Enhancing Mead Aroma Using Non-Saccharomyces Yeast β-Glucosidase Producers Isolated from Honey: A Case Study in the Upper Turi Region

Abstract

The Upper Turi region in the Maranhão Amazon is a significant producer of honeybees, and mead production represents a cost-effective means of adding value to the honey production chain. This study investigates non-Saccharomyces yeasts isolated from honey as β-glucosidase producers to enhance the mead aroma. Sixty-five honey samples from the Upper Turi in Maranhão underwent yeast screening. Biochemical tests identified isolated yeasts, and β-glucosidase-producing strains were selected via esculin agar. Meads were produced using selected strains of Saccharomyces cerevisiae. Fermentation analyses included pH, °Brix, temperature, conductivity, dissolved oxygen, and volatile compounds (GC-MS). Thirty-six yeasts were isolated, with three identified as β-glucosidase producers. Strain 20 (Saccharomycopsis fibuligera) was selected for mead production due to its fermentative capacity, tolerance to pH and ethanol, and its ability to produce β-glucosidase, which hydrolyzes the glycosidic precursors in honey. During alcoholic fermentation, Saccharomycopsis fibuligera exhibited lower fermentative potential compared to Saccharomyces cerevisiae, reducing only 3.7% of the initial soluble solids (°Brix). The pH and temperature remained relatively stable throughout the fermentation for both yeast strains. The levels of dissolved oxygen and conductivity in the fermented mead were higher for S. cerevisiae than for Saccharomycopsis fibuligera. Specifically, S. cerevisiae showed reductions of 52.85% in dissolved oxygen and conductivity of 1115 µS/cm, while Saccharomycopsis fibuligera exhibited reductions of 33.0% in dissolved oxygen and conductivity of 511 µS/cm. Although the β-glucosidase-producing yeast yielded a mead with a low ethanol concentration, it contributes a unique fruity compound (ethyl hexanoate) and avoids high acetic acid production, providing a distinct aromatic profile that can be explored.

1. Introduction

Honey is primarily composed of sugars—mainly fructose and glucose—and water; which together account for approximately 95% of its dry weight. The water content is a critical factor influencing honey’s shelf life. In addition to these major components, honey contains various minor constituents such as amino acids, organic acids, enzymes, vitamins, pigments, minerals, and pollen grains [1].

It also contains a diverse microbiota, including bacteria, molds, and yeasts. These microorganisms are adapted to survive osmophilic conditions, acidity, and the presence of natural antimicrobial compounds, often introduced during harvesting and processing [2,3]. They can be categorized into two groups: (i) those naturally occurring in honey, such as molds and yeasts, which are non-pathogenic and do not affect quality under normal moisture conditions; and (ii) those that indicate hygienic or commercial quality, such as mesophilic bacteria, total coliforms, thermotolerant coliforms, molds, and yeasts [4].

Among these microorganisms, yeasts are particularly relevant due to their role in fermentation. They convert sugar into ethanol and produce enzymes such as pectinases and glycosidases. A prominent species is Saccharomyces cerevisiae, widely used as an inoculum in beverage production. This yeast contributes to the development of flavor and aroma through the generation of volatile compounds, including aldehydes, ketones, alcohols, terpenes, and esters [5].

The aromatic potential of yeasts is closely associated with their enzymatic activity, especially β-glucosidases. These enzymes hydrolyze glycosidic bonds in plant-derived precursors, such as terpenes and polyphenols, releasing volatile aglycones that enhance aroma complexity [6]. While S. cerevisiae is well-established, non-Saccharomyces yeasts are gaining attention for their unique enzymatic capabilities, including the production of higher levels of β-glucosidase, which can yield distinctive flavor profiles in fermented beverages, such as wine and beer [7,8,9].

Interestingly, β-glucosidase activity has also been observed in the honeybee Apis mellifera, which is believed to produce the enzyme in its hypopharyngeal glands. The enzyme is secreted during feeding and transferred to the honey sac [10]. In addition to Apis mellifera, honey produced by the stingless bee Melipona fasciculata, a native species known for its unique floral honey, was also considered to explore broader microbial diversity and regional characteristics.

In beverage production, particularly in mead, which is derived from the alcoholic fermentation of honey, solutions, the selection of yeast strains is a determining factor in defining product characteristics. While traditional methods rely on spontaneous fermentation by indigenous yeasts, controlled fermentation using selected strains allows for enhanced sensory qualities and production consistency. Isolating yeasts directly from honey enables the development of meads with unique regional and floral attributes [11,12].

Based on this context, the present study aimed to isolate and characterize yeasts from honey produced by Apis mellifera and Melipona fasciculata, focusing on strains with high β-glucosidase activity. The primary goal was to evaluate their potential to generate alcoholic fermentation with enriched aromatic profiles. Investigating honey-derived yeasts represents an opportunity to obtain mead with distinct and regionally influenced characteristics.

2. Materials and Methods

2.1. Isolation and Enumeration of Yeasts in Honey Samples

The sample collection was realized in the informal trade from six municipalities within the Upper Turi Region in the countryside of Maranhão State between July and August 2017. A total of 65 honey samples were obtained, without registration or inspection, including 39 honey samples from Melipona fasciculata (Tuiba bees and stingless bees) and 26 honey samples from Apis mellifera (Africanized bees and honeybees). The samples were transported in isothermal boxes to the Food Microbiology Laboratory of the Federal Institute of Maranhão, where microbiological analyses were immediately conducted following APHA (2001) guidelines [13].

