** yeast extract; brewer's spent yeast; autolysis; ultrasonic sonotrode; *Saccharomyces cerevisiae*/*pastorianus*; non-*Saccharomyces* yeast; proteinogenic amino acids; folate vitamers; biological activity

#### **1. Introduction**

Beer production generates large quantities of spent yeast during the fermentation and lagering process. Following primary fermentation, this equates to about 0.7–1.1 kg compressed yeast per hectoliterfinished beer [1]. According to the current state of brewing technology, spent yeast after primary fermentation is only used in small quantities to pitch the next batch [2]. The major share is obtained from a propagation plant, which provides highly viable and vital yeast that ferments vigorously [2]. At the end of the cold lagering process, yeast referred to as "lagering cellar yeast" is generated (0.5–0.9 kg compressed yeast per hl finished beer), together with precipitated turbidity particles and "barm beer" [1].

The spent yeast from the brewing process is suitable for use as an efficient starting material to produce yeast extract [1,3]. Yeast extract is generally defined as the soluble content of a yeast cell that remains once the cell wall has been destroyed and removed [4–6]. The variety of different physiologically valuable substances in yeast cells offers the possibility of using them as yeast extract in different areas of the food industry [3,7]. As "yeast food", these extracts can therefore increase the free α-amino nitrogen (FAN) when fermenting beer worts with a high content of unmalted grains [8] or a high extract content (high-gravity worts) [9,10], and consequently improve the yeast's nutrient supply and fermentation performance [11,12]. Free proteinogenic amino acids supply the majority of the FAN [12]. The quantity and composition of the relevant amino acids are ultimately critical to performance during fermentation [13] and also impact the beer's aroma profile [14]. From a nutritional standpoint, yeast extracts from spent yeast supply a high concentration of essential and semi-essential amino acids for human beings [5,7]. Yeast extracts are also a good source of B vitamins [6,15]. Among these, the various naturally occurring folate vitamers play an essential role in the human diet, with the biologically active form 5-methyltetrahydrofolate (5-CH3-H4folate) fulfilling key metabolic tasks in human cells [16]. The bioactivity of yeast extracts, which is demonstrated in the form of reduction and anti-oxidative potential, also makes these extracts particularly interesting for the food industry [6,7,15,17].

The majority of globally produced beer is manufactured by fermenting high-gravity worts [18]. The extract content of the wort is increased by adding sugar syrup, which modifies the nutrient balance of the wort with respect to all physiologically active components [18]. The altered yeast metabolism during high-gravity fermentation not only changes the quality of the finished beer but also the material composition of the yeast [10,18]. In addition, the biodiversity of the generated spent yeast in breweries is increased through the use of various alternative non-*Saccharomyces* strains as pure starter cultures for beer production [19,20]. Improper storage or handling of the spent yeast can result in contamination with various microorganisms, which can impact the subsequent yeast extract production process [21]. The composition of ingredients in commercially available yeast extracts varies greatly [17]. One reason is the influence of different yeast extract manufacturing methods, which we have evaluated in previous studies [5,6]. Another reason lies in the diversity of yeast starting material [22]. To the best of our knowledge, no research has been undertaken until now about the influence of the biodiversity of spent yeast from the brewing process on the composition of ingredients in yeast extracts.

In this work, it was shown for the first time how the composition of various physiologically valuable substance groups of a yeast extract depends on the biodiversity of the spent yeast from beer production. Therefore, beer was produced on a pilot scale using 12 ◦P wort and different yeast strains (*S. cerevisiae* TUM 68, *S. pastorianus* TUM 34/70, *Saccharomycodes ludwigii* TUM SL 17, *Saccharomycopsis fibuligera* TUM 525, *Brettanomyces bruxellensis* TUM Bret 1 and *Torulaspora delbrueckii* T 90). Furthermore, different wort gravities (12 ◦P, 16 ◦P, 20 ◦P) were fermented with *S. cerevisiae* TUM 68 to investigate the influence of high-gravity brewing on a commercial yeast strain. The spent yeast generated after primary fermentation and lagering was then processed into yeast extract using a mechanical (ultrasonic sonotrode) and autolytic cell disruption method. All yeast extracts were investigated to determine their general composition (protein, fat and ash). The physiologically valuable protein content was analyzed in detail with regard to different free and protein-bound amino acids. The effects on the amino acid spectrum of the yeast extract through contamination of the spent yeast by the obligate beer spoiler *Lactobacillus brevis* were also observed. Additionally, we characterized the biological activity of the yeast extract based on its reduction and anti-oxidative potential. We also showed how the total folate content was allocated between the different folate vitamers (5-CH3-H4folate, 5-CHO-H4folate, 10-CHO-PteGlu, H4folate, PteGlu) in the fermentation medium or in the relevant spent yeast and then how it could be transferred to the yeast extract. These results should increase the knowledge on fluctuating nutritional composition of yeast extracts. Furthermore, the most appropriate brewer's spent yeast could be selected to produce a yeast extract with the desired nutritional composition.

#### **2. Materials and Methods**

#### *2.1. Yeast Propagation and Fermentation*

A sterilized, hopped and standardized all malt wort concentrate (N53940; Döhler GmbH, Darmstadt, Germany) was used to produce the standardized propagation and fermentation wort. The all malt wort concentrate was diluted to an original gravity of 12 ◦P. To adjust the wort to the higher gravities of 16 ◦P and 18 ◦P, respectively, D-(+)-maltose monohydrate (Merck, Darmstadt, Germany) was added to 12 ◦P wort as an adjunct. Before use, worts were heat treated at 100 ◦C for 10 minutes for sterilization. The original gravity of the standardized fermentation wort corresponded to the original gravity of the standardized propagation wort. For the precise composition of the standardized all malt wort refer to Table 1.


