Introduction

Tej is an Ethiopian alcoholic beverage fermented commonly at the household level through traditional methods without inoculating defined starter cultures1,2. It is produced from honey, water and Gesho (Rhamnus prinoides) with some flavoring agents like Weira (Olea europaea) and Grawa (Vernonia amygdalina) for conditioning the fermentation vat3,4. Sugar is added in place of honey in some cases which necessitates the addition of coloring agents to make the final product look yellow5. The final product differs between different producers, but has effervescent nature2,6. The alcohol content of Tej is higher than other non-distilled Ethiopian fermented alcoholic beverages7. The quality of Tej varies depending on the raw materials used by the producers8. Reports show the presence of toxic metabolites such as high concentrations of methanol and fusel oils in the traditionally fermented Tej5. The microorganisms for Tej fermentation mainly come from the ingredients and equipment used for its preparation which varies between producers9. To produce products with consistent organoleptic quality and microbiological stability, starter culture development and optimization of the fermentation process were recommended10.

Starter cultures are microbial preparation of large numbers of cells of at least one type of microorganism to be added to a raw material to produce a fermented food11. They have been successfully applied to a wide verity of African fermented foods and drinks12. The selection of the right types of starter cultures is important to obtain products with the desired sensory attributes. Starter microorganisms should be resistant to low pH, high ethanol, high salt concentration, high temperature and be able to utilize various carbon sources10,11,12,13,14,15,16. Production of biotechnologically important products such as diacetyl and exopolysaccharides are also considered desirable17. Previous researches reported that LAB are mostly involved in the production of Tej18,19,20 and the yeasts of the genus Saccharomyces were responsible for the conversion of sugars to ethanol in Tej1.

Despite the huge economic values for venders which are usually females, and variability in the production processes of Tej from one locality to another, there is limited information about the development of starter culture for this valuable product. There are a few studies regarding the use of mixed starter cultures for production of Tej. A previous study2 used different types of lactic acid bacteria in combination with S. cerevisiae and found promising results for Tej produced by a combination of different lactic acid bacteria with S. cerevisiae. Similarly, our previous report21, for samples collected from different locality showed the potential of combined starter culture of Lactobacillus and Saccharomyces for production of Tej. The findings of the previous researches also indicated that, the potential starters for Tej production could differ depending on the Tej sources. The aim of this study was therefore to develop potential mixed lactic acid bacteria and yeast starter cultures for controlled production of Tej.

Materials and methods

Sampling site, sample collection and processing

Three cities (Debre Markos, DM; Finote Selam, FS and Bure, BU) in the Amhara region, Northern Ethiopia, were selected using lottery method from cities with potential for Tej and honey production. A total of 30 Tej samples, 10 samples from each city, were purposively collected from the best Tej producing and vending houses. Each sample was collected from the local venders with pre-sterilized containers, and the samples were transported to the Microbiology Laboratory, College of Natural and Computational Sciences, Addis Ababa University using Ice Box with a temperature of 4 °C. Then, isolation and enumerations of microorganisms in the Tej samples were done using the appropriate culture medium and culture conditions as described below.

Isolation and enumeration of the LAB and yeast isolates

LAB and yeast count of the Tej samples were analyzed according to the methods reported in Bahiru et al.22. Briefly, 25 mL of each sample was homogenized in 225 mL of sterile peptone water, followed by tenfold serial dilution (10−2 to 10−7) and plating 0.1 mL of the serially diluted sample on pre-sterilized agar media. The LAB was enumerated on pre-sterilized MRS (de Man, Rogosa, and Sharpe) agar after incubation at 32 °C for 24–48 h under anaerobic conditions using anaerobic jar. The yeasts were enumerated on yeast extract peptone dextrose (YPD) agar after incubation at 28 °C for 72 h.

Characterization of the LAB and yeasts

Cultural, morphological characteristics, catalase enzyme, gas and diacetyl production were done both for LAB and yeasts. However, Gram reaction (KOH test) was done only for LAB, and extracellular starch- like (amyloid) compound detection was done only for yeasts. Cultural (colony color, margin and elevation) and morphological (shape and arrangement of cells) were assessed according to the method described by23,24. Production of catalase was evaluated following the previously reported method25. Gas production was done according to the method of26. Production of diacetyl by the LAB and yeasts was done according to the method of27, exopolysaccharide production for the LAB was evaluated according to the method of28 and the production of extracellular starch-like (amyloid) compound was detected according to the method of29. The KOH test was done following the method of30.

