Non-Alcoholic Beverages from Fermented Cereals with Increased Oligosaccharide Content

Fermented foods and beverages are produced worldwide using various manufacturing techniques, raw materials and microorganisms (1). The problem is that production methods for most of traditional fermented foods are not optimised or industrialised, and are rather small-scale, localised processes. It is benefi cial to revisit such traditional products and to conduct biotechnological investigations to explore and enhance their functional properties, and thus promote their commercial exploitation outside of their local market. Kvass is a cereal-based beverage traditionally produced from fermented rye and barley malt, rye fl our, and stale rye bread in Lithuania and other eastern European countries. It is a non-alcoholic beverage, so the ethanol content should be negligible, and it is considered to be spoiled if alcohol accumulates to higher levels. As opposed to sourdough bread, kvass undergoes no heat processing aft er fermentation and thus contains high cell counts of viable yeast and lactic acid bacteria (LAB). The microfl ora of kvass fermentation is normally composed of Saccharomyces cerevisiae and LAB, but the composition of species is expected to be quite variable due to diff erences ISSN 1330-9862 original scientifi c paper


Introduction
Fermented foods and beverages are produced worldwide using various manufacturing techniques, raw materials and microorganisms (1).The problem is that production methods for most of traditional fermented foods are not optimised or industrialised, and are rather small--scale, localised processes.It is benefi cial to revisit such traditional products and to conduct biotechnological investigations to explore and enhance their functional properties, and thus promote their commercial exploitation outside of their local market.
Kvass is a cereal-based beverage traditionally produced from fermented rye and barley malt, rye fl our, and stale rye bread in Lithuania and other eastern European countries.It is a non-alcoholic beverage, so the ethanol content should be negligible, and it is considered to be spoiled if alcohol accumulates to higher levels.As opposed to sourdough bread, kvass undergoes no heat processing aft er fermentation and thus contains high cell counts of viable yeast and lactic acid bacteria (LAB).The microfl ora of kvass fermentation is normally composed of Saccharomyces cerevisiae and LAB, but the composition of species is expected to be quite variable due to diff erences ISSN 1330-9862 original scientifi c paper doi: 10.17113/ft b.54.01.16.4106 Non-Alcoholic Beverages from Fermented Cereals with Increased Oligosaccharide Content in fermentation techniques and fermentation media.To date, microbial associations in kvass fermentation have been poorly investigated (2).Eff orts to use pure starter cultures of LAB to produce acceptable kvass have not been documented.
LAB play a key role in food fermentations where they not only contribute to the development of the desired sensory properties in the fi nal product but also to their microbiological safety (3).The antimicrobial eff ect of LAB is mainly related to the production of organic acids; however, some strains are able to synthesise antimicrobial substances -bacteriocins (3).During recent years the interest in bacteriocin-producing LAB has increased because of their potential to produce natural antimicrobial agents to enhance the safety of food products.The application of such LAB for non-alcoholic cereal-based beverage produc tion provides new opportunities for development of functional fermented products.
Cereals are important source of hemicelluloses, which include arabinoxylans, β-glucans and arabinogalactans, also called dietary fi bre (4).Among commonly grown cereals, rye is considered to be a healthy cereal with its high content of dietary fi bre and associated components.The growing interest in foods and beverages with increased dietary fi bre content is observed worldwide.Such foods and beverages are oft en called prebiotic foods, because of their ability to stimulate the growth of benefi cial colonic microfl ora (5,6).Research on prebiotics and other novel health-promoting food components has been intensive for over a decade.Arabinoxylan-derived arabinoxylooligosaccharides (AXOS), which may have various chemical structures, depending on the xylan source and the degradation method used, stand increasingly in the spotlight as potential prebiotics (7,8).During the past decade, the studies of the possibilities to produce the AXOS by using biocatalytic conversion have received more att ention.The enzymatic hydrolysis allows control of the end-products and to obtain AXOS having the desired degree of polymerization and substitutions (9).Therefore, there is an interest in new enzymatic technologies for partial and selective degradation of arabinoxylan (AX).
Most recently, extrusion technology has been used in the development of novel, value-added foods (10,11).Extrusion cooking is preferable to other food-processing techniques because it is a continuous process with high productivity and signifi cant nutrient retention, owing to the high temperature and short time required.During extrusion cooking, the raw materials undergo many chemical and structural transformations, such as starch gelatinization, protein denaturation, complex formation between amylose and lipids (12).No signifi cant changes were found in dietary fi bre content at mild or moderate extrusion conditions, but some fi bre components could be solubilised (13).In addition, the extrusion denatures undesirable enzymes, inactivates some antinutritional factors, destructs naturally present toxins, and sterilises the fi nal product (14).The application of new raw materials in the extrusion process opens perspectives both for the increase of the assortment of extruded products and the development of novel components for food industry that exhibit specifi c technological properties.
The aim of this study is to develop a new technology for making traditional Lithuanian non-alcoholic beverage kvass from fermented cereals by extending the spectrum of raw materials such as extruded rye and applying new biotechnological resources such as xylanolytic enzymes and antimicrobially active LAB to improve the functional properties.