Initially, decimal dilutions of the samples were prepared in a 0.85% NaCl saline solution. Aseptically, 25 g of each sample was weighed into a flask containing 225 mL of 0.85% NaCl saline solution (10⁻¹ dilution). From this dilution, 10⁻² and 10⁻³ dilutions were prepared, and subsequent analyses were performed on these dilutions. Yeasts were isolated and counted from the decimal dilutions by plating on Malt Extract agar supplemented with chloramphenicol (100 mg/L) using the spread-plate technique. The plates were incubated at 30 °C for 4 days. Results were expressed in Colony Forming Units (CFU/25 g). The morphologically different and representative colonies were selected and isolated on yeast extract peptone glucose (YEPG) agar medium at 30 °C for two days. The yeast colonies were grouped based on morphological characteristics (color, shine, format, margin, surface, elevation, and consistency) on agar and fermentability of carbohydrates (glucose, fructose, galactose, sucrose, raffinose, lactose, and maltose), according to the protocol described by Kurtzman et al. [14].

2.2. Selection of ß-Glucosidase-Producing Yeasts

After isolation, the selection of yeasts producing β-glycosidases was conducted on plates containing esculin as a substrate in an esculin glycerol (EG) agar medium (1 g/L esculin, 0.3 g/L ferric chloride, 1 g/L hydrolyzed casein, 25 g/L yeast extract, 8 g/L glycerol, 20 g/L agar). The culture medium was autoclaved at 121 °C for 15 min and then transferred to Petri dishes until solidification. Each plate was inoculated with four yeast strains, incubated at 25 °C, and examined after 2, 4, 6, and 8 days. Yeast strains exhibiting a dark brown halo, indicative of esculin hydrolysis, were considered positive for β-glucosidase activity, following the methodology of Peréz et al. [6]. The diameter of these halos was measured using a millimeter ruler. This methodology, utilizing esculin as a substrate to screen for β-glucosidase-producing yeasts, is a rapid, easy application and well-established technique reported in the literature for cacao and wine fermentation and has proven effective in selecting yeasts capable of producing volatile compounds in alcoholic fermentation [15].

Cultures of isolated yeast strains incubated on YEPG liquid medium at 30 °C for 2 days were used for growth and fermentation capacities using different carbohydrates at 2% (glucose, fructose, galactose, sucrose, raffinose, lactose, maltose) by inoculating 100 µL of yeast culture in test tubes containing Durham tubes and 5 mL of liquid medium, yeast extract medium (4.5 g/L), peptone (7.5 g/L), and phenol red indicator. Resistance to different temperatures (37 and 45 °C), tolerance to different concentrations of ethanol (5, 10, and 15% ethanol), and growth at different pH (2.5, 4.5, and 7.5) were tested on tubes containing 5 mL of the YEGP liquid medium (3 g/L yeast extract, 5 g/L peptone, and 10 g/L glucose) and 100 µL yeast culture and were assessed according to the methods recommended by Kurtzman et al. [14]. The tubes were incubated at 30 °C, and yeast growth was monitored by visual observation every 24 h for 4 days. Yeast growth was determined by the presence of turbidity in the medium and gas production within the Durham tubes. The absence of yeast growth was considered when no changes were observed in the medium.

2.3. Mead Production with ß-Glucosidase-Producing Yeast

To produce alcoholic fermented mead from bee honey, a single yeast strain was selected based on its high potential for ethanol and β-glycosidase production. Fermentations were carried out in triplicate, and a control fermentation using Saccharomyces cerevisiae yeast was also conducted. Additionally, a wort sample without an inoculum was evaluated to observe spontaneous fermentation.

The fermentation process followed the protocols described by Duarte et al. [16] and Sadoudi et al. [17]. The must was obtained by diluting honey in distilled water at a ratio of 1:10 (weight/volume) to achieve a sugar content of up to 16 °Brix. The pH was adjusted to 4.5 with sodium carbonate solution, and potassium metabisulfite was added at a concentration of 100 mg/L to inhibit bacterial growth. Yeasts were pre-cultured in YM broth (1% glucose, 0.5% peptone, 0.3% malt extract, 0.3% yeast extract) at 30 °C for 48 h. The pasteurized wort was then inoculated with the yeast pre-culture and incubated at 22 °C without stirring. Fermentation progress was monitored through sugar concentration (Brix degree measurements), CO₂ production, and temperature every 24 h to 144 h. The end of fermentation was determined when the Brix degree remained stable.

At the end of fermentation, the first racking was performed. This involved transferring the fermented product to another container and eliminating the solid residues that had settled at the bottom of the fermenter container. Subsequently, the fermented product was stored in an incubator at a temperature of 10 °C. The second racking took place 10 days after the conclusion of fermentation, and, after 30 days, the third racking was conducted with aeration of the fermented must. The last racking occurred 10 days later, followed by filtration using cellulose filters. The clarified product was then stored in glass bottles at a temperature of 10 °C.

2.4. Physicochemical Analyses of the Mead

To assess the fermentation process, a series of physical and chemical analyses, including pH, temperature, Brix degree, and aromatic compounds in the mead, were conducted following the method outlined by Instituto Adolfo Lutz [18].