Fermentation of carbohydrates

Four milliliters aliquot of the nitrogen base medium (4.5 g yeast extract, 7.5 g peptone, 0.5 g of bromothymol blue in 1000 mL distilled water) was dispensed in test tubes containing inverted Durham’s tube. After filtration, 1 mL aliquot of the sugar solutions was aseptically added to the test tubes containing the nitrogen base. A loop full of young cultures were inoculated into the tubes and incubated at 30 °C for 72 h. A blank consisting of inoculated basal medium devoid of any carbon source served as a control. The accumulation of gas in the Durham’s tubes and the color change of the indicator were recorded as positive results for carbohydrates fermentation31. The carbohydrates evaluated were glucose, fructose, galactose, lactose, sucrose, maltose, mannose, starch, trehalose and xylose.

Assimilation of carbon sources

Assimilation of carbon source was determined by the auxanographic carbohydrate assimilation method29 with slight modification, about 2 mL from 10% stock solution of carbon sources (mannose, lactose, trehalose, starch, maltose, fructose, sucrose, glucose, galactose, xylose and glycerol) were placed on the dried agar surface and inoculated with 10 µL of 48 h young yeast cultures. The Petri dishes were incubated at 30 °C for 10 days and examined for growth.

Physiological stress tests for LAB and yeast

The physiological stress tests were assessed for LAB and yeast isolates on MRS and YPD broths, respectively. Fresh overnight cell cultures were centrifuged (3000 rpm for 15 min at 24 °C) and washed twice with 0.1% w/v peptone water (pH 7). The cell cultures (OD600 of 0.2) in peptone water corresponding to ~ 107 cell/mL of yeast and LAB were centrifuged and re-suspended in every stress experiment and each experiment was conducted in triplicate.

pH tolerance

Tolerance of the yeasts and the LAB to low pH was tested using the method of32. One milliliter of the 24 h old yeast and LAB cultures centrifuged, washed and re-suspended in peptone water were separately inoculated into test tubes containing 5 mL of YPD and MRS broth adjusted to pH 2.5, 3.5 and 4.5 using 1N HCl and NaOH, respectively. The test tubes were incubated for 24 h at 32 °C, and optical density was measured at 600 nm using UV–visible spectrophotometer (Jenway, 6405, Felsted, Dunmow, UK).

Salt tolerance

Tolerance of the yeasts and the LAB to NaCl was tested using the method of33. One milliliter of the 24 h old yeast and LAB cultures centrifuged, washed and re-suspended in peptone water were separately inoculated into test tubes containing 5 mL of YPD and MRS broth supplemented with 5.5, 7.5 and 10% (w/v) NaCl, respectively. The test tubes were incubated for 24 h at 32 °C and optical density was measured at 600 nm using UV–visible spectrophotometer (Jenway, 6405, Felsted, Dunmow, UK).

Ethanol tolerance

Tolerance of the yeast and the LAB to ethanol was tested using the methods of33,34. One milliliter of the 24 h old yeast and LAB cultures centrifuged, washed and re-suspended in peptone water were separately inoculated into test tubes containing 5 mL of YPD and MRS adjusted to 10, 15 and 20% ethanol, respectively. The test tubes were incubated for 24 h at 32 °C and optical density was measured at 600 nm using UV–visible spectrophotometer (Jenway, 6405, Felsted, Dunmow, UK).

Temperature tolerance

Temperature tolerance of the yeasts and the LAB was tested using the method of35. One milliliter of the 24 h old yeast and LAB cultures that were centrifuged, washed and re-suspended in peptone water were separately inoculated into test tubes containing 5 mL of YPD and MRS, respectively. The tubes were then incubated at 35 °C, 40 °C and 45 °C for 24 h, and optical density was measured at 600 nm using UV–visible spectrophotometer (Jenway, 6405, Felsted, Dunmow, UK).

