Edited by: Jian Zhao, University of New South Wales, Australia
Reviewed by: Stefan Junne, Technical University of Berlin, Germany; Graciela Liliana Garrote, National University of La Plata, Argentina; Volkmar Passoth, Swedish University of Agricultural Sciences, Sweden
†ORCID: Sarah Köhler,
This article was submitted to Food Microbiology, a section of the journal Frontiers in Microbiology
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
A process development from a traditional grain-based fermentation to a defined water kefir fermentation using a co-culture of one lactic acid bacterium and one yeast was elaborated as a prerequisite for an industrially scalable, controllable, and reproducible process. Further, to meet a healthy lifestyle, a low ethanol-containing product was aimed for. Five microbial strains—
Traditionally fermented beverages enjoy increasing popularity against the background of a conscious and healthy lifestyle (
Generally, the so-called cross-feeding occurs when LAB and yeast are fermented in co-culture meaning that both groups of organisms can profit from the metabolic products of one another (
Besides the consortium of microorganisms, also the fruit component is rather complex. Traditionally, e.g., figs, dates, raisins, and/or lemons, were used to produce WK (
In summary, industrializing the production of WK with a complex mixture of raw materials and microorganisms is very difficult in terms of final product properties and especially reproducibility (
In order to get a deeper insight which microorganisms originating from natural WK grains contribute synergistically to the fermentation progress resulting in an organoleptic characteristic beverage, systematic experiments to reduce the microbial complexity were conducted. Furthermore, the main focus was the production of a lactic acidic beverage by co-culture fermentation, which should preferably contain a low amount of alcohol meeting the consumer demands for a healthy drink. Thereby, the influence of the available nutrients, such as amino acids and trace elements, on the microbial interactions and the sensory properties represented by different microbially built chemical compounds as well as possible limitations were considered. This paper finally presents a comparison of a traditionally produced complex WK and a fermented beverage product produced with specific starter cultures mimicking WK regarding the fermentation parameters as well as the final product properties.
In the present studies, WK grains were used. These have been used successfully for many years for the traditional production of WK via back-slopping processes. The microbial strains used as defined cultures were isolated and identified from these mentioned grains. Therefore, grains were homogenized by using the ULTRA-TURRAX® T25 basic (IKA® -Werke GmbH & CO. KG, Staufen, Germany), and different dilutions of the homogenized liquid were plated on yeast extract (YE) dextrose agar as well as MRS agar according to
Two different media, herein referred to as basis medium and modified medium, were used for the fermentation studies. The basis medium was established on a common recipe for WK production. In detail, it was composed of the following (per liter): 1 dry fig (Alesto, Lidl Dienstleistung GmbH & Co. KG, Neckarsulm, Germany), 60 g of brown cane sugar (Fairglobe, Lidl Dienstleistung GmbH & Co. KG), and 7.5 mL of lemon juice concentrate (Solevita Zitrone, Lidl Dienstleistung GmbH & Co. KG). The modified medium was composed of the following (per liter): 9 mL of fig syrup (Schoeneberger, Salus Pharma GmbH, Bruckmühl, Germany), 60 g of brown cane sugar, 7.5 mL of lemon juice concentrate, and a defined concentration of either X-SEED KAT YE or X-SEED Peptone (Ohly GmbH, Hamburg, Germany), a yeast peptone. Both yeast derivatives were of food grade quality. For media preparation, sugar, dry fig/fig syrup, and YE when applicable were dissolved in boiling tap water (80% of needed volume). After being cooled down to room temperature (20°C), lemon juice concentrate was added, and batches were filled up to desired volume with lukewarm tap water.