The relative density at 20 °C/20 °C was determined using a pycnometer that had been thoroughly rinsed with alcohol and then with ether before weighing the sample and distilled water at 20 °C. The density was calculated using the following formula: Density = (Mass of the pycnometer with the sample − Mass of the empty pycnometer)/(Mass of the pycnometer with water − Mass of the empty pycnometer).

The actual alcohol content was determined by transferring 100 mL of the sample into a distillation flask connected to the distillation system. Three-quarters of the initial volume was removed, and the final volume was adjusted to 200 mL with distilled water. After stirring and cooling, the distillate was transferred to a measuring cylinder for alcohol content measurement at 20 °C using an alcoholmeter.

The pH was measured using a pre-calibrated pH meter with buffer solutions. The °Brix readings of both the must and the final fermented product were measured using a digital refractometer.

To assess the efficiency of sugar conversion into alcohol, sugar concentration was determined using the following equation: Sugar Concentration (g/L) = °Brix × 10.33 + 1.445. The concentration of sugars (g/L) was determined following the method by Torres Neto et al. [19].

The levels of dissolved oxygen and conductivity during the alcoholic fermentation of mead were determined using a multiparameter analyzer. One hundred (100) milliliters of sample for each mead produced was collected and used for the direct determination of pH, temperature, conductivity, and dissolved oxygen. For calibration, a 0.1 M KCl solution (Metrohm). This method is for rapid monitoring of the alcoholic fermentative process [20,21].

2.5. Determination of Volatile Compounds in Mead

To analyze the volatile compounds in the mead, a gas chromatograph coupled with a mass spectrometer was employed. The procedure involved capturing volatile compounds from the mead using the solid-phase microextraction (SPME) method. This method consisted of transferring 5 mL of the mead sample into a 40 mL glass vial equipped with a sealed lid and a Teflon septum. A magnetic bar was added to facilitate homogenization. The 40 mL vial was then heated in a water bath at 40 °C on a shaker-heater. After an equilibration time of 10 min, a Carboxen/Divinylbenzene/PDMS fiber (50/30 µm) was exposed to the sample for 30 min. Subsequently, the fiber was injected into the gas chromatograph coupled to a mass spectrometer for analysis. This analysis was conducted by the analytical center at the University of São Paulo.

2.6. Statistical Analysis

Graphs and tables were generated using Excel software (Microsoft 365). The similarity among yeast isolates (based on biochemical tests) and the relationship between mead samples and volatile compounds were assessed by hierarchical clustering analysis (Ward’s method, Euclidean distance) and principal component analysis (PCA) using Past software, version 3.8 [22].

3. Results

3.1. Quantification of Yeasts in Honey Samples

The quantification of yeasts in 65 honey samples from six municipalities revealed microbial counts ranging from 0 to 928 CFU/g (

Table 1. Quantification of yeast from honey in municipalities from Upper Turi region, Maranhão, Brazil.

N° of Samples	Municipalities	Bee Type	Specie Bee	Yeast (CFU/ g)
1	Viana	Tiúba	Melipona fasciculata	928
2	Newton Bello	Africanized	Apis mellifera	88
3	Nova Olinda	Africanized	Apis mellifera	548
4	Santa Luzia	Tiúba	Melipona fasciculata	62
5	Santa Luzia	Africanized	Apis mellifera	57
6	Santa Luzia	Africanized	Apis mellifera	117
7	Viana	Africanized	Apis mellifera	0
8	Viana	Tiúba	Melipona fasciculata	22
9	Nova Olinda	Tiúba	Melipona fasciculata	138
10	Santa Luzia	Tiúba	Melipona fasciculata	7
11	Viana	Africanized	Apis mellifera	94
12	Viana	Africanized	Apis mellifera	308
13	Nova Olinda	Tiúba	Melipona fasciculata	195
14	Santa Luzia	Tiúba	Melipona fasciculata	102
15	Santa Luzia	Africanized	Apis mellifera	170
16	Santa Luzia	Tiúba	Melipona fasciculata	37
17	Santa Luzia	Tiúba	Melipona fasciculata	36
18	Santa Luzia	Africanized	Apis mellifera	0
19	Maranhãozinho	Africanized	Apis mellifera	316
20	Maranhãozinho	Africanized	Apis mellifera	272
21	Nunes Freire	Africanized	Apis mellifera	164
22	Nunes Freire	Africanized	Apis mellifera	174
23	Nunes Freire	Africanized	Apis mellifera	594
24	Viana	Africanized	Apis mellifera	1
25	Viana	Tiúba	Melipona fasciculata	5
26	Viana	Africanized	Apis mellifera	13
27	Viana	Africanized	Apis mellifera	0
28	Viana	Tiúba	Melipona fasciculata	9
29	Newton Bello	Africanized	Apis mellifera	0
30	Nova Olinda	Africanized	Apis mellifera	104
31	Santa Luzia	Tiúba	Melipona fasciculata	1
32	Viana	Africanized	Apis mellifera	0
33	Viana	Tiúba	Melipona fasciculata	0
34	Viana	Tiúba	Melipona fasciculata	0
35	Newton Bello	Africanized	Apis mellifera	0
36	Viana	Tiúba	Melipona fasciculata	13
37	Nova Olinda	Africanized	Apis mellifera	0
38	Nova Olinda	Africanized	Apis mellifera	0
39	Nova Olinda	Africanized	Apis mellifera	0
40	Nova Olinda	Africanized	Apis mellifera	0
41	Nova Olinda	Africanized	Apis mellifera	0
42	Nova Olinda	Africanized	Apis mellifera	14
43	Nunes Freire	Tiúba	Melipona fasciculata	0
44	Nunes Freire	Tiúba	Melipona fasciculata	0
45	Nunes Freire	Tiúba	Melipona fasciculata	0
46	Nunes Freire	Tiúba	Melipona fasciculata	0
47	Nunes Freire	Tiúba	Melipona fasciculata	0
48	Nunes Freire	Tiúba	Melipona fasciculata	0
49	Nunes Freire	Tiúba	Melipona fasciculata	0
50	Maranhãozinho	Tiúba	Melipona fasciculata	0
51	Maranhãozinho	Africanized	Apis mellifera	0
52	Maranhãozinho	Africanized	Apis mellifera	0
53	Maranhãozinho	Africanized	Apis mellifera	0
54	Maranhãozinho	Africanized	Apis mellifera	0
55	Maranhãozinho	Tiúba	Melipona fasciculata	0
56	Maranhãozinho	Tiúba	Melipona fasciculata	0
57	Maranhãozinho	Africanized	Apis mellifera	0
58	Newton Belo	Tiúba	Melipona fasciculata	2
59	Newton Belo	Tiúba	Melipona fasciculata	2
60	Newton Belo	Tiúba	Melipona fasciculata	0
61	Newton Belo	Tiúba	Melipona fasciculata	0
62	Newton Belo	Tiúba	Melipona fasciculata	0
63	Newton Belo	Tiúba	Melipona fasciculata	0
64	Newton Belo	Tiúba	Melipona fasciculata	0
65	Newton Belo	Tiúba	Melipona fasciculata	0