Molecular identification of the LAB and yeast isolates

Molecular identification of LAB

Extraction of the total DNA of the LAB was conducted following the method described in36. For identification of the LAB isolates, the full length 16S rRNA gene was amplified using the primer pair 27 F: 5’-AGA GTT TGA TCC TGG CTC AG-3′ and 1492 R: 5’ GGT TAC CTT GTT ACG ACT T-3. The PCR reaction parameters were: 94 °C for 10 min followed by 32 cycles at 94 °C for 30 s, 55 °C for 20 s, 72 °C for 55 s, and final extension at 72 °C for 5 min. The total reaction mixture per tube was 25 μL. The amplicons were sequenced using Sanger sequencing using the 27 F and 1492 R primers using the ABI PRISM® BigDye™ Terminator cycle sequencing kit. The ABI PRISM Sequencer platform 3730XL at the Beijing Sangon Biotechnology Company, China. The top match for the 16S rRNA sequence of the bacteria was searched against the GenBank database using the Basic Local Alignment Search Tool (BLAST). Sequence identity with a threshold of 97% or above was used to describe the bacterial species using the rRNA gene region37,38,39.

Molecular identification of the yeasts

The yeast isolates were grown overnight for 24 h at 30 °C on YPD medium and DNA extraction was performed following the method40. Primers NL1: 5’-GCA TAT CAA TAA GCG GAG GAA AAG-3’ and NL4: 5’-GGT CCG TGT TTC AAG ACG G-3’ were used to amplify the large sub unit (LSU) rDNA D1/D2 domain. The PCR conditions were: an initial denaturation at 98 °C for 2 min followed by 35 cycles at 98 °C for 10 s, annealing at 55 °C for 15 s, 72 °C for 15, and final extension at 72 °C for 5 min. The total volume of the reaction mixture per tube was 25 μL, and amplicon sequencing was performed methods described41, with minor modifications. Cycle sequencing was performed using the ABI PRISM® BigDye™ Terminator cycle sequencing kit. An ABI PRISM Sequencer platform 3730XL at the Beijing Sangon Biotechnology Company, China. The top match for the sequence of the yeasts in the current study was searched against the GenBank database using the Basic Local Alignment Search Tool (BLAST). Sequence identity with a threshold of 99% or above was used to describe the yeast species using the rRNA gene region37,38,39.

Starter culture development

Compatibility test

The candidate isolates’ inter-compatibility was evaluated by cross streaking each isolate against each other on YPD (for yeast) and MRS (for LAB) agar plates, and incubating at 32 °C for 48 h42. Isolates considered compatible were able to grow on the same substrate without an inhibition zone when cross streaked against each other.

Starter culture formulation

The combinations of the LAB and yeasts used for production of Tej were determined using Design Expert version 13.0.05, 2021 Statistical software. Accordingly, 5% viable cells of the pure isolate were mixed in a flask containing 25 mL of pasteurized honey (33.3%) dissolved in distilled water to determine the suitable combination to produce Tej43.

Propagation of the isolates

Pure cultures of LAB and yeasts were separately inoculated into MRS and YPD broth media, respectively. The LAB were incubated at 32 °C for 48 h and the yeasts were incubated at 30 °C for 72 h. Cells of the LAB and yeast were harvested from the media by centrifugation at 3000 rpm for 15 min. The viable counts of LAB (109 CFU/mL) and yeasts (107 CFU/mL) were determined using spread plate technique. The pellets (2 mL) were rinsed with peptone water and inoculated into 100 mL pasteurized honey in a 250 mL flask, and the mixture was incubated for 48 h at 32 °C44.

Laboratory preparation of Tej using starter cultures

For laboratory preparation of Tej using the starter cultures, 250 g of honey pasteurized at 62.5 °C for 15 min was mixed with 500 mL of sterile distilled water at a ratio of 1:2 in 1000 mL flasks. The mixture was inoculated with 5% of LAB and yeast suspension, and the flasks were tightly covered with aluminum sheet. After 3 days of fermentation, 10 g of sterile Rhamnus prenoides (Gesho) was added to the must and the flasks were covered again with sterile aluminum sheet and fermented continuously for 15 days at ambient temperature. After the fermentation period the prepared Tej was filtered through sterile muslin cloth to remove the sediments and the Rhamnus prenoides.

Sensory analysis

Panelists consisting of Tej producers and consumers (n = 10) were selected to determine the sensory acceptability of the product. Panelists ranked acceptability of various attributes of different Tej produced under laboratory conditions using 5-point hedonic scale.

Data analysis

The data from the current study were analyzed using SPSS Software version 25.0.0, 2021. One way ANOVA was used to compare different means and mean comparison was done using Tukey’s HSD test. Statistical significance was determined at p < 0.05. The possible combinations of the isolates used for production of Tej under the laboratory conditions were identified using Design Expert software version 7.0.0 (2005).