Raw materials and enzymes
Rye wholemeal extruded by a single-screw extruder was supplied by the miller Ustukiu Malunas (Pasvalys, Lithuania).Extrusion conditions were: feed rate 110 kg/h, temperature 45/65/110 °C (individual zones), screw speed 350-410 rpm.The extruded rye fl our was passed through a 0.5-mm sieve.

Lactic acid bacteria and yeast
Lactobacillus sakei KTU05-6 (optimal temperature 30 °C) and Pediococcus pentosaceus KTU05-10 (optimal temperature 37 °C) from culture collection of the Department of Food Science and Technology (Kaunas University of Technology, Kaunas, Lithuania) were used for beverage fermentation.Tested LAB were isolated from spontaneous rye sourdough and selected due to their high antimicrobial activity (19).LAB were cultured in MRS broth (CM0359; Oxoid Ltd, Hampshire, UK) at appropriate temperatures for 48 h prior to use.For parallel studies dry baker's yeast Saccharomyces cerevisiae (Fermentis, Marcqen-Baroeul, France) was used.

Arabinoxylan enzymatic hydrolysis
For hydrolysis of AX, 2.5 mg of each rye sample were suspended in 1 mL of 0.1 M sodium acetate buff er (pH=5) and incubated with four diff erent enzyme preparations separately (600 xylanase units (XU) per g of cereals) at 55 °C for 5 h.Enzymatic reaction was stopped by heating the reaction mixture at 100 °C for 10 min.The degree of AX degradation was determined by high-performance anion--exchange chromatography (HPAEC).

Preparation of beverages
Modifi ed technological scheme for traditional kvass production was used to prepare a non-alcoholic beverage from extruded rye under laboratory conditions.The beverage mash was prepared from extruded wholemeal rye and tap water (55 °C) at a ratio of 1:50 and treated by enzymatic hydrolysis.Simultaneous enzymatic liquefaction and saccharifi cation step was performed in 2 h at 55 °C by adding selected amounts of α-amylase (48.6 amylase units (AU) per g of cereals) and amyloglucosidase (20 glucosidase units (GU) per g); additionally β-xylanolytic enzyme preparation Ceremix Plus MG (12.5 XU/g) was added for the degradation of non-starch polysaccharides.The activities of enzymes were selected according to the producer's recommendations and the results of initial tests.Aft er liquefaction, the mash was boiled at 100 °C for 10 min to inactivate the enzymes, and cooled down to a temperature suitable for LAB (30 and 37 °C) or yeast (30 °C).The mash was inoculated with 10 % of starter culture suspension (10 8 colony-forming units (CFU) per g, moisture content 65 %) or dry yeast (0.6 g per L of mash).Fermentation was carried out for 48-60 h and samples for oligo-and monosaccharide analysis were collected every 12 h and fi ltered with Eppendorf fi ltration system (Eppendorf, Hamburg, Germany).Sweetness of fermented mash was adjusted by adding sugar, and caramelised malt was used to obtain brown colour of the drink.The beverage was then cooled, fi ltered and samples were taken for quality evaluation.The rest of the beverage was dispensed into 250-mL amber glass bott les.The bott les were closed with screw caps and stored at 0-6 °C for 30 days for microbial analysis.Kvass produced from commercial rye malt extract and fermented by yeast under the same conditions served as the control.All fermentations were performed in duplicates.