). Brazilian legislation does not establish microbiological standards for honey from native bees. Although some states have their own regulations, Maranhão lacks specific legislation regarding the microbiological quality of honey from native bees.

Low counts of yeast were detected in honey samples from the Upper Turi region of Maranhão, Brazil. Only 21.5% (n = 14) of samples presented yeast populations above 100 CFU/g. However, honey samples from Apis mellifera (24.6%) exhibited a higher prevalence of yeasts (Supplementary Materials, Figure S1). At the state level, Rio Grande do Norte [23], Paraná [24], São Paulo [25], and Santa Catarina [26] have set a maximum limit of 10⁴ CFU/g as the microbiological criterion for molds and yeasts. Based on these criteria, all samples analyzed in this study are within the established limits. Therefore, most samples presented low levels of contamination by yeasts, demonstrating satisfactory conditions from a food safety point of view.

3.2. Selection of ß-Glucosidase-Producing Yeasts

The morphological characterization of the 33 yeast strains isolated from honey samples was conducted based on Kurtzman et al. [14]. These colonies displayed various morphological traits, with differing sizes, circular shapes, and elevations ranging from convex to planar. Most colonies had regular margins, smooth and glossy or opaque surfaces, and colors that varied between yellow, white, pink, or cream, with a creamy consistency (Supplementary Materials, Table S1).

Most yeast strains demonstrated the ability to ferment sucrose, representing 69% of isolates. Galactose and glucose were the second most fermented sugars, at 39.4% and 33.3%, respectively. Fructose, lactose, maltose, and raffinose were fermented to a lesser extent, with 24.2%, 21.2%, 21.2%, and 12.1%, respectively (Supplementary Materials, Table S2).

Regarding stressful conditions, such as variations in pH, temperature, and ethanol concentrations, most yeast strains exhibited limited fermentative activity with glucose but maintained growth. Approximately 33.3 and 27.3% of the isolates showed fermentative activity within the pH range of 4.5 to 7.5, respectively. However, only 21.2% of the yeast strains displayed fermentative activity under extremely low pH conditions (pH 2.5). The pH significantly affected yeast growth and fermentative capabilities. The honey analyzed had a pH of 3.43, falling within the pH range of 2.5 to 4.5, where most isolated yeasts displayed higher fermentative capacity (Supplementary Materials, Table S3).

Only 15.2% of the isolates exhibited fermentative activity at 5% (v/v) ethanol. Strain 11 exhibited ethanol tolerance and fermentative capacities across all analyzed concentrations. Most ethanol-resistant yeast strains (19%) grew at the lowest ethanol concentration (5%, v/v). All yeast strains were sensitive to temperatures above 45 °C (Table S3).

After conducting the tests, 13 strains were classified into six species based on morphological characteristics, carbohydrate fermentation, fermentative capacity at different temperatures, and pH (

Table 2. Identification and ß-glycosidase production by yeasts isolated from honey in the Upper Turi region.

N° of Strains	Bee Specie	ß-Glycosidase Producer	Identification
1	M. fasciculata	-	*
2	M. fasciculata	-	*
3	M. fasciculata	-	*
4	M. fasciculata	-	*
5	M. fasciculata	-	*
6	M. fasciculata	-	Rhodoturola
7	M. fasciculata	-	*
8	M. fasciculata	-	Rhodoturola
9	M. fasciculata	-	*
10	A. mellifera	-	Kluyveromyces marxianus
11	A. mellifera	-	Kluyveromyces lactis var lactis
12	A. mellifera	-	*
13	A. mellifera	-	*
14	A. mellifera	-	*
15	A. mellifera	-	*
16	A. mellifera	-	*
17	A. mellifera	-	*
18	A. mellifera	-	Candida auringiensis
19	M. fasciculata	+	*
20	M. fasciculata	+	Saccharomycopsis fibuligera
21	A. mellifera	+	Rhodoturola
22	A. mellifera	-	Rhodoturola
23	M. fasciculata	-	*
24	M. fasciculata	-	Candida taylori
25	M. fasciculata	-	*
26	A. mellifera	-	*
27	A. mellifera	-	*
28	A. mellifera	-	Candida auringiensis
29	A. mellifera	-	Rhodoturola
30	A. mellifera	-	Candida taylori
31	A. mellifera	-	Candida taylori
32	A. mellifera	-	*
33	A. mellifera	-	*

Note: * Yeast strains not identified for used method.