Ethical approval

Ethical clearance was obtained from the Institutional Review Board of the College of Natural and Computational Sciences, Addis Ababa University with Ref No. SF/MCMB/345/14/2. After the approval of the proposal, informed consent was obtained from the panelists. The panelists were informed of the right to withdraw at any time if not comfortable and their participation was completely confidential. The ethical issue was conducted in accordance with the declaration of Helsinki and its amendments.

Results and discussion

Isolation, enumeration and characterization of the LAB and yeasts

LAB and yeasts were enumerated form Tej samples collected from the three study sites, and the mean counts (log CFU/mL) of both yeasts and LAB were significantly higher (p < 0.05) for the Tej samples collected from Finote Selam compared to the other two study sites (Supplementary Table 1). Other researchers also reported similar counts for LAB and yeasts22,45.

A total of 359 isolates of yeast (193) and LAB (166) were isolated from traditionally fermented Tej samples from the three sampling sites. The number of the isolates were reduced based on physiological stress tolerance tests, and four LAB and four yeasts which were compatible were used for production of Tej under the laboratory conditions. The details of their morphological, cultural and biochemical characteristics were summarized in Supplementary Tables 2 and 3, and identified using molecular techniques, respectively. The total number of each strain isolated was n = 1 for Lactiplantibacillus pentosus, n = 2 for Lacticaseibacillus paracasei and n = 1 for Lentilactobacillus hilgardii (Supplementary Fig. 1). All the yeast isolates used in the fermentation (n = 4) were Saccharomyces cerevisiae (Supplementary Fig. 2). The sequences were deposited in the NCBI database with the accession number provided in the bracket after the species name; Lactiplantibacillus pentosus (L2) (PV423076), Lacticaseibacillus paracasei (L6) (PV423077), Lacticaseibacillus paracasei (L7) (PV423078), Lentilactobacillus hilgardii (L8) (PV423079), and Saccharomyces cerevisiae (Y2) (PQ410281), Saccharomyces cerevisiae (Y4) (PQ410276), Saccharomyces cerevisiae (Y5) (PQ410274) and Saccharomyces cerevisiae (Y6) (PV416743). The predominance of Saccharomyces in the Tej samples of the current study was also in agreement with the report1. Various studies reported the presence of different types of LABs in Tej2,22.

Both the LAB and yeasts were able to produce diacetyl but only the LAB produced exopolysaccharides. Other researchers also reported the production of diacetyl by Lactobacillus species46. In yeasts, the concentration of diacetyl is linked to the degree of growth and the level of degradation of fermentable sugars47.

Carbon source utilization

All the yeast isolates were able to ferment glucose, and assimilate both glucose and fructose. Mannose, lactose and starch were neither assimilated nor fermented by the isolates. None of the yeast isolates utilized lactose, mannose and starch (Table 1), a similar study was also reported48. The differences in carbon source utilization depend on the presence of specific enzymes and sugar preferences by the yeasts.

Table 1 Carbohydrate utilization of the yeast isolates.
Full size table

Physiological stress tests for LAB and yeasts

pH tolerance

The LAB and yeasts were able to tolerate all the pH values tested. Significantly higher tolerances (4.6 ± 0.20) and (6.0 ± 0.14) were recorded, respectively at pH 3.5 and pH 4.5 for Lacticaseibacillus paracasei (L7) (Table 2). Other researchers also reported low pH tolerances of LAB and yeasts49,50,51,52,53. Microorganisms are reported to have acid tolerance mechanisms such as acid shock proteins, and the condition is strain specific54

Table 2 pH tolerance of the LAB and Yeast isolates.
Full size table

Ethanol tolerance

Ethanol tolerances of the LAB and yeasts decreased as the concentrations increased. The highest tolerances of 0.86 ± 0.01 and 0.56 ± 0.03 were recorded respectively for Lacticaseibacillus paracasei (L7) and the yeast Saccharomyces cerevisiae (Y4) at 10% alcohol concentration (Table 3). High ethanol tolerances of the LAB55 and yeasts15,56 were also reported. Tolerance to ethanol depends on having specific tolerance mechanisms, and is conserved within species55. High ethanol tolerances in Lacticaseibacillus paracasei has been linked the SMN-LBK gene family57. In Saccharomyces cerevisiae, ethanol tolerance was associated with gene expression responses, pathway-based mechanisms and regulatory networks58.

Table 3 Ethanol tolerance of the LAB and yeast isolates.
Full size table

Salt tolerance

Salt tolerances of the LAB and yeasts decreased with increase in the salt concentration from 5.5 to 10%. The highest salt tolerances were recorded for Lacticaseibacillus paracasei (L6) and Saccharomyces cerevisiae (Y6) at 5.5% salt concentration (Table 4). This is not in line with the report by48. However, halophilic LAB were reported to be widely used as starter cultures and were the dominant microorganisms in many fermented foods59,60. The presence of highly conserved pathways that mediate salt tolerance was reported for Saccharomyces cerevisiae61. The differences in salt tolerance could be attributed to having mechanisms such as Na+ and K+ fluxes, modulation of cell wall properties and synthesis of compatible solutes33.

Table 4 Salt tolerance of the LAB and yeast isolates.
Full size table

Temperature tolerance

A general decreasing trend in tolerance of the LAB and yeasts was observed. Best tolerance was recorded at 35 °C for both LAB (L8) and yeast (Y2) (Table 5). Similar pattern of tolerance was recorded for LAB62. Strain dependence of temperature tolerance in LAB was reported by63. Significant growth of Saccharomyces cerevisiae at 42 °C was reported by64. But, a report by48 showed the complete absence of yeast growth at 40 °C. The differences in temperature tolerances of yeasts, and LAB could be attributed to differences in heat shock responses65, and mutations linked to high temperature growth was associated with high temperature tolerances in Saccharomyces cerevisiae53.

Table 5 Temperature tolerances of the LAB and yeast isolates.
Full size table

Starter culture development

Compatibility of the isolates

The isolates that were able to tolerate the stresses were assessed for compatibility in cross streaking method. Four isolates from the yeast and four from the LAB were found to be compatible (Fig. 1).

Fig. 1
figure 1

Compatibility of the yeast (a) and LAB isolates (b).

Full size image

Starter culture formulation

There were eight possible combinations generated by the Design expert software (F1 to F8). But, two of the combinations contained either the LAB or the yeast and weren’t included in the laboratory production of Tej. The LAB and yeast isolates used in the formulation were: F1 (L2 and Y4); F2 (L7 and Y5); F3 (All combinations); F4 (L8 and Y2); F5 (L6 and Y6) and the control was fermented without spontaneously without the addition of starter culture (Table 6).

Table 6 Formulation of starter culture using design expert software.
Full size table

Propagation of inoculants and production of Tej

Inoculants were prepared by mixing 2 mL of washed pellets of LAB and yeast isolates into 25 g of honey diluted with 50 mL of distilled water (Fig. 2a). The Tej produced using the starter cultures before filtration (Fig. 2b) and after filtration (Fig. 2c). The produced Tej samples were yellow in color and had pleasant honey odor.

Fig. 2
figure 2figure 2

Inoculates produced by mixing honey, water and the isolates (a), Tej produced under laboratory conditions using formulates (b), the final filtered Tej (c).

Full size image

Sensory analysis

Panelists consisting of Tej producers and consumers (n = 10) were selected to determine the sensory acceptability of the products. Panelists ranked acceptability of various attributes of different Tej samples using a 5-point hedonic scale. The sensory attributes of the Tej such as color, flavor, odor, and overall sensory acceptability were assessed. The overall mean sensory acceptability of the Tej types ranged from 1.72 to 4.15 (Table 7). The difference between the best rated Tej samples (T4 and T5) and the control was not statistically significant. The second Tej sample (T2) was the least rated by the panelists.

Table 7 Sensory characteristics of Tej made with the formulated starter cultures.
Full size table

Conclusion

Yeasts and LAB were the dominant microorganisms in the Tej samples collected from the Northern part of Ethiopia. Among the total of 359 isolates, four potential isolates of LAB and four yeasts were selected for the purpose of testing their starter culture potential in Tej production based on their stress tolerance capabilities. Two of the LAB isolates were identified to be Lacticaseibacillus paracasei, one each of Lactiplantibacillus pentosus and Lentilactobacillus hilgardii. All the yeasts belonged to Saccharomyces cerevisiae. The isolates were able to pass all the hurdles tested and were able to produce Tej under controlled conditions. From the sensory analysis, Tej (T5) produced by the combination of Lacticaseibacillus paracasei (L6) and Saccharomyces cerevisiae (Y6) was the most rated compared to all other combinations of Tej produced under laboratory conditions. Thus, we are able to produce acceptable Tej using defined mixed starter cultures under the laboratory conditions which has strong implications for consumers and vendors with regard to the quality and safety of the product. Further testing of the isolates at larger scales is important to scale up Tej production.