The strains were pre-cultured in two steps. In both steps, yeast strains were pre-cultured in 250-mL shaking flask at 130 rpm and 30°C for 72 h in YE dextrose (YED) medium containing 50 g/L of glucose (AppliChem GmbH, Darmstadt, Germany) and 10 g/L of YE (SERVABACTER®, SERVA Electrophoresis GmbH, Heidelberg, Germany), pH 5.5. LAB were pre-cultured as stand culture in closed 250-mL bottles at 30°C for 72 h in MRS medium according to
In the first part of the studies, fermentations with different combinations of two microbial strains in co-culture were investigated. Thereby, 1 ⋅ 106 yeast cells/mL and 1 ⋅ 107 LAB/mL were applied in co-culture. The ratio yeast/LAB (1:10) was derived from the respective cell concentrations in the supernatant of WK fermented by grains. For comparison, fermentations with WK grains were investigated simultaneously. Thereby, 100 g of WK grains per liter was used for the fermentation. For the investigations, the basis medium, which contained dried fruits, was used. The fermentations were conducted in 1-L glass bottles with loosely sealed caps at 22°C (±2°C) (room temperature) as stand cultures. At the start of the fermentation and before sampling, the samples were mixed by mild horizontal shaking. The sampling was executed once per day for offline analyses, which included determination of cell count, high-performance liquid chromatography (HPLC) analysis, pH measurement, and sensory analysis. After 7 days of fermentation, the products showed the best obtainable characteristics and were analyzed toward a wide range of volatile components (see section “Analyses of Volatile Components”). The fermentation studies were executed in biological duplicates. The most appropriate combination of microorganisms was then applied in further trials.
In further experiments, the components of the basis medium were adapted to make the composition more defined, further called modified medium. Thereby, the impact of the supplementation of either X-SEED KAT YE or X-SEED Peptone (Ohly GmbH), a yeast peptone, was tested in the concentration of 0.2 and 1.0 g/L each. Roughly, these two yeast supplements differed in their protein content (75% in X-SEED KAT vs. 67% in X-SEED Peptone) and composition [higher proportion of short-chain peptides and amino acids in X-SEED KAT (available at X-SEED product sheets)]. Furthermore, the modified medium contained fig syrup in a defined volume (cf. section “Media Composition and Preparation”) to avoid the usage of whole dried fruits. The sampling was executed once per day for offline analyses, which included determination of cell count, HPLC analysis, pH measurement, and sensory analysis. After 7 days of fermentation, aroma analyses of the final products took place. The fermentation studies were executed in biological duplicates.
Finally, fermentations in 5-L bioreactors with WK grains as well as the co-culture combination Leu + Han with the modified medium were performed. For this purpose, 4.5 L of modified medium including 1.3 g/L of X-SEED KAT and 4.5 L of modified medium without YE was used in the case of the co-culture Leu + Han and WK grains, respectively. Fermentations took place in 5-L Biostat® B Twin bioreactors (Sartorius AG, Göttingen, Germany), at 26°C with a low stirring rate of 20 rpm and no aeration. The pH value and dissolved oxygen (pO2-%-saturation) were measured online via EASYFERM PLUS VP PH/RX 325 and VISIFERM DO ECS 325 H0 (Hamilton Germany GmbH, Gräfelfing, Germany) probes. The sampling was executed three times a day for offline analyses, which included determination of cell count, HPLC analysis, pH measurement, and once-per-day sensory analysis. Fermentations were carried out in biological triplicates. Data from identical sampling times were shown as the mean value of a triple, double, or single sampling and were marked accordingly.
The total cell concentration was analyzed by using impedance measurement (MultisizerTM 3, Beckman Coulter GmbH, Brea, CA, United States). In the case of samples including grains, only the liquid fraction was analyzed. Ten microliters of the sample was diluted in 10 mL of Isoton II, and 50 μL thereof was analyzed using a 30 μL capillary. With the use of the MultisizerTM 3 Software Version 5.53, the pulse data were converted to size features. In the case of co-culturing, LAB and yeast strains were used, which differed in their cell sizes; and thus, a discrimination of the respective populations was possible. Particles with a size of 0.6–2.0 μm were considered as LAB cells, whereas particles with a size of 2.0–10.0 μm were considered as yeast cells based on cell size distribution in pure cultures, respectively. Their concentration per mL was calculated. A potential overlap of the LAB and yeast populations by crossing cell sizes was in neither case significant and thus was neglected for the determination of the real population distribution ratio, which was a good approximation. For the strains used, a chain formation could not have been microscopically verified in the conditions applied in the experiments.