). These species included Rhodoturola (n = 5), Kluyveromyces marxianus (n = 1), Kluyveromyces lactis var. lactis (n = 1), Candida auringiensis (n = 2), Candida taylori (n = 3), and Saccharomycopsis fibuligera (n = 1).

The selection of ß-glycosidase producers was conducted using esculin as a substrate. Only 9% of strains (n = 3), namely strains 19, 20, and 21, produced this enzyme extracellularly. Esculin hydrolysis led to the formation of glucose and esculetin, with the latter reacting with ferric ions, resulting in dark or grayish-black coloration [6]. Notably, yeast strains from the Upper Turi Region, Maranhão, exhibited low ß-glycosidase enzyme production and fermentative capacity.

Principal Component Analysis (PCA) was applied to the dataset, considering the fermentative capacities of the yeast isolates across a range of carbohydrates (glucose, fructose, galactose, sucrose, raffinose, lactose, and maltose), β-glucosidase enzyme production, and their tolerance to different stress conditions (pH, ethanol, and temperature). This analysis aimed to identify underlying patterns and classify the yeast strains based on these multifaceted characteristics. The first two principal components (PC1 and PC2) accounted for 34.63% and 20.80% of the total variance, respectively, providing a significant representation of the data’s variability (Figure 1).

The analysis revealed that PC1 is strongly influenced by the fermentative capacity of the yeast strains across several carbohydrates, with high positive loadings observed for fructose (0.411), glucose (0.405), and maltose (0.400), and to a lesser extent for raffinose (0.284) and lactose (0.277) (Figure S2, Supplementary analysis). This indicates that strains with higher positive values along PC1 demonstrate a greater overall ability to ferment these specific carbohydrates. PC2, on the other hand, is primarily driven by the tolerance of the yeast isolates to a broad spectrum of pH conditions, as evidenced by strong positive loadings for pH 2.5 (0.484), pH 4.5 (0.556), and pH 7.5 (0.357). Strains exhibiting higher positive values on PC2 display a greater capacity for tolerance in acidic, moderately acidic, and neutral to slightly alkaline environments.

The dendrogram of Figure 2, constructed using Ward’s method and Euclidean distance, visually corroborated the groupings suggested by the PCA while also revealing similarities between isolates. Based on PCA and corroborated by dendrogram analysis, four distinct groups of yeast isolates were identified based on their functional characteristics: (i) Group 1 (pH/ethanol tolerant): Strains 11, 13, 15, 16, 30, 31, and 33, located towards positive PC2, exhibited high tolerance to a wide pH range and varying ethanol concentrations, with diverse carbohydrate fermentation profiles; (ii) Group 2 (Efficient carbohydrate fermenters): Strains 12, 18, 20, 25, and 28, located towards negative PC2 and positive PC1, demonstrated strong fermentative capacity of various carbohydrates and formed a clear cluster in the dendrogram, suggesting lower tolerance to acidic pH; (iii) Group 3 (Diverse characteristics): Strains 1, 2, 3, 5, 9, 19, 21, 22, 23, 26, and 27 displayed variable PC1/PC2 values and growth at elevated temperatures, with a more dispersed pattern and smaller subgroups in the dendrogram, indicating a tolerance characteristic to different pH; (iv) Group 4 (Limited carbohydrate fermentation and pH tolerance): Strains 6, 14, 21, 22, and 32, located towards negative PC1, showed limited fermentation of tested carbohydrates but exhibited some tolerance at different pH.

Although strain 11 does not produce the β-glucosidase enzyme, its position in the PCA plot (positive PC2, moderate PC1) and its location within the pH-tolerant group in the dendrogram highlight its strong tolerance to a wide pH range and a moderate capacity for carbohydrate fermentation. This combination of stress tolerance and fermentative potential across various carbohydrates under diverse conditions positions strain 11 as a promising candidate for further in-depth studies.

3.3. Mead Production with ß-Glucosidase-Producing Yeast

For mead production, the Saccharomycopsis fibuligera strain isolated from honey in the Upper Turi region was selected due to its ability to produce β-glucosidase and its favorable fermentation characteristics, including efficient utilization of glucose, sucrose, and maltose within 24 h, as well as its tolerance to ethanol concentrations of up to 5%, which are also desirable traits for mead production. β-glucosidase enzymes from yeasts play a crucial role in releasing volatile aroma compounds by hydrolyzing glycosidic bonds in various fermentations, including wine production from cacao and grapes and the processing of beer must [7,15,27]. Since honey also contains glycosylated compounds [28], the application of yeasts with β-glucosidase activity during honey fermentation is likely to liberate similar aromatic compounds, significantly influencing the aroma and flavor of mead.