Chromatographic analysis of saccharides
HPAEC analysis of oligosaccharides extracted from extruded rye was performed on HPLC system, equipped with Waters 717 Plus autosampler, two Waters 515 HPLC pumps (Waters, Milford, MA, USA), analytical column CarboPac PA-100 (250 mm×4 mm) and guard column CarboPac PA-100 Guard (50 mm×4 mm; Dionex, Sunnyvale, CA, USA).Decade pulsed amperometric detector (PAD) (Waters) with a gold electrode (Antec Leyden, Zoeterwoude, The Netherlands) was used in pulse mode at 30 °C and Millenium 32 soft ware (Waters) for instrument control and data handling.
The eluents for gradient analysis of the oligosaccharides were A: 1 M sodium acetate in 100 mM sodium hydroxide (Sigma-Aldrich) and B: 100 mM sodium hydroxide.The fi rst isocratic phase was 15 min in 100 % B, followed by a linear gradient from 100 % B to 88 % B and 12 % A in 20 min.The second isocratic phase was 5 min in 88 % B and 12 % A. The second linear gradient was back to 100 % B in 5 min.The last isocratic phase was 5 min in 100 % B, in which the column was stabilised before the next injection.The total analysis time was 50 min and the fl ow rate was 1 mL/min.
The eluents for gradient analysis of the monosaccharides were A: 200 mM sodium hydroxide and B: water.The fi rst isocratic phase was 13 min in 18 % A and 82 % B, followed by a linear gradient to 100 % A in 9 min.The second isocratic phase was 15 min in 100 % A. The second linear gradient was back to 18 % A and 82 % B in 8 min.The last isocratic phase was 10 min in 18 % A and 82 % B. The total analysis time was 55 min and the fl ow rate was 1 mL/min.External standards were used in both oligo-and monosaccharide analysis for accurate identifi cation and quantifi cation of the formed sugars.For oligosaccharide analysis, the mixtures of XOS (from X 2 to X 6 ) and maltooligosaccharides (from M 3 to M 7 ) (Megazyme International Ireland) were used.AXOS (A 3 X, A 2 X, A 2 XX, A 2,3 X, A 2,3 XX) were prepared at the University of Helsinki, Finland.The preparation, purifi cation and identifi cation of these AXOS are described by Rantanen et al. (20) and Pastell et al. (21).The amount of isolated AXOS was analysed in those papers and these values were used as standard to quantify unknown AXOS mixtures.Additionally, maltose, cellobiose, 1,3-1,4-β-glucotriose, two glucotetroses (diff erent in structure) and galactotetrose were used (Megazyme International Ireland).The peak heights (mV) were used in the quantifi cation of carbohydrates.For the monosaccharides, an external standard mixture containing arabinose, glucose, fructose, xylose (Merck KGaA, Darmstadt, Germany) and sucrose (Sigma-Aldrich) was used.The external standard mixture was injected aft er every four samples in both methods to monitor the retention times and detection responses.

Beverage quality evaluation
Total titratable acidity (TTA) was analysed by titration with 0.1 M NaOH solution, using phenolphtalein as indicator, and expressed in degrees.One degree (1°) of acidity corresponds to 1 mL of 1 M NaOH required to neutralise the organic acids present in 100 mL of sample.The pH value was measured and recorded using a pH electrode (PP-15; Sartorius, Goett ingen, Germany).Volatile acid content was determined by steam distillation and aft er titration with 0.1 M NaOH it was expressed in g of acetic acid per L. Concentration of lactic acid (g/L) was calculated from the diff erences between TTA and volatile acidity.
Contents of ethanol (% by volume) and extractable materials (% by mass) were determined aft er distillation by den sitometric method using pycnometer at 20 °C.Byproducts of fermentation potentially considered as fl avour compounds or as impurities (methanol, higher alcohols and esters) were analysed using a Hewlett -Packard 5890 gas chromatograph (Hewlett -Packard, Wilmington, DE, USA) equipped with a split-splitless injector and a fl ame ionisation detector (FID).The injection temperature was 200 °C and helium served as a carrier gas with a fl ow rate of 1.2 mL/min.Compounds were separated on a column Zebron ZB-WAX (30 m×0.25 mm×0.25 m, 100 % polyethylene glycol; Phenomenex, Torrance, CA, USA).The temperature program used was: initial temperature of 40 °C for 5 min, rising to 100 °C at a rate of 4 °C/min, and a 2 min hold.The detection temperature was set at 250 °C.Qualitative analysis was done by comparison of retention times of the chromatogram of the kvass samples to that of the standard solution of 6 alcohols and esters (methyl acetate, ethyl acetate, methanol, propanol, isobutanol, isoamyl alcohol; Sigma-Aldrich).A standard solution was prepared by dissolving the analysed compounds in 96 % ethanol.Quantitative analysis was conducted using generated calibration curves.

Microbial analysis of beverages
The eff ect of LAB on the microbiological characteristics of beverages was determined by counting their number and the total number of microorganisms aft er 1, 10, 20 and 30 days of storage.LAB counts were determined on MRS agar (Liofi lchem, Roseto degli Abruzzi, Teramo, Italy) using plate count techniques.Plates were incubated for (72±4) h under anaerobic conditions (using atmosphere generation system AnaeroGen; Oxoid, Basingstoke, UK).The experiment was carried out in triplicate.Microbiological contamination of fermented products was evaluated by calculating the total number of microorganisms and was expressed in CFU/mL using plate count agar (Liofi lchem) according to ISO 4833-2:2013 method (22).