**Table 1.** Wort composition.

The propagation procedure described below was used for all the yeast strains in the study. *Saccharomyces cerevisiae* TUM 68 (hereinafter Scer), *Saccharomyces pastorianus* TUM 34/70 (hereinafter Spas), *Saccharomycodes ludwigii* TUM SL17 (hereinafter Slud), *Saccharomycopsis fibuligera* TUM 525 (hereinafter Sfib), *Brettanomyces bruxellensis* TUM Bret1 (hereinafter Bbru) and *Torulaspora delbrueckii* TUM T90 (hereinafter Tdel) were sourced from the Yeast Center at the Weihenstephan Research Center for Brewing and Food Quality (RCW) of the Technical University of Munich (TUM) on agar slant. An inoculation loop of a pure agar slant colony was transferred to 40 mL standardized wort and incubated for 48 h at 20 ◦C on an orbital shaker (80 rpm). The 40 mL transferred into 400 mL standard wort and incubated again for 48 h at 20 ◦C on an orbital shaker (80 rpm). This process was repeated from 400 mL to 4 L standardized wort followed by an incubation at 48 h at 20 ◦C on an orbital shaker (80 rpm).

The fermentation procedure described below was used for all the yeast strains and standardized worts in the study. Standardized laboratory-scale brewing trials were performed using stainless steel vessels of 10 cm diameter × 33 cm height (2.5 L) with 20% headspace and clamped down lids according to Meier-Dörnberg et al. [23]. The propagated yeast was pitched in standardized and aerated (10 mg O2/L) wort in a Cornelius container (20 L) with a living cell count of 15 million CFU/mL. Each batch was then divided into three fermentation vessels. Fermentation took place at 18 ◦C and was unpressurized until final attenuation. The viscous primary fermentation yeast was cropped at the bottom of the vessel immediately after the final attenuation. The fermented supernatant was then stored in a carbonated and pressurized state (0.6 bar) for 14 days at 2 ◦C. Lagering cellar yeast was cropped from the bottom at the end of the cold lagering process again. The fermented supernatant and the spent yeast of each batch was immediately used for analysis and yeast extract production.

#### *2.2. Yeast Pre-Treatment*

After the yeast was cropped, it was immediately subjected to three washing processes to remove residual wort components. Each washing process was performed as follows: The viscous cropped spent yeast was diluted with distilled water to 10% dry matter, passed through a yeast sieve (mesh size 0.5 mm), centrifuged (1000 g, 5 min, 18 ◦C, 500 mL centrifuge tube) and the supernatant was discarded. The sedimented yeast in the centrifuge tube was then resuspended with distilled water for 5 minutes to 10% dry matter and the washing procedure was started afresh. The washed yeast was subsequently collected and diluted to 7% dry matter in distilled water before being fed into the disruption process.

#### *2.3. Yeast Quality Control*

The propagation yeast, the spent yeast, the washed spent yeast before disruption and the macerated yeast suspension were only used after passing quality control. Tests for quality control were already described in detail in our previous work [5,6] and had to give negative results for foreign yeasts and microorganisms.

#### *2.4. Yeast Cell Disruption Methods*

#### 2.4.1. Ultrasonic Sonotrode

Cell disruption using cavitation was carried out using the ultrasonic homogenizer SONOPLUS HD 3400 (Bandelin). The sonotrode diameter was 25 mm with an operating frequency of 20 kHz. In a stainless steel vessel (400 mL) the process suspension (200 mL) was subjected for 30 minutes to a constant ultrasonic output of 400 W without pulsation. The resulting process heat was removed by means of a glycol-cooling bath to maintain a constant temperature of 7 ◦C. This disruption process was adopted according to Jacob et al. [5,6].

#### 2.4.2. Standard Autolysis

To autolyze the yeast cells, 200 mL of the yeast suspension was heated in a reaction vessel (400 mL) for 24 h at 50 ◦C with constant stirring (100 rpm). Sodium chloride (0.086 mol/L) and ethyl acetate (0.051 mol/L) were added at the start of the process. This disruption process was adopted according to Jacob et al. [5,6].

#### 2.4.3. Autolysis with Contamination of *Lactobacillus brevis*

*Lactobacillus brevis* BLQ 6 (sourced from the RCW) was cultivated in MRS broth medium (Sifin Diagnostics GmbH, Germany) for 3 days at 28 ◦C and harvested by centrifugation (2500 g, 10 min). The cells were washed with sterile water and centrifuged again. *Lactobacillus brevis* was added to an autolysis process (resulting in a final concentration in the autolysis suspension of 106 CFU/mL) that was conducted as described in Section 2.4.2.

#### 2.4.4. Autolysis to Improve γ-aminobutyric Acid (GABA) Production

Process parameters of Masuda et al. were used to improve GABA production during autolysis [24]. Spent yeast (*S. cerevisiae*) after primary fermentation (12 ◦P) was washed (see Section 2.2) and added at a dry matter content of 7% to a solution containing sterile distilled water, monosodium glutamate (0.060 mol/L) (Merck, Darmstadt, Germany) and D-(+)-glucose monohydrate (0.266 mol/L) (Merck, Darmstadt, Germany). Following this, 200 mL of the reaction solution was adjusted to pH 6 with 2N HCl or 2N NaOH and incubated at 37 ◦C for 72 h with constant stirring (100 rpm). After 72 h the reaction solution was heated for 15 min at 85 ◦C. For the control, the process was conducted without monosodium glutamate or D-(+)-glucose monohydrate.

#### *2.5. Production of Yeast Extract*

After the yeast cell disruption process (Sections 2.4.1–2.4.4) cell wall components first had to be separated from the cell extract. Therefore, the samples were centrifuged for 20 minutes at 10,000 g and 4 ◦C. The supernatant was carefully pipetted out the centrifuge tubes and freeze-dried (Christ Alpha 1–4 LSCbasic, condenser temperature: −55 ◦C, vacuum: 0.1 mbar, ice condenser capacity: 4 kg/24 h). In this way, a yeast extract powder was produced for the subsequent analyses that offered a constant basis for comparison.