The analyses of low-molecular sugars (in detail glucose, fructose, and sucrose) and ethanol and organic acids (more detailed lactic acid and acetic acid) were conducted via HPLC (Knauer Wissenschaftliche Geräte GmbH, Berlin, Germany) applying 10 μL of sample on a Nucleogel® Ion 300 OA column (Macherey-Nagel GmbH & Co. KG, Düren, Germany) at 40°C column temperature. The separation of the target compounds was achieved using an isocratic elution with 5 mmol/L of H2SO4 at a flow rate of 0.4 mL/min. The sugar and ethanol detection was performed in a refractive index detector. Organic acids were detected via multiple wavelength detector at a wavelength of 210 nm. The residual sugar was defined as the sum of glucose, fructose, and sucrose.
The final fermented products as well as the respective unfermented beverage bases were tested for a number of volatile components and fatty acids, which are described in detail as follows.
The quantification of acetaldehyde, higher aliphatic and aromatic alcohols, and acetate esters was determined by static headspace gas chromatography with flame ionization detection (HS-GC-FID) according to method 9.39 outlined in the European Brewery Convention (EBC) (
Ethyl butanoate, ethyl hexanoate, ethyl octanoate, ethyl decanoate, and ethyl dodecanoate were determined by headspace–solid-phase microextraction (HS-SPME) gas chromatography–mass spectrometry (GC-MS). The GC-MS system consisted of a Shimadzu GC 2010 interfaced with a MS-QP2010 Plus (Shimadzu Corp.) equipped with a Gerstel MPS 2XL auto sampler (Gerstel, Linthicum Heights, MD, United States) for automated HS-SPME sampling. Data evaluation was done using the GCMSsolution software Version 4.45 SP1 (Shimadzu Corp.). Esters were extracted from 2 mL of liquid sample using a 50/30 μm divinylbenzene/carboxen/polydimethylsiloxane (DVB/CAR/PDMS) fiber (Supelco, Bellefonte, PA, United States). The column used for chromatographic separation was a HP-5MS UI column [30 m × 0.25 mm i.d. × 0.25 μm film thickness from Agilent (Santa Clara)]. The extraction and GC parameters were used as described by
The analyses of volatile sulfur components were accomplished by application of two separate HS-SPME GC-MS/MS methods that were both run on an Agilent Technologies 7890B gas chromatograph interfaced to a 7000C Triple Quadrupole mass spectrometer (Agilent). This GC-MS/MS setup was equipped with a Gerstel MPS 2XL sampler (Gerstel) for automated HS-SPME sampling, and the column used for chromatographic separation was a HP-5ms Ultra Inert GC Column (30 m × 0.25 mm, × 0.25 μm film thickness from Agilent). Agilent MassHunter Workstation Software—Qualitative Analyzes (ver. B.07.00) was used for data analyses. Method 1 covered methanethiol, ethanethiol, propane-1-thiol, and butane-1-thiol. The second method covered the thioesters and sulfides, namely,
Short- and medium-chain fatty acids, herein referred to as fatty acids, were determined by HS-SPME-GC-FID. The GC-FID system used was a Shimadzu GC-2010 (Shimadzu Corp.) equipped with an Agilent CP Wax 58 FFAP (50 m × 0.32 mm × 0.32 μm). HS-SPME sampling was done using a Gerstel MPS 2XL auto sampler (Gerstel) equipped with an 85 μm polyacrylate fiber (Supelco). To extract the fatty acids, 2 mL of aliquots of the liquid samples was acidified by addition of 80 μL of 1 M HCl in 10-mL amber headspace vials. To enrich components in the headspace above the sample, sealed headspace vials we incubated for 15 min at 50°C (500 rpm), followed by an extraction for 15 min at 50°C. The loaded fiber was then desorbed for 1 min at 250°C using a split ratio of 2. A temperature program starting at 60°C, followed by a ramp of 17°C/min to 150°C, and followed by a ramp of 8°C/min to 220°C was used to separate the target analytes. The calibration ranges were as follows: butanoic acid (0.15–12 mg/L), 3-methyl butanoic acid (0.05–4 mg/L), pentanoic acid (0.03–2.4 mg/L), hexanoic acid (0.05–4 mg/L), octanoic acid (0.1–8 mg/L), decanoic acid (0.04–3.2 mg/L), and dodecanoic acid (0.03–2.4 mg/L). The lowest concentration of the calibration range was defined as limit of quantification for each component. 4-Methyl pentanoic acid was used as internal standard; it was added to the samples at a concentration of 3 mg/L prior to acidification with HCl.