The mead was prepared using must with optimal physicochemical properties for fermentation, including an average °Brix of 27.2, a density of 1.752, a pH of 3.3, and a temperature of 24.8 °C. During fermentation, the Saccharomycopsis fibuligera strain has low potential for fermentation of the sugar honey, with °Brix varying from 27.2 to 26.2, corresponding to only 3.7% of the initial °Brix after 144 h of fermentation. In contrast, S. cerevisiae has a higher capacity for fermentation of sugar honey, with °Brix varying from 27.2 to 11.4°Brix, corresponding to only 58.1% of the initial °Brix in the same fermentation time.

Comparing volatile compounds produced during alcoholic fermentation by Saccharomycopsis fibuligera and S. cerevisiae (

Table 3. Volatile compounds of mead produced by Saccharomycopsis fibuligera and S. cerevisiae.

Volatile Compounds	Flavors	A/H
		Saccharomycopsis fibuligera	S. cerevisiae
Alcohols (n = 4)
Ethanol	-	-	23.54
Isobutyl alcohol	Sweet	11.52	10.55
3-Methyl-1 butanol (isoamyl alcohol)	Fruity and whisky	10.58	9.55
2-Phenyl ethanol	Sweet and floral	-	2.90
Esters (n = 7)
Ethyl acetate	Fruity sweet ether	-	6.19
Ethyl propionate	Fruity	10.40	8.93
Ethyl isobutyrate	Fruity	-	4.08
Ethyl butanoate	Apple, Pineapple, and Blue cheese	-	5.13
Ethyl hexanoate	Fermented vanilla and fruity	2.62	-
Ethyl octanoate	Apple Fruity	2.23	2.22
Ethyl decanoate	Pineapple Fruity	2.44	3.89
Organic acid (n = 1)
Acetic acid	Acid	1.17	5.01
Nitrogen compounds (n = 3)
1,2-Propanediamine	Ammoniacal odor	-	7.54
Dimethylamine	-	10.14	-
Methyl hydrazine	-	26.55	-

), we found significant differences. The Saccharomycopsis fibuligera strain exhibited limited fermentation of honey and must and did not produce ethanol, unlike the commercial S. cerevisiae strain.

The mead produced with Saccharomycopsis fibuligera featured nine volatile compounds, including nitrogen compounds like methyl hydrazine and dimethylamine, along with substances imparting fruity aromas, such as isobutyl alcohol, isoamyl alcohol, ethyl propionate, octanoate, decanoate, and ethyl hexanoate.

In contrast, the mead produced with commercial S. cerevisiae featured twelve volatile compounds, with ethanol being the predominant component, followed by isobutyl alcohol, isoamyl alcohol, and ethyl propionate. Additionally, this sample contained esters like ethyl acetate, ethyl isobutyrate, ethyl butanoate, and phenyl ethyl alcohol, contributing to fruity aromas. Nitrogen compound 1,2-propane diamine was also identified.

Our study quantified fifteen volatile compounds in mead produced from honey in the Upper Turi region, Maranhão State, Brazil. These compounds include alcohol, esters, nitrogen compounds, and organic acids, with ten of them contributing to fruity, sweet, or floral notes in the alcoholic beverage.

Acetic acid was detected in both mead samples, but its concentration was five times higher in the mead fermented exclusively with S. cerevisiae. The production of acetic acid is a result of yeast metabolism and, when present in significant concentrations, can negatively affect the quality of mead beverages. Furthermore, it is noteworthy that the ester ethyl hexanoate was detected only in the latter sample, while other esters were found in higher proportions in the mead produced with commercial S. cerevisiae [29].

Figure 3 displays the data on mead from fermentation with Saccharomycopsis fibuligera and the mead from fermentation with blend yeasts (commercial S. cerevisiae and Saccharomycopsis fibuligera). The mead fermented exclusively with Saccharomycopsis fibuligera displayed lower electrical conductivity, ranging from 867 to 511 µS/cm. However, in the mead fermented with both S. cerevisiae and Saccharomycopsis fibuligera, electrical conductivity increased substantially, ranging from 867 to 1115 µS/cm.

During the fermentation of mead using the Saccharomycopsis fibuligera strain, as well as the combination of S. cerevisiae and Saccharomycopsis fibuligera, the dissolved oxygen content decreased considerably in the first 24 h, from 94.6% to 27.2% and 19.2%, respectively. At the end of the fermentative process, the ODs were 44.6 and 63.4, with a reduction of 52.85% and 33.0% for S. cerevisiae and Saccharomycopsis fibuligera. This rapid reduction indicates good adaptation of the yeast, as the strains efficiently consumed oxygen during the initial phase of fermentation.

4. Discussion

4.1. Quantification of Molds and Yeasts in Honey Samples

Several studies have indicated that honey produced and collected immediately after production exhibits low yeast contamination rates [30]. In contrast, honey obtained through informal trade, without registration or inspection, shows elevated levels of yeast contamination [31]. The higher prevalence of molds and yeasts in Apis mellifera honey may originate from the bees’ own intestinal microbiota or be introduced through floral nectar [32].

Ribeiro et al. [33] conducted a study in the Alto Turi region, analyzing the pollen composition of 130 geopropolis samples collected by Melipona bees, including M. subnitida, M. seminigra, M. flavolineata, and M. fasciculata. A total of 148 pollen types, distributed among 49 plant families and 108 genera, were identified, including Attalea speciosa, Anacardium, Borreria verticilatta, and others, reflecting the pollen profile from Maranhense Amazon, in the municipality of Santa Luzia do Parua.