The dry yeast extracts enabled the disruption methods to be directly compared in terms of the following analysis without the need to consider the effectiveness of the different methods. Results on the effectiveness of the three disruption methods can be reviewed in a previous work by Jacob et al. [5]. An overview of the sample description and related process details can be seen in Table 2.

**Table 2.** Sample overview and process details of cropped surplus yeast after primary fermentation (F.) and cold lagering (L.); data are expressed as mean values; confidence limits were determined to be lower than 5% of the average value.


#### *2.6. Analysis*

#### 2.6.1. Protein and Amino Acids

Nitrogen content in the yeast extract was determined using the Kjeldahl method described in the MEBAK (Central European Brewing Technology Analysis Commission) brewing technology analysis methods (Method 2.6.1.1) [25]. Protein content of the yeast extracts was estimated by multiplying its nitrogen content by the factor 5.5 [26].

Free proteinogenic amino acids (except proline and cysteine) were quantified using high performance liquid chromatography (HPLC) according to MEBAK Method 2.6.4.1 [27]. The detailed procedure for proline and cysteine (Method 4.11.1) was taken from *Buch für chemische Untersuchung von Futtermitteln* (The chemical analysis of feedstuffs) [25]. To determine all free and protein-bound amino acids (total amino acid quantity), the resuspended yeast extracts underwent acid hydrolysis before measurements were taken according to Method 4.11.1 of the chemical analysis of feedstuffs [25].

#### 2.6.2. Fat

Crude fat was determined according to Method 5.1.1 from the *Methodenbuch für chemische Untersuchung von Futtermitteln* (The chemical analysis of feedstuffs) [25].

#### 2.6.3. Water and Ash Content

Water content was determined using MEBAK Method 2.2 [27], the ash content similarly according to Method 8.1 from the *Methodenbuch für chemische Untersuchung von Futtermitteln* (The chemical analysis of feedstuffs) [25].

#### 2.6.4. Folate

Vitamin B9 (total folate) and folate derivatives (5-CH3-H4folate, 5-CHO-H4folate, 10-CHO-PteGlu, H4folate, PteGlu) were analyzed using LC-MS/MS according to the recently published method by Striegel et al. [28].

#### 2.6.5. Antioxidative and Reduction Potential

The antioxidative potential of yeast extracts was measured using a Sigma-Aldrich kit, in which the antioxidants from a sample inhibit the formation of radical cations. Spectrophotometry was used to measure this inhibition proportionally by means of a color reaction. Trolox (TE), a vitamin E analog, was used as the control antioxidant.

Reduction potential of the yeast extract can be determined by MEBAK Method 2.15.2 [27]. The reductones of the sample reduce a certain quantity of Tillmann's reagent (2,6-dichlorophenolindophenol, DPI) within a certain period, which can be measured spectrophotometrically (520 nm).

#### 2.6.6. Wort Density

Wort density was measured using a DMA™ 35 Basic portable density meter (Anton Paar GmbH, Ostfildern, Germany), and the gravity was expressed in degree Plato (◦P). 1 ◦P corresponded to 1 g of extract per 100 g of liquid solution, where extract included both fermentable sugars and non-fermentable carbon sources.

#### 2.6.7. Calculations of Apparent Attenuation

Apparent attenuation (%) of the wort was the proportion of the wort dissolved solids (extract), which was fermented during fermentation:

Apparent attenuation (%) = [(original gravity (◦P) − final gravity (◦P))/(original gravity (◦P))] × 100.

#### *2.7. Statistical Evaluation*

All experiments were performed in triplicate and the relevant results given as arithmetic means. At a confidence level of 95%, the expected range (confidence interval) for each mean was calculated from the variance using Student's t-distribution. A single factor analysis of variance (ANOVA) and a paired t-test were performed to demonstrate differences between the results. "Significant" differences were described as having a *p*-value < 0.05. A test by Dixon was used to evaluate the results.

#### **3. Results and Discussion**

#### *3.1. General Nutrient Composition*

The nutrient value of the generated yeast extracts showed great variability. Table 3 lists the analysis results of the general nutrient composition of yeast extracts produced using the mechanical (ultrasonic sonotrode) method. To calculate the total protein content of the yeast extracts, the nitrogen quantity determined via Kjeldahl analysis was multiplied by the conversion factor 5.5 proposed by Reed et al. [26] This was proven to be a suitable conversion factor in our previous published papers in relation to yeast extracts [5,6] and was also used by Caballero-Cordoba et al. [29]. The factor of 6.25, which is generally used, overestimates the protein content as the total nitrogen volume contains the RNA nitrogen quantity (ribonucleic acids, 5–10% of the dry mass of yeast extract) as well as the proteinogenic nitrogen [5,6]. In the statistical evaluation of the obtained results, we established that the total protein content of the mechanically produced yeast extracts did not differ significantly (ANOVA *p*-value > 0.05) from the protein content of the autolytically produced yeast extracts (all nutritional data of the autolytically produced yeast extracts (Table S1) can be found in the supplementary materials). This observation was already noted and discussed in one of our previous papers [6]. Yeast extracts produced using primary fermentation yeast with different original wort contents (Scer 12 ◦P, Scer

16 ◦P, Scer 20 ◦P), differ significantly (ANOVA *p*-value < 0.05), with no significant difference between the fermentations Scer 16 ◦P and Scer 20 ◦P (*t*-test *p*-value > 0.05) (Table 3). This fact is justified in that adding maltose to the high-gravity worts (Scer 16 ◦P, Scer 20 ◦P) modified the nutrient balance compared with the normal-gravity worts (Scer 12 ◦P), and this caused a higher osmotic pressure at the start of fermentation and a higher alcohol content at the end of fermentation. Consequently, the yeast's vitality and viability dropped [30], which is associated with reduced specific growth and fermentation rates [18,31]. For high-gravity fermentations, this is also linked to reduced amino acid uptake rates, higher residual FAN (freely available amino nitrogen) and increased accumulation of trehalose and glycogen [1,18]. Presumably, this reduces the protein content of the yeast cell dry mass and ultimately results in a lower total protein content in the yeast extract. In our trials, when using spent yeast from high-gravity fermentations (Scer 16 ◦P, Scer 20 ◦P), we established a protein content in yeast extract reduced by 8.5% when compared with spent yeast from normal-gravity (Scer 12 ◦P) fermentations.