Sensory evaluations of the beverages
Sensory evaluations were conducted by 11 panellists, consisting of KTU staff and students, using a 7-point rating scale for the colour, odour and fl avour att ributes.Value of 1 corresponded to the lowest intensity and of 7 to the highest intensity of the att ribute.Overall acceptability was evaluated using a 9-point hedonic scale where 9 is 'like extremely' and 1 'dislike extremely'.Coded samples were served and water was provided for rinsing between the sensory evaluations of the samples.

Statistical analysis
All analytical determinations were performed at least in triplicates.Results were expressed as mean value±standard deviation (S.D.).Statistical data analysis was conducted using a Microsoft Excel spreadsheet, and the statistical program for Windows (SPSS, v. 16.0;IBM Corporation, Armonk, NY, USA) was used for comparison of the mean values by one-way analysis of variance.The results were signifi cant at p<0.05.

Eff ect of enzyme preparations on rye arabinoxylan hydrolysis
All tested enzyme complexes were capable of degrading rye AX main chain, releasing various monosaccharides and AXOS.d-Arabinose (Ara), d-xylose (Xyl) and d-glucose (Glc) were determined as a result of full AX and starch degradation (Fig. 1a).The results showed that Depol 680P and Depol 692L gave the highest yields of Glc (602.0 and 382.2 mg/g, respectively) and Xyl (23.2 and 23.4 mg/g).Thermomyces lanuginosus-derived xylanase Pentopan Mono BG had the highest activity, but unexpectedly poor release of Xyl, which can be explained by this xylanase assignation to family GH11.According to the results of other authors, a GH11 family xylanase needs the presence of other xylanolytic enzymes to act effectively on AX for complete degradation of polymeric chain (23,24).Kormelink et al. (25) reported that pure GH10 family endo-1,4-β-d-xylanases (from Aspergillus awamori) converted up to 72 % of the wheat fl our AX into mono-, di-and oligosaccharides with the degree of polimerisation (DP) of 3-10, whereas endo-1,4-β-d-xylanases from GH11 family hydrolysed only 50 % of AX into oligosaccharides with DP=5-10.Shorter AXOS were also formed by GH10 family endo-1,4-β-d-xylanases, compared to those produced by GH11 family endo-1,4-β-d-xylanases (7,26).
Higher value of cereal-based products is related to XOS and AXOS, which according to other authors have shown promising prebiotic properties (27,28).However, linear oligosaccharides may be fermented rapidly in the colon and thus those with more complex carbohydrate structures are currently studied to fi nd more slowly fermenting prebiotics (29,30).Van Crayeveld et al. (31) determined that AXOS with DP≤3 caused increased colonic acetate and butyric concentrations and increased the amount of bifi dobacteria.However, litt le information is available on which AXOS would be the most benefi cial, because it is complicated to obtain enough purifi ed and well-defi ned AXOS preparations.

Saccharide profi les in fermented rye mash
Lactic acid bacteria P. pentosaceus and L. sakei and baker's yeast were used for preparation of a non-alcoholic beverage from extruded rye.Diff erent enzymatic treatments were tested for extruded rye mash preparation: (i) with enzyme composition of α-amylase from A. oryzae, glucoamylase from A. niger and xylanolytic enzyme complex from H. insolens and B. licheniformis (as the best rye AX-degrading enzyme preparation), and (ii) only with xylanolytic enzyme complex.Control mash was prepared from non-enzymatically treated extruded rye.Effi ciency of fermentation process was evaluated based on the changes in pH and TTA, saccharide mass fraction as well as the amount of ethanol and some volatile compounds.
Changes in saccharide mass fraction during fermentation were analysed by HPAEC every 12 h (Fig. 2).The consumption rate of Glc and sucrose (Suc) aft er 24 h of fermentation showed that the LAB used Glc fi rst as the main substrate, and only when the mass fraction of this monosaccharide in fermentation medium was low, Suc was used (Table 1).In non-enzymatically treated mash, the LAB decreased the mass fraction of Glc by 62.1 % (P.pentosaceus) and 69.0 % (L.sakei); additionally, Suc mass fraction was decreased by 62.3 and 61.1 %, respectively.Aft er treatment with xylanolytic enzyme complex, P. pentosaceus and L. sakei consumed 88.5 and 75.7 % of Glc, respectively, and Suc mass fraction remained intact.In the samples treated with amylolytic and xylanolytic enzyme complexes, Glc mass fraction decreased only by 19.3 % (P.pentosaceus) and 5.2 % (L.sakei), the amount of Suc was not changed.The mass fractions of Ara (10.9-15.2mg/g) and Xyl (8.1-13.7 mg/g) were almost similar in all enzymatically treated samples fermented with L. sakei and P. pentosaceus and remained unchanged until the end of fermentation.