Three trained experts in tasting of sour fermented beverages analyzed sensory properties of the fermentation products. For this purpose, a descriptive and evaluative analysis scheme for sour fermented beverages established at the VLB was used. After sampling, all beverages were stored in closed screw-cap tubes for a maximum of 2 days at 4°C and brought to room temperature shortly before tasting. Samples were tasted in a non-blinded manner.
In the following, the fermentation progress of the trials examining different microbial combinations (section “Different Lactic Acid Bacteria–Yeast Combinations for Producing Water Kefir Beverages”) as well as adapted media (sections “Using Fig Syrup Compared With Dry Figs Based on Fermentation With Water Kefir Grains” and “Effect of Different Yeast Extracts and Concentrations on the Performance of Leu + Han During Water Kefir Fermentation”) aiming for a more defined WK production process is shown. Finally, 5-L bioreactor fermentations applying the appropriate microorganism combination compared with fermentations with original WK grains were performed (section “Five-Liters Bioreactor Fermentation With Water Kefir Grains and Defined Co-culture Leu + Han”).
Firstly, systematic experiments to reduce the microbial complexity were conducted. Therefore, combinations of one lactic acid bacterium and one yeast strain were applied in basis medium against the usage of grains as a control. The focus was on the production of a lactic acidic beverage, which should preferably contain a low amount of alcohol besides an overall balanced sensory perception.
Course of bacteria
Course of produced ethanol, lactic, and acetic acid and consumed sugar during the WK fermentations with different LAB–yeast combinations
Concentrations of different selected volatile components in the unfermented basis medium and beverages produced with different LAB–yeast combinations and WK grains after 7 and 4 days of fermentation, respectively. The quantified aroma-relevant analytes included higher aliphatic alcohols and esters
The grains showed the fastest rise of bacterial cell growth compared with the defined co-cultured bacteria cells (1 vs. 2 days to end exponential growth phase). However, at the end of the fermentation, similar bacterial cell concentrations were reached (∼1 ⋅ 108 cells/mL). In contrast, lower final yeast cell counts of 7.64 ⋅ 106 cells/mL than in co-culture (∼2.5 ⋅ 107 cells/mL on average), which showed similar progress of the yeast cell counts, were observed. In the co-culture, the bacteria cell growth depended on the combined yeast. Consequently, the combinations including Dek showed the lowest yeast cell growth compared with the other combinations but the fastest bacteria growth in case of co-culture, which resulted in an amount of 1.8 g/L of lactic acid (cf.
Combinations with Han and Sac led to ethanol concentrations of ∼6 g/L (∼0.76 Vol.-%), which was sensory recognizable in the case of Liq + Han. The combination of Leu + Sac showed the lowest pH decrease to finally 4.3 (cf.
Sensory properties of products fermented with different LAB–yeast combinations after 7 days of fermentation. Shown are the mean data of two samples produced in biological duplicates. The rating scale of the descriptive analysis
In contrast to fruity aroma components, thiols and sulfurous components were found in a detectable range in all combinations as well as in the basis medium, which is displayed in
Although the combinations of Dek with Liq and Leu (orange lines in
Further, the blue lines in
Summarizing, the fermentations with the co-cultivation of LAB and yeasts lasted longer to get a final product than the fermentations with grains. The fermentation with WK grains was faster and led to a more complex taste and a higher diversity in perceptible sensory properties but resulted in a higher content of alcohol in the final product. By using defined starter cultures in a co-cultivation, the formation of ethanol was lower and might be more controllable than by fermentations with grains.