A key physicochemical characteristic of stingless bee honey is its higher moisture content compared to honey from Africanized bees, making it more susceptible to microbial growth, particularly yeasts that are part of its native microbiota. Consequently, this honey can undergo natural fermentation due to the presence of osmophilic yeasts, classified as ‘matured honey’, which naturally ferments after extraction [34,35].

4.2. Selection of ß-Glucosidase-Producing Yeasts

Yeast species play a predominant role in the fermentation process, producing essential components such as ethanol, carbon dioxide, and other secondary compounds that significantly influence the flavor, aroma, and quality of alcoholic beverages. Despite this, the isolation and selection of native yeast species from honey, particularly those capable of producing ß-glucosidases for alcoholic fermentation, have been relatively unexplored in the literature. New microbial isolates, especially yeast strains, are continually sought after to meet the biochemical requirements and withstand the stress conditions encountered during environmental fermentations [36,37].

Resistance to stress conditions is a critical consideration because factors such as temperature, ethanol concentrations, and pH levels can impact not only the kinetics of fermentation but also the overall yeast metabolism. In our study, all yeast strains isolated exhibited varying degrees of tolerance to specific stressors. Notably, in our study, all isolated yeast strains exhibited growth at both 37 °C and 45 °C, indicating tolerance to these temperatures. However, the ability to tolerate ethanol concentrations of 10% and 15% (v/v) and pH 2.5 varied among the strains. The least stressful conditions were observed at 5% ethanol and pH levels between 4.5 and 7.5. Similar findings were reported by Silva et al. [38], where temperature was identified as the most limiting condition, followed by ethanol concentration, with optimal conditions found at 6% ethanol and 30 °C.

The Saccharomycopsis fibuligera is known for its enzyme production capabilities, such as amylases, and notably prominent levels of ß-glucosidase [39]. The production of ß-glucosidase is intricately linked to ethanol production, as demonstrated by Tang et al. [40].

The identified yeast species are commonly associated with honeybees or are found in their habitats. Carvalho et al. [41] analyzed 200 samples of Portuguese honey and isolated a limited number of yeast strains (n = 24), including species such as Rhodotorula mucilaginosa, Candida magnoliae, and Zygosaccharomyces mellis.

Da Fonseca Meireles et al. [42] obtained 532 yeast isolates associated with the nests of Melipona interrupta and Cephalotrigona femorata bees. They identified 15 yeast strains through molecular identification methods, including species like Candida sp., Hyphopichia sp., Rhodotorula sp., Wickerhamiella versalitis, Debaryomyces sp., Candida orthopsilosis, Candida apicola, Metschnikowia sp., Zygosaccharomyces siamensis, Hanseniaspora opuntiae, Pichia sp., Pichia kluyveri, Trichosporon asahii, Kodamaea ohmeri, and Aureobasidium sp.

The biodiversity of cultivable yeasts found in honey and pollen reflects the intricate ecological relationships within a beehive, as a recent study demonstrates significant sharing in yeast species between honey and pollen from the same beehive. Nine yeast species were identified in the pollen of Nannotrigona testaceicornes and Tetragonisca angustula, including Candida maltosa, Candida norvegica, Kazachstania telluris, Schizosaccharomyces pombe, Scheffersomyces insectosus, Meyerozyma guilliermondii, Brettanomyces bruxellensis, Kazachstania exigua, and Starmerella lactis-condensi. Similarly, seven of these species (Kazachstania telluris, Schizosaccharomyces pombe, Scheffersomyces insectosus, Meyerozyma guilliermondii, Brettanomyces bruxellensis, Kazachstania exigua, and Starmerella lactis-condensi) were found in honey from three species of stingless bees [43].

This rich diversity of yeasts is likely influenced by the vast floral biodiversity of the Amazon region, where the unique composition of pollen and nectar from various plant species provides diverse ecological niches for yeast colonization. In the municipality of Santa Luzia do Paruá, pollen types such as Mimosa pudica, Borreria, and an unidentified species from the Arecaceae family were reported in honey samples collected during the rainy season, while Hyptis and an unidentified Asteraceae species were present during the harvest season [44].

The diversity of yeasts found in honey, including non-Saccharomyces species, reveals the untapped potential of these microorganisms for mead fermentation. Saksinchai et al. [45] identified a wide variety of yeasts in honeys from different bee species (Apis and stingless bees), including genera such as Candida and Starmerella, with genetic variations among isolates from different sources. The presence of Starmerella meliponinorum in various types of honey and its ability to metabolize sugars suggest its active role in the honey environment.

Silva et al. [46] also isolated yeasts from honey and pollen of stingless bees and Apis mellifera, identifying species such as S. cerevisiae, known for high ethanol production and stress tolerance, as well as S. meliponinorum. These findings highlight that both honeybees and stingless bees harbor a diverse yeast microbiota with potential for biotechnological applications, including the fermentation of beverages such as mead. However, no studies have identified yeast species in honey from bees in the Upper Turi region of Maranhão, located in the southwestern Amazon.