Infl uence of microorganisms on beverage quality
The changes in quality parameters (pH and TTA) of beverages from enzymatically treated (with amylolytic and xylanolytic enzyme complex) extruded rye during fermentation by LAB or yeast are presented in Fig. 3.It was determined that pH of beverages fermented by LAB reached lower values (3.25-3.54)compared to yeast-fermented (4.34).On increased during fermentation, and higher amount of organic acids was detected in LAB-fermented beverages (2.4-5.8 °) than in the yeast-fermented ones (0.9 °).Among LAB, P. pentosaceus produced twice more organic acids than L. sakei.According to the pH and TTA values, the in-crease of acidity was almost the same as aft er 24 h.Small changes in pH and TTA values aft er 48 h of LAB fermentation as compared to 24 h showed that beverage fermentation using LAB can be completed within 24 h.Fermentation of traditional home-made kvass by yeast usually provides from 3 to 4 % (by volume) of alcohol in the end product, but in non-alcoholic fermented beverage it should not exceed 1.2 % (by volume).The replacement of yeast by LAB allowed decreasing alcohol content in the beverage to an appropriate amount and fulfi lling this requirement for non-alcoholic beverages.Determination of alcohol content (Fig. 4a) showed that both LAB produced some amount of alcohol (0.1-0.3 %), but it did not exceed the limits for non-alcoholic beverages, while yeast produced 4.5 % of alcohol.
LAB fermentation as compared to yeast provided more fl avour compounds that participate in the formation of product aroma and taste (Fig. 4b).The main product of LAB fermentation is lactic acid, but some other compounds, like acetic acid, diacetyl, methylacetate or For estimation of LAB antimicrobial properties, lactic acid bacteria cell count and total microorganism cell count were evaluated during beverage storage for 30 days at 0-6 °C (Fig. 5).Analysis showed that aft er 30 days, viable cells of L. sakei decreased in the beverage more than 4 times (from 2.8•10 7 to 6.2•10 6 CFU/mL), but viability of LAB remained high (more than 10 6 CFU/mL).The use of L. sakei resulted in a reduced total microorganism growth in the beverage during storage and confi rmed that this LAB is capable of retarding the growth of extraneous microorganisms.

Sensory characteristics and acceptability of beverages
Statistically signifi cant (p≤0.05)diff erences between commercial kvass made from rye malt extract (control) and beverages made from extruded rye fermented by yeast or LAB were determined in sweetness, sourness, colour and odour intensity (Fig. 6).The use of extruded rye for beverage production increased odour and decreased colour intensity of the product.Furthermore, using ex-truded raw material for fermentation, intensity of brown colour was lost, compared to the control kvass.Beverages fermented by LAB also changed fl avour properties: sweetness and bitt erness decreased, while sourness increased.
Acceptability (Fig. 7) of beverages fermented by LAB was lower (p≤0.05)compared to commercial kvass, but it was higher (p≤0.05)than of beverage fermented by yeast.Comparing both LAB, higher acceptability was determined of the beverage fermented by L. sakei than by P. pentosaceus.Assessors preferred L. sakei-fermented beverage due to its higher fl avour and odour intensity and lower sourness.

Conclusions
Extruded rye, xylanolytic enzymes and LAB can be used for production of novel, safe and higher value (with increased amount of oligosaccharides) non-alcoholic beverages.Using the selected xylanolytic enzyme complex Ceremix (Humicola insolens and Bacillus licheniformis) the yield of AXOS and XOS in beverage mash was increased by 300 and 1100 mg/L, respectively, which remained intact during fermentation by LAB.Beverages fermented by the LAB (Pediococcus pentosaceus and Lactobacillus sakei) with antimicrobial activity had lower pH values (3.3-3.5) and higher TTA values (2.4-5.8°),higher concentrations of volatile compounds (0.22-0.45 g/L) and less alcohol (0.1-0.3 % by volume) than the beverage fermented by yeast (pH=4.3,TTA=0.9°,γ(volatile compounds)=0.09g/L and ϕ(alcohol)=4.5%).The acceptability of the beverage from extruded rye fermented by L. sakei was higher than of the yeast-fermented beverages and close to the acceptability of commercial kvass made from rye malt extract.

Table 1 .
Mass fraction of saccharides determined in extruded rye mash during fermentation with Pediococcus pentosaceus and Lactoba-