**Table 3.** General nutritional composition of yeast extracts made from spent yeast of beer production via a mechanical disruption method (ultrasonic sonotrode); influence of original gravity (12 ◦P, 16 ◦P, 20 ◦P), time of yeast cropping (after primary fermentation vs. after cold beer lagering (L)) and yeast strain (Scer, Spas, Slud, Tdel, Bbru, Sfib); for results of ANOVA and pairwise t-test, see text; data are expressed as mean values ± confidence limits; Cal.: calculated sum parameter of carbohydrates, RNA, non-nitrogen fraction and others components.


We could also see a significant reduction (*t*-test *p*-value < 0.05) of the protein content (3%) in the yeast extract when using lagering cellar yeast (Scer 12 ◦P L) compared with primary fermentation yeast (Scer 12 ◦P) (Table 3). At the end of primary fermentation, part of the yeast was still suspended and settled as "lagering cellar yeast" only once cold lagering began. In this connection, Powell et al. showed that the unsettled yeast had a lower cell age with reduced fermentation performance and flocculation tendency, suggesting a modified physiological state [32]. A modified physiological state could explain the different protein contents of the yeast extract produced from lagering cellar yeast (Scer 12 ◦P L) and primary fermentation yeast (Scer 12 ◦P), too. During longer lagering phases, proteinogenic material in the yeast could also be lost via excretion, as established by Steckley et al. [33].

The protein content of the yeast extracts produced from primary fermentation yeast with different yeast strains also differed significantly (ANOVA *p*-value < 0.05). The non-*Saccharomyces* yeast strain *S. fibuligera* TUM 525 (Sfib 12 ◦P) gave the highest value at 598 mg/g yeast extract. The strain often used to produce alcohol-free or low-alcohol beer, *S. ludwigii* TUM SL 17 (Slud 12 ◦P), had the second-highest value at 574 mg/g yeast extract. The yeast strain *B. bruxellensis* TUM Bret 1 (Bbru 12 ◦P) provided 508 mg protein per g yeast extract. Only for the non-*Saccharomyces* yeast strain *T. delbrueckii* TUM T 90 (Tdel 12 ◦P) and the two commercially used yeast strains *S. cerevisiae* TUM 68 (Scer 12 ◦P) or *S. pastorianus* TUM 34/70 (Spas 12 ◦P) the protein content was less than 500 mg/g yeast extract. Spas 12 ◦P gave the lowest value overall in our trials (411.54 mg/g). For a *S. pastorianus* yeast strain, Vieira et al. determined values of 698 mg/g and 765 mg/g yeast extract when reusing the yeast two to four times in the fermentation process [34]. In another work, the same group determined a protein content of 641 mg/g yeast extract [15]. The sampling time during fermentation was unknown and the higher conversion factor of 6.25 was applied to calculate the protein content. For a brewery spent yeast,

Podpora et al. established a protein content of 625 mg/g or 638 mg/g yeast extract, without giving details on the yeast strain or process conditions. Protein from the fermentation medium was also recorded for the yeast extract production and the higher conversion factor of 6.25 was applied [35].

There was no significant difference in the ash content of mechanically and autolytically produced yeast extracts (ANOVA *p*-value > 0.05), as demonstrated in our previous work when investigating various disruption methods [6]. The use of spent yeast from high-gravity fermentations (Scer 16 ◦P, Scer 20 ◦P) gave significantly lower values (*t*-test *p*-value < 0.05) for the ash content in the yeast extract than a spent yeast from normal-gravity fermentation (Scer 12 ◦P). This observation can be presumably attributed to the same effect that reduced the protein content in the yeast extract. Therefore, the ash percentage of the total dry mass could presumably be reduced by an increased trehalose and glycogen content due to the modified yeast metabolism. Another significantly lower result (*t*-test *p*-value < 0.05) was the ash concentration in the yeast extract caused by using lagering cellar yeast (Scer 12 ◦P L) rather than spent yeast obtained following primary fermentation (Scer 12 ◦P). The different physiological state of these two starting yeasts mentioned earlier is also suspected of influencing the cell ingredient composition. The ash content of the yeast extracts of all investigated yeast strains differed significantly (ANOVA *p*-value < 0.05). Only Bbru 12 ◦P and Scer 12 ◦P showed no significant difference (*t*-test *p*-value > 0.05). In literature, for the ash content, a range from 78 mg/g to 140 mg/g yeast extract is found [15,34,35]. Due to different yeast strains, fermentation media and yeast extract production processes, a direct comparison is not possible.

The fat content of all autolytically produced yeast extracts was between 0.04–0.05% of the dry mass and did not differ significantly (ANOVA *p*-value > 0.05). By using the mechanical cell disruption method, the fat content of the yeast extracts reached a maximum of 18.2 mg and a minimum of 9.68 mg per g yeast extract (Table 3). Neither the time of cropping the spent yeast (Scer 12 ◦P vs. Scer 12 ◦P L) or the original wort content (Scer 12 ◦P, Scer 16 ◦P, Scer 20 ◦P) in the fermentation process had any significant influence on the fat content of the yeast extract (ANOVA and *t*-test *p*-value > 0.05). Only the yeast strains Slud 17 12 ◦P and Bbru 12 ◦P differed significantly from the others (*t*-test *p*-value < 0.05). In general, the fat content of the yeast extracts was very low as already established in other studies [6,15,34].

#### *3.2. Amino Acid Composition*

From a physiological point of view, the proteinogenic material composition is crucial. The essential amino acids (His, Thr, Val, Met, Ile, Phe, Leu Lys) are indispensable for human nutrition as these cannot be synthesized by the body and must be absorbed via food [36]. The rapid usability of proteinogenic material in microbiological culture media is especially assured for yeasts if this material is present in the form of free amino acids, i.e. individual amino acids are not linked via peptide bonds [13]. This ensures the amino acids can be transported via various mechanisms through the cell wall and cell membrane and then to be metabolized [13]. Specific amino acids are preferentially absorbed by the cell [13]. In addition, individual amino acids can significantly affect the aroma metabolism of a yeast and thereby influence the overall aroma of a fermentation by-product [14]. In a previous study, we showed how different cell disruption methods impacted the amino acid composition and the proteinogenic material [5]. Following an autolytic process, the content of free amino acids in the yeast extract was significantly higher than that produced via mechanical disruption methods due to enzymatic degradation processes [5]. This was also confirmed in the trials presented here, as revealed by the comparison between mechanical (Figure 1) and autolytic (Figure 2) methods for the relevant trial series (*t*-test *p*-value < 0.05).

**Figure 1.** Free and protein-bound amino acids in yeast extracts made from spent yeast of beer production via mechanical disruption method (ultrasonic sonotrode); influence of original gravity (12 ◦P, 16 ◦P, 20 ◦P), time of yeast cropping (after primary fermentation vs. after cold beer lagering (L)) and yeast strain (Scer, Spas, Slud, Tdel, Bbru, Sfib); for results of ANOVA and pairwise t test, see text; data are expressed as mean values ± confidence limits.

**Figure 2.** Free and protein-bound amino acids in yeast extracts made from spent yeast of beer production via autolysis; influence of original gravity (12 ◦P, 16 ◦P, 20 ◦P), time of yeast cropping (after primary fermentation vs. after cold beer lagering (L)) and yeast strain (Scer, Spas, Slud, Tdel, Bbru, Sfib); for results of ANOVA and pairwise *t* test, see text; data are expressed as mean values ± confidence limits.

However, the aim of this work was to evaluate the influence of the starting material on the amino acid composition of the yeast extract according to the two disruption methods. It was shown that producing the yeast extract via sonotrode (Figure 1) generated significant differences (ANOVA *p*-value < 0.05) in the content of free amino acids in the yeast extract when using spent yeast from fermentation processes with different gravities (Scer 12 ◦P, Scer 16 ◦P, Scer 20 ◦P). No significant difference was recorded between the Scer 16 ◦P and Scer 20 ◦P test series (*t*-test *p*-value > 0.05). Using lagering cellar yeast (Scer 12 ◦P L) also generated a significantly lower content of free amino acids (*t*-test *p*-value < 0.05). In the comparison of the test series Scer 12 ◦P, Scer 12 ◦P L, Scer 16 ◦P, Scer 20 ◦P, it was also apparent that lower total quantities of all amino acids resulted in lower quantities of free amino acids. The content of free amino acids following mechanical disruption was likely to be derived largely from the free amino acid pool in the cell [1,5], which is influenced by the extraction process [37]. Amino

acids that were enzymatically released from the protein (despite a constant process temperature of 7 ◦C), were also present in the mechanically produced yeast extract as already evidenced in a previous paper [5]. The content of free amino acids in the yeast extract of yeast strains Scer 12 ◦P, Spas 12 ◦P, Slud 12 ◦P, Sfib 12 ◦P, Bbru 12 ◦P and Tdel 12 ◦P were significantly different (ANOVA *p*-value < 0.05 (Figure 1). There was no correlation between the total quantity of all amino acids (or bound amino acids) and the free amino acids in the yeast extract.

Significant differences were established for the test series Scer 12 ◦P and Scer 12 ◦PL(*t*-test *p*-value < 0.05) or Scer 12 ◦P, Scer 16 ◦P and Scer 20 ◦P (ANOVA *p*-value < 0.05) and no difference between Scer 16 ◦P and Scer 20 ◦P (*t*-test *p*-value > 0.05) following the autolytic process (Figure 2). When comparing the test series Scer 12 ◦P, Scer 12 ◦P L, Scer 16 ◦P, Scer 20 ◦P, it was also shown that the lower the total quantity of all amino acids, the higher the quantity of free amino acids. This indicated that the higher gravity in the fermentation process or due to the longer lagering period the spent yeast obtained during the autolytic yeast extract production process had a greater hydrolytic potential. The increased release of amino acids from the cell protein during the autolytic process was presumably attributed to a higher quantity of various proteinases in the yeast cells. An increased amount of proteinases in the fermentation medium was therefore reported for high-gravity fermentations, which is caused by excretion from the living yeast cell as well as by cell autolysis [38,39]. Fukal et al. also reported that yeast proteinases have high thermostability at a temperature of 50 ◦C [40], which corresponds to the autolysis process temperature selected in this study. It could also be evidenced that low yeast vitality is associated with greater proteinase excretion [38]. Yeast vitality at the end of lagering also drops substantially with beer production, whereby proteinase is released [38]. This could explain the increased proteolytic activity of the lagering cellar yeast during autolytic yeast extract production. Consequently, adding spent yeast from high-gravity fermentations and using lagering cellar yeast results in a higher amount of free amino acids in the yeast extract. From Figure 3 it is obvious that individual amino acids in the relevant test series (Scer 12 ◦P, Scer 12 ◦P L, Scer 16 ◦P, Scer 20 ◦P) are also released from the protein in different percentages. This means that not only the total quantity of free amino acids differed between the individual test series, but also the spectrum of individual free amino acids in the yeast extract. Yeast cells contain a variety of different proteolytic enzymes [41,42], which are likely to be present and active in different quantities in the respective test series. The range of individual free amino acids in the yeast extract therefore depends not only on the production method, as already shown [5], but also on the original wort content and the lagering period of the beer, from which the spent yeast originates. For the different yeast strains Scer 12 ◦P, Spas 12 ◦P, Slud 12 ◦P, Sfib 12 ◦P, Bbru 12 ◦P and Tdel 12 ◦P, we could show a significant difference (ANOVA *p*-value < 0.05) in the content of free amino acids in the autolytically produced yeast extracts (Figure 2). However, there was no correlation between the total quantity of amino acids and the free amino acids. The yeast extracts from spent yeast of non-*Saccharomyces* yeast contained a maximum of 200 mg free amino acids per g yeast extract. In contrast, the two commercially used yeast strains Scer 12 ◦P and Spas 12 ◦P provided 340 mg/g. Berlowska et al. determined 449.7 mg free amino acids per g yeast extract for a *S. cerevisiae* yeast strain [22]. For the analyzed non-*Saccharomyces* yeast species *K. marxianus, S. stipitis* and *P. angusta*, a free amino acid content was ranging of between 101.4 mg and 405.3 mg per g yeast extract [22]. It is not possible to directly compare these results with the current study due to the different production process and fermentation conditions. In the exemplary comparison, Figure 4 presents the detailed spectrum of free amino acids of yeast extracts, produced from spent yeast of the commercial yeast strain Spas 12 ◦P and the alternative non-*Saccharomyces* yeast strain Slud 12 ◦P. Significant differences (*t*-test *p*-value < 0.05) were found for the amino acids Asp, Glu, Asn, Ser, Gly, Thr, Tyr, Val, Trp, Ile, Phe and Leu. The amount of all individual amino acids (total, free) of conducted experiments can be found in Figures S1–S4 of the supplementary materials.

**Figure 3.** Percentage share of free amino acids of yeast extract (from *S. cerevisiae* TUM 68) protein released via autolysis; influence of original gravity (12 ◦P, 16 ◦P, 18 ◦P) and time of yeast cropping (after primary fermentation vs. after cold beer lagering (L)).

**Figure 4.** Free proteinogenic amino acid spectrum in yeast extracts made from spent yeast (*S. pastorianus* TUM 34/70 and *S. ludwigii* TUM SL 17) after primary fermentation via autolysis; for results of ANOVA and pairwise t-test, see text; data are expressed as mean values ± confidence limits.

The proteinogenic amino acid composition of the yeast extract depends on the production method [5] and, as shown above, on the starting yeast. If hygiene standards are not maintained when producing beer and storing yeast, the starting yeast could potentially be contaminated with microorganisms. The most common beer-spoilage organism in early stages of the beer production process is the species *Lactobacillus brevis* [43]. The ability of this bacteria species to convert glutamic acid into the nutritionally valuable γ-aminobutyric acid (GABA) has been proven [44]. Masuda et al. demonstrated that this reaction also proceeds during the autolysis of different yeast strains [24]. An intrinsic enzymatic mechanism with the enzyme glutamate decarboxylase is responsible for this reaction [24]. An excess of glutamic acid as a substrate and glucose as an energy supplier can increase GABA formation [24]. In a previous study we could establish an increased GABA concentration in the yeast extract for the strain *S. cerevisiae* TUM 68 [5] and also presumed this was caused by the mechanism postulated by Masuda et al. [24]. As all other yeast strains in this study had lower

GABA concentrations (see supplementary material) in their autolytically produced yeast extracts than *S. cerevisiae* TUM 68 (Scer 12 ◦P, 50 ◦C), we attempted to further increase GABA formation for the autolysis of *S. cerevisiae* TUM 68 (Scer 12 ◦P + Gluc + Glu, 37 ◦C), using the same process parameters as Masuda et al. At the same time, we investigated whether contamination of the yeast with *L. brevis* (Scer 12 ◦P + L, 50 ◦C) also produced an elevated GABA concentration in the autolytically produced yeast extract, or altered the proteinogenic amino acid composition. However, contamination with *L. brevis* (Scer 12 ◦P + L, 50 ◦C) influenced neither the proteinogenic amino acid spectrum of the yeast extract, nor did the GABA concentration differ significantly from the control sample (Figure 5) (*t*-test *p*-value > 0.05). The results coincided with the data from our previous study [5]. Champagne et al. also reported that they observed no significant influence of bacterial contamination on the extract yield, the total nitrogen, the FAN or the turbidity of an autolytically produced yeast extract [45]. The solvent ethyl acetate presumably inhibited the contaminants, reported by Champagne et al. [45] and in this study. Barrette et al. could therefore show a reduced viability in a bacteria population during autolytic yeast extract production using ethyl acetate [21]. Under the autolysis conditions of Masuda et al., the GABA concentration could be increased in the yeast extract of *S. cerevisiae* (Figure 5), which suggests an enzymatic mechanism with the enzyme glutamate decarboxylase being active. While the GABA concentration in our yeast extract of *S. cerevisiae* roughly doubled, Masuda et al. was able to increase the values for various *Candida* and *Pichia* strains by more than tenfold [24].

**Figure 5.** Gamma-aminobutyric acid in yeast extracts made from surplus yeast (*S. cerevisiae* TUM 68) after primary fermentation via autolysis (37 ◦C, 72 h or 50 ◦C, 24 h); Scer 12 ◦P (control); Scer 12 ◦P + Gluc + Glu (Scer 12 ◦P + glucose + glutamic acid); Scer 12 ◦P + L (Scer 12 ◦P + *Lactobacillus brevis*); for results of ANOVA and pairwise t-test, see text; data are expressed as mean values ± confidence limits.

#### *3.3. Folate Vitamer Distribution*

Yeasts and yeast extracts are known for their high content of different B vitamins, which include folates (vitamin B9) [6,15]. Folates, especially tetrahydrofolate-(H4folate)-polyglutamates, play an essential role in various metabolic pathways such as amino acid synthesis in mitochondria and DNA replication in the cell nucleus [28]. The human body cannot generate vitamin B9 itself and therefore needs to obtain an adequate supply via diet [28]. In a previous study, we already showed that yeast extracts can be produced with a folate content between 1.35 and 4.94 mg/100 g using a spent yeast (*S. cerevisiae* TUM 68) from a top-fermenting brewing process [6]. In this context, we presented the influence of different mechanical and autolytic extraction methods on the content of the folate vitamers 5-CH3-H4folate, 5-CHO-H4folate, 10-CHO-PteGlu, H4folate, and PteGlu. Hjortmo et al. evidenced a total folate content in a range of 4000–1,4500 μg/100 g yeast dry matter for various yeast strains, wherein the sample material came from the exponential growth phase of the yeast population and a synthetic culture medium was used [46]. The spent yeast from beer preparation is cropped at the end of the fermentation process or as lagering cellar yeast and as such is already in the stationary phase of yeast cell growth. In this phase, the folate content of the yeast dry mass drops sharply and remains

at a lower level than in the exponential growth phase [47]. The material composition of the culture medium also plays a critical role [47].

In order to provide a comprehensive picture of the process, we also established the folate content of the fermented medium and the spent yeast in addition to the autolytically and mechanically produced yeast extract. This showed that all fermented media had a low total folate content (6–17 μg/100 g), calculated from the folate vitamers 5-CH3-H4folate, 5-CHO-H4folate, 10-CHO-PteGlu, H4folate and PteGlu. No significant difference could be determined between the investigated fermented media of any test series (Scer 12 ◦P, Scer 12 ◦P L, Scer 16 ◦P, Scer 20 ◦P, Bbru 12 ◦P, Slud 12 ◦P, Tdel 12 ◦P, Spas 12 ◦P, Sfib 12 ◦P) (ANOVA *p*-value > 0.05). Hjortmo et al. could not show any folate in the fermented medium of a *S. cerevisiae* strain [47]. This can be explained by the fact that there were no more yeast cells in the supernatant [47]. In our work, however, the supernatants still contained suspended yeast cells, which were not removed by centrifuging before folate analysis. Regarding the yeast extract, the original wort content (Scer 12 ◦P, Scer 16 ◦P, Scer 20 ◦P) had no significant influence (ANOVA *p*-value > 0.05) on the total folate content of the spent yeast (Figure 6). High-gravity fermentations generally lead to lower specific yeast growth rates than normal-gravity fermentations [18]. For a continuous yeast culture (chemostat), Hjortmo et al. demonstrated a significant positive correlation between specific growth and the total folate content of a yeast population [47]. This has been evidenced for the exponential growth phase of a yeast population [47]. The primary fermentation yeast is in the stationary phase. In this case, the total folate content of the spent yeasts from the normal and high-gravity fermentations did not differ and presumably settled at around the same level, irrespective of the effect of the different specific growth rates experienced previously. In contrast, the precipitated lagering cellar yeast (Scer 12 ◦P L) had a significantly lower total folate content than that of the primary fermentation yeast (*t*-test *p*-value < 0.05). During the cold lagering process, the non-precipitated primary fermentation yeast stays in the stationary phase and precipitates within this period as lagering cellar yeast (12 ◦P L). No more cell division occurs during this period, which presumably caused the constant degradation of the folate vitamers required for amino acid synthesis and DNA replication. The ratio of the individual folate vitamers in the spent yeast (Scer 12 ◦P, Scer 12 ◦P L, Scer 16 ◦P, Scer 20 ◦P) did not change significantly (ANOVA *p*-value > 0.05). The yeast extracts produced via sonotrode had higher folate contents in each case than the corresponding spent yeasts. Removing the insoluble cell components (mainly cell walls) enriched the folate content. Once again, no significant influence (ANOVA *p*-value > 0.05) of the original wort content (Scer 12 ◦P, Scer 16 ◦P, Scer 20 ◦P) on the total folate content or the distribution of the individual folate vitamers could be observed. The total folate content of the autolytically produced yeast extracts was between 800 and 1400 μg/100 g and did not differ significantly (ANOVA *p*-value > 0.05).

In a comparison of the different yeast strains (Figure 7), Scer 12 ◦P with 3640 μg/100 g yeast dry mass had the highest total folate content in the spent yeast. A value of 2970 μg/100 g could be determined for the spent yeast of the yeast strain Bbru 12 ◦P. We established the lowest total folate content (1930 μg/100 g) for the yeast strain Spas 12 ◦P and could determine no significant difference to Sfib 12 ◦P (*t*-test *p*-value > 0.05). Between the strains Slud 12 ◦P and Tdel 12 ◦P there was also no significant difference in the total folate content of the spent yeast (*t*-test *p*-value > 0.05) (Figure 7). The quantity of the physiologically valuable folate vitamer 5-CH3-H4folate varied in the spent yeast of all the investigated yeast strains (ANOVA *p*-value < 0.05). It was also observed that the ratio of 5-CH3-H4folate to the other folate vitamers, 5-CHO-H4folate, 10-CHO-PteGlu, H4folate and PteGlu, differed significantly in the spent yeast of the particular yeast strains (ANOVA *p*-value < 0.05) (Figure 7). At 2200 μg/100 g the proportion of 5-CH3-H4folate was approximately 60% of the total folate content of the spent yeast Scer 12 ◦P, whereas this was only around 20% for Tdel 12 ◦P. The total folate content of the yeast extracts (sonotrode) from the spent yeasts Scer 12 ◦P, Bbru 12 ◦P and Spas 12 ◦P differed significantly from the other mechanically produced yeast extracts (*t*-test *p*-value < 0.05). The yeast extracts Slud 12 ◦P, Tdel 12 ◦P and Sfib 12 ◦P, however, did not differ significantly (*t*-test *p*-value > 0.05). The high total folate content (6000 μg/100 g) of the yeast extract from the spent yeast Bbru 12 ◦P was

striking with regard to the total folate content of the corresponding spent yeast (2970 μg/100 g). The proportion of the physiologically valuable folate vitamers 5-CH3-H4folate of the total folate constant was also not constant in the mechanically produced yeast extracts. The total folate content of the autolytically produced yeast extracts in Figure 7 was between 800 and 2100 μg/100 g. Only the Tdel 12 ◦P yeast extract differed significantly (*t*-test *p*-value < 0.05) from the others. The influence of the different production methods (autolysis, sonotrode) on the total folate content and the folate vitamer distribution has already been discussed in our previous work [6].