4.3. Mead Production with ß-Glucosidase-Producing Yeast

Fermentation of honey from the Alto Turi region using Saccharomycopsis fibuligera, selected for its β-glucosidase production and ability to metabolize honey sugars, resulted in low fermentative efficiency, with minimal sugar conversion and no ethanol production—contrasting with the high fermentative performance observed with Saccharomyces cerevisiae. Volatile compound analysis revealed a distinct profile for the mead produced with S. fibuligera, characterized by the presence of nine compounds, including fruity esters such as ethyl propionate, ethyl octanoate, ethyl decanoate, and the exclusive ethyl hexanoate, as well as nitrogen-containing compounds. In comparison, the mead fermented with S. cerevisiae contained twelve volatile compounds, with ethanol as the predominant component, along with other fruity esters.

The production of fermented beverages with complex and distinctive sensory profiles is a central goal in enological research and other fermentation fields, such as mead production. The use of non-Saccharomyces yeasts has proven to be a promising strategy for achieving such diversity, as demonstrated by Bi et al. [11] in pomegranate-based fermentations. In that study, fermentation with different non-Saccharomyces yeast strains resulted in volatile compound profiles significantly different from those produced by Saccharomyces cerevisiae, with some strains promoting increased production of ethyl and acetate esters associated with fruity and floral notes.

Similarly, the yeast Saccharomycopsis fibuligera used in the mead fermentation in our study also influenced the volatile compound profile, although it showed a lower ethanol yield compared to S. cerevisiae. The presence of fruity esters such as ethyl propionate, ethyl octanoate, and ethyl decanoate, observed in the mead inoculated with S. fibuligera, suggests a contribution of this yeast to the beverage’s aroma. Comparable results were reported by Pereira et al. [47] in mead produced from multifloral honey, where esters such as isoamyl acetate, ethyl hexanoate, and ethyl octanoate were identified as the main contributors to fruity and floral aromas.

This ability to modulate the aromatic profile is often associated with the activity of the β-glucosidase enzyme, which can release volatile compounds from glycosylated precursors present in the raw material. Thus, the use of β-glucosidase-producing non-Saccharomyces yeasts, such as S. fibuligera in our study, represents a valuable tool for directing the aromatic composition of fermented beverages, even if this may be accompanied by a lower ethanol yield, as observed in our case.

The activity of the enzyme β-glucosidase emerges as a key factor in diversifying the aromatic profile of fermented beverages, with consistent evidence from studies on wine [7], cocoa fermentation [15], and craft beer [27]. This enzyme catalyzes the hydrolysis of glycosidic bonds found in aroma precursors, releasing volatile aglycones that enrich the final flavor. In mead, honey is the primary source of glycosides, making the β-glucosidase activity of fermentative yeasts a potential contributor to its aroma.

The study by Han et al. [27] demonstrated a significant impact of non-Saccharomyces yeasts with high β-glucosidase activity on the increase in floral and fruity terpenes in beer, highlighting the potential of this enzyme to generate unique sensory profiles. Similarly, Saccharomycopsis fibuligera, isolated in our study and selected for its β-glucosidase production, may have contributed to the aromatic profile of mead through the release of volatile compounds from glycosylated precursors present in honey.

A detailed analysis of the volatile compounds identified in the mead fermented with this strain revealed an increase in esters and other aromatic compounds, which may provide direct evidence of the role of β-glucosidase in flavor formation (Figure 4). Comparing these findings in the literature may position mead produced with selected yeasts as a beverage with distinct and potentially innovative aromatic characteristics.

Volatile amines detected in fermented beverages, such as wine, often originate from the microbial decarboxylation of amino acids, likely due to their use as nitrogen sources by yeast. Notably, dimethylamine and methyl hydrazine were produced exclusively by the Saccharomycopsis fibuligera strain. It is worth mentioning that low concentrations (below 50 μg/L) of dimethylamine are unlikely to negatively impact the wine’s aroma, given the higher content of ethanol and other volatile compounds, such as esters, terpenes, and higher alcohols [40].

The use of electrical conductivity for monitoring alcoholic fermentation processes and determining their endpoints is a cost-effective and straightforward approach. A study by Li et al. [48] demonstrated that increased ionic and ethanol concentrations during alcoholic fermentation led to higher electronic conductivity. In our study, it became evident that conductivity increased with rising pH levels, indicating that as the fermentation process and ethanol production progressed, the production of organic acids also occurred.

The amount of dissolved oxygen required to ensure good growth is determined by the strain of yeast used in the fermentation process [49]. The growth of facultative anaerobic yeasts, such as S. cerevisiae, requires molecular oxygen in the initial stage of fermentation for the synthesis of sterols and unsaturated fatty acids, which can improve the fermentation process by increasing cell viability, increasing the fermentation rate, and shortening the total fermentation time [50,51]. However, Saccharomyces cerevisiae is a facultative anaerobic and fermentative organism, meaning it can grow on sugars without needing an alternative respiratory pathway when the classical respiratory chain is inhibited [51].

5. Conclusions

A low number of honey samples from the Alto Turi Region (n = 14) from Maranhão showed the presence of yeast. Four yeast isolates were able to produce β-glycosidase, and the strain 20 (Saccharomycopsis fibuliger) during biochemical tests demonstrated the fermentative capacity from glucose, sucrose, maltose, and galactose and grew at pH 7.5 and 5% ethanol, being selected for mead production. This yeast strain demonstrated potential for mead production by generating a distinct volatile profile, notably including the fruity ester ethyl hexanoate and limiting acetic acid accumulation. In our study, the application of this yeast in the fermentative process suggests that fruity characteristics can be enhanced from fermentation; however, more studies are necessary to associate co-culture together with yeast-produced ethanol.