Sensory-wise, the combination of Leu + Han was preferred by the panelists, which was based on a well-balanced characteristic between sourness (lactic acid) and sweetness and the presence of desired attributes such as fruity and carbonized. Based on the results, the combination was chosen for further trials.
In order to implement more defined media components in the production process, investigations with fig syrup were executed. For this purpose, different volumes of fig syrup were tested to get the most appropriate concentration equal to the usage of one dry fig per liter. In a test row, basis media were investigated, which differed in the volumes of liquid syrup. Sugar and lemon juice concentrate, as was used above, were added, and sensorial tests were performed. In conclusion, 9.0 mL of fig syrup per liter corresponded to the usage of one dry fig per liter regarding the taste as the basis. Based on this, fermentations with WK grains with the chosen amount of fig syrup compared with the usage of dry figs were performed, and results can be seen in
In order to avoid the noticeable presence of sulfurous components in the final product when applying defined starter cultures, two different YEs, developed for LAB cultivation especially, were used as a sufficient supplement of nutrients. The fermentation performance of the microorganism combination Leu + Han without the supplementation of YE was not as good as in the first study (cf.
Course of cell concentrations
Like previously shown, the fermentations with the combination Leu + Han without adding YE differed when using fig syrup or dried fruits. These results (cf.
Concentrations of different volatile components in modified media (references) and in 7-day fermented beverages produced with the Leu + Han combination and different types and amounts of YE. The quantified aroma-relevant analytes included higher aliphatic alcohols and esters
The comparison of the volatile components, displayed in
The analyses of the volatile components of the products fermented by Leu + Han with the usage of different concentrations of YE and the replacement of dried figs by fig syrup showed a smaller amount of esters and higher aliphatic alcohols, e.g., acetaldehyde and ethyl acetate and can be seen in
The fatty acid concentration in the products during the supplementation of YE decreased similarly to the product previously described, and no influence of different YEs as well as concentrations was observed. Interestingly, the concentration of butanoic acid of the fermented product without YE was the highest in all experiments at about 24 mg/L. The associated rancid bad taste was clearly identified in the sensory evaluation as well (cf.
Sensory properties of the products fermented with the microbial combination Leu + Han under the investigation of 0, 0.2, and 1 g/L of two different YEs after 7 days of fermentation. Shown are the mean data of two samples produced in biological duplicates. The rating scale of the descriptive analysis
Of particular note during these experiments was the presence of sulfurous components, recognizable in detectable volatile components (
The final products showed improved sensory properties, recognizable in a higher pronounced sourness, more full-bodied products, and more fruitiness as compared with the study without YE in co-cultivation (dotted blue line in
In order to estimate the reproducibility of the WK production by co-cultivation compared with WK grains, fermentations in 5-L bioreactors were performed (
Off-line data of WK fermentations in 5-L bioreactors with grains
It is of particular note that the grain-based fermentations gave desirable sensorially similar products after 3, 6, and 5 days, whereas beverages produced by co-cultures were sensorially similar after 7 days of fermentation each. This was not obvious from the measured data but is of importance for later industrial application. The reached cell concentrations of yeast and bacteria were higher in co-cultivation than during the fermentation with grains (∼3-fold for bacteria and 10-fold for yeast). Incidentally, the yeast concentration decreased after 3 days of fermentation by using grains, which is most probably reasoned by the low stirring rate and the impeded homogenization caused by an increasing grain mass during fermentation. The gray lines in
The usage of grains led to higher ethanol concentrations (23 g/L at 3 days; ∼3 Vol.-%) than the co-cultivation of Leu + Han, accompanied by a nearly complete consumption of the sugar at the fourth day of fermentation, which is comparable with the fermentations in section “Different Lactic Acid Bacteria–Yeast Combinations for Producing Water Kefir Beverages” in 1-L scale. After 6 days of fermentation with co-cultures, the amount of ethanol increased to 3.3 g/L (0.42 Vol.-%), and to a final value of 7 g/L on average after 1 week of fermentation rapidly. When comparing the three fermentations with co-cultures, the sugar consumption as well as the ethanol content in the end deviated (14.8 vs. 2.7 and 3.5 g/L), although staying below 3 g/L until 6 days of fermentation, which is why the mean value of ethanol was that high. The lactic acid concentration reached higher values by co-cultivation than by the fermentation with grains (4.98 vs. 1.33 g/L), which was recognizable in the taste of the final products.