**Figure 6.** Distribution of the folate vitamers 5-CH3-H4folate, 5-CHO-H4folate, 10-CHOPteGlu, H4folate and PteGlu in spent yeast (*S. cerevisiae* TUM 68) and in the corresponding yeast extract (via sonotrode or autolysis); influence of original gravity (12 ◦P, 16 ◦P, 20 ◦P), time of yeast cropping (after primary fermentation vs. after cold beer lagering (L)); for results of the pairwise t-test, see text; data are expressed as mean values ± confidence limits.

**Figure 7.** Distribution of the folate vitamers 5-CH3-H4folate, 5-CHO-H4folate, 10-CHOPteGlu, H4folate and PteGlu in spent yeast, yeast extracts (via sonotrode or autolysis) ; influence of yeast strain (Scer, Spas, Slud, Tdel, Bbru, Sfib); for results of the pairwise t-test, see text; data are expressed as mean values ± confidence limits.

#### *3.4. Biological Activity*

Yeast cells have a wide range of different functional components, which impart bioactive properties to a yeast extract once extracted from the cell. These include peptides, amino acids, flavonoids, polyphenols and carotenoids [15,34,48]. Through autolysis, Vieira et al. increased the anti-oxidative potential of a mechanically produced yeast extract due to the release of phenolic components and amino acids [49]. When the thermal load is too high, the anti-oxidative potential drops, which, according to Vieira et al., is likely attributed to the breakdown of phenolic components, vitamins and bioactive peptides [49]. We evidenced this correlation in an earlier work as follows. We observed a significant drop in concentration of the bioactive peptide glutathione or the total polyphenol content following an autolytic yeast extract production method exposed to thermal stress compared with a mechanical process at an appropriate thermal load [6]. The process parameters of the two yeast extract production methods from our previous study [6] corresponded to those in these trials. Since the yeast extract of the autolytic process exhibited low bioactivity in the previous study, we only investigated the reductive and anti-oxidative potential of the mechanically produced yeast extract in this work (Figure 8). We found no significant difference in the reduction potential of the test series Scer 12 ◦P and Scer 12 ◦P L (*t*-test *p*-value > 0.05). There was a significant difference between Scer 12 ◦P, Scer 16 ◦P and Scer 20 ◦P (ANOVA *p*-value < 0.05), but with no significant difference between Scer 12 ◦P and Scer 16 ◦P (*t*-test *p*-value > 0.05). The results were similar for the anti-oxidative potential, with no significant difference between Scer 16 ◦P and Scer 20 ◦P. The biological activity of the yeast extracts of different yeast strains varied considerably, with (Scer 12 ◦P) having the highest values, followed by Tdel 12 ◦P and Spas 12 ◦P. Significant differences (ANOVA *p*-value < 0.05) could be established for all the investigated yeast strains in relation to the reductive and anti-oxidative potential. Only the reduction potential of Spas 12 ◦P and Sfib 12 ◦P were not significantly different (*t*-test *p*-value > 0.05). A wide range of different components such as peptides, vitamins, phenolic components and enzymes were responsible for the biological activity [6,15,49].

**Figure 8.** Reduction potential of yeast extract made from spent yeast (*S. cerevisiae* TUM 68) of beer production via mechanical disruption method (sonotrode); influence of original gravity (12 ◦P, 16 ◦P, 20 ◦P), time of yeast cropping (after primary fermentation vs. after cold beer lagering (L)) and yeast strain (Scer, Spas, Slud, Tdel, Bbru, Sfib); for results of ANOVA and pairwise t-test, see text.

#### **4. Conclusions**

This study showed that the biodiversity of spent yeast from the brewing process means that it can substantially influence the composition of physiologically important ingredients in the resulting yeast extract. The yeast strain (commercial *Saccharomyces* and alternative non-*Saccharomyces* yeast strains), the original wort content of the fermentation medium and the spent yeast cropping time, have a direct

impact on different components of the yeast extract. The general nutrient composition (protein, fat and ash content) of the yeast extracts displayed significant differences after using various spent yeasts. We have evidenced in detail that the release of proteinogenic amino acids during autolysis of the spent yeasts differed greatly and thereby influenced the yeast extracts' FAN. The respective proteinogenic amino acid spectrum also varied. Contamination of the spent yeast with the beer spoiler *L. brevis* had no impact on the amino acid profile of the yeast extracts. Using relevant autolytic process conditions made it possible to increase the GABA concentration in the yeast extract of *S. cerevisiae* TUM 68, however a commercial use in this context is doubtful. It was possible to influence both the total folate content and the proportion of individual folate vitamers by using different yeast strains. Lagering cellar yeast as a starting material to produce yeast extract resulted in lower total folate contents in the yeast extract than primary fermentation yeast. The original wort content of the fermentation medium had no significant influence on the total folate of the spent yeast or the yeast extract. The biological activity (reduction and antioxidative potential) of the yeast extracts also depended on which spent yeast was used from the brewing process. The top-fermenting yeast strain *S. cerevisiae* TUM 68 gave particularly high values. In conclusion, the results indicate that brewer's spent yeast should be carefully selected to produce a yeast extract with a defined nutritional composition. A further research objective would be to adapt the autolytic or mechanical production process for brewer's spent yeast to influence the content of physiologically important ingredients in a yeast extract.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2311-5637/5/2/51/s1.

**Author Contributions:** F.F.J. designed and performed the experiments, analyzed the data and wrote the paper; L.S. and M.R. contributed reagents/materials/method/analysis tools for folate analysis and supported the practical implementation. M.H. selected/provided yeast and bacteria strains and isolated some of the strains in previous studies. M.R., M.H. and F.-J.M. revised the conception and manuscript, and agreed for submission.

**Funding:** This research received no external funding.

**Acknowledgments:** The authors thank Martin Neumeier (Technische Universität München–Lehrstuhl für Analytische Lebensmittelchemie, Freising, Germany) and Steffen Pfeil (Technische Universität München–Forschungszentrum Weihenstephan für Brau- und Lebensmittelqualität, Freising, Germany) for technical assistance.

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

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