Additional online analyses of the pH, dissolved oxygen, and redox value are illustrated in
Sensory properties
The aim of this study was to expand the knowledge about the impact of different microbial consortia representing different degrees of complexity on the characteristics of WK. Therefore, LAB and yeasts isolated from WK grains were used for fermentations as defined starter cultures in co-cultivation against the original grains as a control. The focus of the trial was firstly to reduce the complexity of WK fermentations in order to provide a more controllable and reproducible process and secondly to monitor differences in cell growth, the corresponding metabolization of sugars, production of organic acids, and the perceptible taste in combination with chemical analysis of the microbial volatile components.
Based on previous internal studies regarding isolated defined strains from traditional WK grains (data not shown), combinations of two different LAB,
In summary, the fermentations with grains were much faster than in co-cultivation, which could be explained by a high adaptation of the higher number of microorganisms in grains among one another in contrast to only two species that might not be able to fully compensate for all microbial interactions (
Combinations including
Different gas chromatographic methods were conducted to quantify volatile components in the fermented products.
In contrast to the presence of fruity notes, in all products except for the combinations with
Irrespective of the occurred off-flavors, the combination of the yeast
Regardless of the type of the chosen starter cultures, in addition to a carbon source (mostly raw cane sugar or pure sucrose), most dried fruits (e.g., figs or cranberries) are used for the production of WK. Although the latter serves as a source of amino acids, vitamins, and minerals, the small applied amount of dried fruits compared with the richness in nutrients of fresh fruits leads to a WK medium relatively poor in nutrients (
In order to overcome nutritional limitations in co-culture approaches, the supplementation of YE in different concentrations was investigated.
Concluding, the addition of at least 1 g/L of YE, especially the applied X-SEED KAT YE, promoted a nutritional supply and might balance a lack of nutrition, which avoids the production of high amounts of sulfurous components when defined starter cultures are used. Furthermore, this effect allowed the replacement of dried figs with fig syrup. These steps led to a more defined process of the WK production.
However, the concentration of YE could be increased in order to decrease the fermentation time and therefore the production process bearing in mind avoiding noticeable yeasty flavor components in smell or taste. The usage of X-SEED KAT YE showed slightly higher lactic acid concentrations (
The development of a defined process for WK production instead of using complex grains enables the possibility to produce a WK-like beverage by using the co-cultivation of one LAB and yeast. In order to compare the production of WK with grains with the usage of defined co-cultivation under well-monitored and scalable conditions, 5-L bioreactor fermentations were performed in triplicates (see
Summarizing, the WK beverage gained by co-cultivation with
The work presented herein described steps from a traditionally produced process to a defined water kefir fermentation process applying a co-culture of only one LAB and one yeast, respectively. It could be shown that the main characteristics of WK—a fruity, aromatic, and acidic beverage made by fermentation of characteristic strains—were achieved by the use of two microbes
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
AT, MSc, ML, and SK conducted the experiments. NR was responsible for the aroma analyses. MSe and SK performed the data analyses. MSe, MSc, and SK drafted the manuscript. ML reviewed and contributed to the structure and content of the manuscript. All authors contributed to the article and approved the submitted version.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The reviewer SJ declared a shared affiliation with one of the authors MSc, to the handling editor at time of review.
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Parts of this research was funded by the German Federal Ministry for Economic Affairs and Energy (program INNO-KOM; research projects: CoKuFerm, funding no. 49VF200039 and sfAFG, funding no. 49MF190041).
We thank Christian Schubert and Sarah Thörner for the analyses of volatile components and data evaluation. We also acknowledge support by the German Research Foundation and the Open Access Publication Fund of TU Berlin.
The Supplementary Material for this article can be found online at: