Art 2014 Manufacture and characterization of a

SAVE OFFLINE

    Manufacture and characterization of a yogurt-like beverage made with oat flakes fermented by selected lactic acid bacteria Nionelli Luana, Coda Rossana, Curiel Jos´e Antonio, Poutanen Kaisa, Gobbetti Marco, Rizzello Carlo Giuseppe PII: DOI: Reference:

S0168-1605(14)00223-2 doi: 10.1016/j.ijfoodmicro.2014.05.004 FOOD 6531

To appear in:

International Journal of Food Microbiology

Received date: Revised date: Accepted date:

18 December 2013 27 April 2014 4 May 2014

Please cite this article as: Luana, Nionelli, Rossana, Coda, Antonio, Curiel Jos´e, Kaisa, Poutanen, Marco, Gobbetti, Giuseppe, Rizzello Carlo, Manufacture and characterization of a yogurt-like beverage made with oat flakes fermented by selected lactic acid bacteria, International Journal of Food Microbiology (2014), doi: 10.1016/j.ijfoodmicro.2014.05.004

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

T

Manufacture and characterization of a yogurt-like beverage made with oat flakes fermented

IP

by selected lactic acid bacteria

SC R

Nionelli Luana1, Coda Rossana2, Curiel José Antonio1, Poutanen Kaisa 2,3, Gobbetti Marco1, Rizzello Carlo Giuseppe*1

Department of Soil, Plant and Food Science, University of Bari Aldo Moro, 70126 Bari, Italy

2

VTT, Techinical Research centre of Finland, 02150, Espoo, Finland.

3

Department of Clinical Nutrition, University of Eastern Finland, Kuopio Campus, P.O. Box 1627,

MA

NU

1

CE P

TE

D

FIN-70211 Kuopio, Finland

*Corresponding author: Department of Soil, Plant and Food Sciences, University of Bari Aldo

AC

Moro, via Amendola 165/a, 70126 Bari, Italy, Tel.: +39.080.5442948; fax: +39.080.5442911. E-mail address: [email protected] Running title: oat flakes fermented beverage

1

ACCEPTED MANUSCRIPT

T

Abstract

IP

This study aimed at investigating the suitability of oat flakes for making functional beverages.

SC R

Different technological options were assayed, including the amount of flakes, the inoculum of the starter and the addition of enzyme preparations. The beverage containing 25% (wt/wt) of oat flakes

NU

and fermented with L. plantarum LP09 was considered optimal on the basis of sensory and technological properties. The enzyme addition favored the growth of the starter, shortened the time

MA

needed to reach pH 4.2 to ca. 8 h, and favored a decrease of the quotient of fermentation. Fermentation increased the polyphenols availability and the antioxidant activity (25 and 70%

D

higher, respectively) and decreased the hydrolysis index in vitro. Sensory analyses showed that

TE

fermented oat flakes beverage had the typical features of a yogurt-like beverage, enhancing the overall intensity of odor and flavor compared to the non-fermented control. Selection of proper

CE P

processing and fermentation condition allowed the obtainment of a beverage with better nutritional

AC

and sensory properties.

Keywords: oat flakes, fermentation, starters, lactic acid bacteria

2

ACCEPTED MANUSCRIPT 1. Introduction The demand of consumers for non-dairy milk substitutes with high acceptance and functionality is

T

increasing (Mårtensson et al., 2000). Cereal-based beverages have a huge potential either to fulfill

IP

this expectation and to act as potential vehicles for functional compounds such as antioxidants,

SC R

dietary fiber, minerals, prebiotics and vitamins (Kreisz et al., 2008).

A variety of technologies (e.g., cooking, sprouting and milling) are routinely used to process cereals, but fermentation still remains one of the best choice to improve nutritional and sensory

NU

properties, and shelf-life (Mattila-Sandholm, 1998). A large proportion of cereals is traditionally

MA

and currently processed into foods and beverages through fermentation (Nout, 2009). Although several preparations remain like a house art, especially in African countries, the raw grain materials and/or the type of fermentation are the main criteria to classify cereal-based fermented beverages.

TE

D

Lactic acid bacteria (LAB, Lactobacillus and Pediococcus spp.), Enterobacter spp., yeasts (Candida, Debaryomyces, Endomycopsis, Hansenula, Pichia, Saccharomyces and Trichosporon

CE P

spp.) and filamentous fungi (Amylomyces, Aspergillus, Mucor and Rhizopus spp.) are mainly used for the manufacture of cereal-based alcoholic beverages (e.g., tchoukoutou, jnard), non-alcoholic

AC

beverages (e.g., uji, ben-saalga), porridges (e.g., mawè), and cooked gels (e.g., kenkey, idli, and mifen) (Nout, 2009).

The selection of appropriate starter cultures for each variant of cereal beverage is an industrial need to drive, accelerate and standardize the fermentation (Coda et al., 2014). Selected starters, through their complex enzyme systems, generate metabolites (volatile and non-volatile) that provide peculiar flavor attributes to fermented cereal-based foods (Salmeròn et al., 2009; Salmeròn et al., 2014). Moreover, the mechanisms by which LAB fulfill the role of efficient cell factory for the production of functional biomolecules and food ingredients to enhance the quality of cereal based beverages, were largely demonstrated (Waters et al., 2013). The failure of the fermentation may lead to spoilage and/or to survival and contamination of pathogens, thereby creating unexpected healthy risks (Coda et al., 2014). Fermentation by LAB is an effective tool to prevent microbial 3

ACCEPTED MANUSCRIPT contamination and it may positively affect the nutritional and functional features of cereal-based beverages (Katina et al., 2005). This also allows to develop market strategies to build up nutritional

T

claims that respond to consumer awareness towards healthy diet (Coda et al., 2014).

IP

Oat, mostly as flakes, is included in the human diet because of the healthy status, which is mainly

SC R

related to the high concentration of β-glucans. Oat is an excellent source of energy and unsaturated fatty acids (Klensporf and Jeleń, 2008), and it contains dietary fibers, high quality proteins and volatile compounds (Heiniö et al., 2001; Heiniö et al., 2002). Different classes of natural

NU

antioxidants, like tocols, phenolic compounds and avenanthramides are also largely found in oat

MA

(Peterson, 2001). Although still debated in some countries, clinical data showed that oat may be included in the gluten-free diet (Sontag-Strohm et al., 2008), as recommended by Food and Drug Administration in the USA (Sontag-Strohm et al., 2008).

TE

D

When collected from the field, oat lacks flavor, and the development of the aroma inevitably requires heat treatments (Heydanek and McGorrin, 1981). Without suitable heating, oat products

CE P

retain flat, green, raw and bitter taste (Klensporf and Jeleń, 2008). The distinctive flavor results from lipid oxidation and n-heterocyclic compounds, which are synthesized during thermal

AC

processing of groats (Heydanek and McGorrin, 1986). Simultaneously, the thermal treatment inactivates lipolytic enzymes (Moltenberg et al., 1986), which are significantly more active in oat than in other cereals like barley or wheat (O’ Connor et al., 1992). The high lipolytic activity causes the rapid release of free fatty acids, which are further subjected to oxidation, leading to an increase of rancidity (Moltenberg et al., 1986). Flaking is the typical processing for oat. It includes steam stabilization to inactivate enzymes, followed by kiln- or drum-drying to generate flavor compounds (Klensporf and Jeleń, 2008). The use of cereals as ingredient for beverage has been largely proposed in literature, but the selection of suitable fermentation conditions and starters, were only partially investigated. This study aimed at manufacturing and characterizing the physical, chemical, functional and sensory properties of a non-alcoholic yogurt-like beverage made with oat flakes, which were 4

ACCEPTED MANUSCRIPT subjected to fermentation by LAB. Different starters, technology options and enzyme preparations

T

were assessed.

IP

2. Materials and methods

SC R

2.1 Bacterial strains, culture media and enzymes

Lactobacillus plantarum LP01, LP06, LP09, LP32, LP39, LP40, LP48 and LP51; Lactobacillus casei LC10, LC11 and LC03; and Lactobacillus paracasei LPC02 and LPC16 (Sacco Srl,

NU

Cadorago, CO, Italy) were singly used as starters for fermentation. LAB were cultivated in

MA

modified MRS (mMRS), prepared with 1% (wt/v) maltose and 5% (v/v) fresh yeast extract, final pH of 5.6 (Oxoid Ltd, Basingstoke, Hampshire, England). Fresh yeast extract was prepared by resuspending baker’s yeast (60 g) in deionized water (300 ml). After sterilization (120°C for 20 min),

TE

D

the suspension was centrifuged at 6,000 x g for 20 min, and the supernatant was recovered and added to mMRS prior to sterilization (Coda et al., 2010).

CE P

Preliminarily, different options to inoculate LAB were considered: (i) lyophilized preparation (ca.5 x 1011 cfu/g) (100 mg/L of beverage); (ii) cell suspension in the tap water used for making

AC

beverages, after cultivation in mMRS (late exponential phase of growth, ca. 10 h), harvesting (centrifugation at 9,000 x g for 10 min at 4°C) and washing (twice in 50 mM pH 7.0 phosphate buffer, 4°C); and (iii) pre-culture (24 h of incubation at 30°C) in oat flakes (25% wt/wt in tap water) to be inoculated (5% wt/wt) into beverage. For all these conditions, the cell density after the inoculum was 5 x 107 cfu/ml of beverage. The enzyme preparations Depol 740L (xylanase activity, 55 µkat/g; endoglucanase activity, 3.6 µkat/g; ferulic acid esterase activity, 0.12 µkat/g; βglucanase activity, 27.7 µkat/g; Biocatalyst Ltd., Great Britain) and Grindamyl 1000 (amylase activity, 120 µkat/g; Danisco, Denmark) were used (Anson et al., 2009).

2.2 Manufacture of cereal beverages

5

ACCEPTED MANUSCRIPT Oat flakes (OF) were supplied by Barilla Spa (Parma, Italy). The characteristics of OF were as follows: moisture, 11.3%; protein (N x 5.70), 9.8% of dry matter (d.m.); fat, 5.5% of d.m.; ash, 2.9

T

% of d.m; and carbohydrates, 70.4% of d.m, respectively.

IP

Oat flakes were milled by Ika M20 Universal Mill (Sigma Chemical Co., Milan, Italy) to obtain oat

SC R

flakes flour (OFF). The amount of OFF used for making oat flakes beverages (OFB) was set up based on sensory analysis. LAB were used alone or in combination with Depol 740L (dosage 50 nkat xylanase/g of OFF) and Grindamyl (dosage 75 nkat α-amylase /g of OFF). OFB, without

NU

microbial inoculum, was incubated under the same conditions and used as the control (Ct-OFB). All After fermentation

MA

beverages were incubated at 30°C under stirring conditions (100 rpm). beverages were pasteurized at 63°C for 30 min and stored at 4°C for 30 days.

TE

D

2.3 Determination of pH, total titratable acidity (TTA) and kinetics of acidification The values of pH were determined on-line by a pHmeter (Model 507, Crison, Milan, Italy) with a

CE P

food penetration probe. Total titratable acidity (TTA) was determined on 10 g of beverage, which were homogenized with 90 ml of distilled water, and expressed as the amount (ml) of 0.1 M NaOH

AC

to get pH of 8.3.

Kinetics of growth and acidification were determined and modelled in agreement with the Gompertz equation, as modified by Zwietering et al. (1990),: y= k + A exp{- exp[(μmax or Vmax e/A)(λ-t) + 1]}; where y is the growth expressed as log cfu/g or the acidification rate expressed as dpH/dt (units of pH) at the time t; k is the initial level of the dependent variable to be modelled (log cfu/g or pH units); A is the cell density or pH (units) variation (between inoculation and the stationary phase);

max

or Vmax is the maximum growth rate expressed as

maximum acidification rate expressed as dpH/h, respectively;

log cfu/g/h or the

is the length of the lag phase

measured in hours. The experimental data were modelled by the non-linear regression procedure of the Statistica 8.0 software (Statsoft, Tulsa, USA).

6

ACCEPTED MANUSCRIPT 2.4 Microbiological analysis The number of presumptive LAB was estimated by plating on mMRS agar (Oxoid) at 30°C for 48

T

h. Total bacteria were determined on Plate Count Agar (PCA, Oxoid) at 30°C for 48 h, yeasts were

IP

counted on Yeast extract-Peptone-Dextrose agar (YPD, Oxoid), supplemented with 150 ppm

Glucose Agar (VRBGA, Oxoid) at 37°C for 24 h.

SC R

chloramphenicol, at 30°C for 72 h, and total enterobacteria were determined on Violet Red Bile

NU

2.5 Water holding capacity, water activity, viscosity, total dry matter, and color

MA

Water holding capacity (WHC) was measured according to the method described by Remeuf et al. (2003). After fermentation and storage, 10 g of OFB were centrifuged (5,000 rpm for 40 min at 7 °C). The expelled water was removed and weighed. The percentage of WHC was defined according

TE

D

to the equation: WHC = [(Sample weight - Expelled water) / Sample weight] * 100. Water activity (aw) was determined at 25ºC by the Aqualab Dew Point 4TE water activity meter (Decagon Devices

CE P

Inc., USA).

An aliquot (100 ml) of OFB held at 30°C was used for viscosity measurements. A rotary

AC

viscometer Myr VR3000 (Viscotech, El Vendrell, Spain) model V1-L at disc spindle speeds of 30, 40 and 60 rpm was used. Readings were taken after 3 min of revolution. The appropriate disc spindle was selected so that the torque readings were not below 10% of the total scale. Total dry matter was determined on 100 ml of beverage at 105°C for 24 h (AOAC, 1985). The chromaticity co-ordinates of the beverages (obtained by a Minolta CR-10 camera) were reported in the form of a color difference, dE*ab, as follows: dE*ab = (dL) 2  (da) 2  (db) 2 where dL, da, and db are the differences for L, a, and b values between sample and reference (a white ceramic plate having L = 93.4, a = –1.8, and b = 4.4).

2.6 Organic acids, ethanol and free amino acids

7

ACCEPTED MANUSCRIPT Water/salt-soluble extracts (WSE) from beverages were prepared following the method of Weiss et al. (1993). An aliquot of beverage (containing 7.5 g of flour) was diluted with 30 ml of 50 mM Tris-

T

HCl (pH 8.8), held at 4°C for 1 h, vortexing at 15-min intervals, and centrifuged at 20,000 x g for

IP

20 min. The supernatant, containing the water/salt-soluble fraction, was filtered through a Millex-

SC R

HA 0.22-m pore size filter (Millipore Co., Bedford, MA) and used for analyses. Organic acids and ethanol were determined by HPLC, using an ÄKTA Purifier system (GE Healthcare) equipped with an Aminex HPX-87H column (ion exclusion, Biorad), a UV detector

NU

operating at 210 nm, and a Perkin Elmer 200a refractive index detector. Elution was at 60°C with a

MA

flow rate of 0.6 ml/min, using 10 mM H2SO4 as the mobile phase (Zeppa et al., 2001). Peaks were identified by comparing elution times and spiking samples with known quantities of standard

D

solutions of acetic and lactic acid and ethanol. Total and individual free amino acids were analysed

TE

by a Biochrom 30 series Amino Acid Analyzer (Biochrom Ltd., Cambridge Science Park, England),

(2008).

CE P

with a Na-cation-exchange column (20 by 0.46 cm inner diameter) as described by Rizzello et al.

AC

2.7 Total phenols and antioxidant activity The 1,1-diphenyl-2-picrylhydrazyl (DPPH) radical scavenging activity of OFB was determined both on methanol extract (ME) and WSE. Five milliliters of OFB were mixed with 50 ml of 80% methanol to get ME. The mixture was purged with nitrogen stream for 30 min, under stirring condition, and centrifuged at 4,600 × g for 20 min. ME were transferred into test tubes, purged with nitrogen stream and stored at ca. 4°C before analysis. The concentration of total phenols was determined as described by Slinkard and Singleton (1997). It was expressed as gallic acid equivalent. The free radical scavenging capacity was determined using the stable 2,2-diphenyl-1picrylhydrazyl radical (DPPH˙), as reported by Yu et al. (2003). The reaction was monitored by reading the absorbance at 517 nm every 2 min for 30 min. A blank reagent was used to verify the stability of DPPH˙ over the test time. The absorbance value measured after 10 min was used for the 8

ACCEPTED MANUSCRIPT calculation of the µmoles DPPH˙ scavenged by extracts. The absorbance value in the presence of the extract was also determined over 30 min and compared with 75 ppm butylated hydroxytoluene

T

(BHT) as the antioxidant reference.

IP

The scavenging effect of freeze dried WSE on DPPH˙ free radical was measured according to the

SC R

method of Shimada et al. (1992), with some modifications. Freeze-dried samples were first dissolved in 0.1 M phosphate buffer pH 7.0 at the final concentration of 1 mg/ml of peptides, and then 2 ml were added to 2 ml of 0.1 mM DPPH, which was dissolved in 95% ethanol. The mixture

NU

was shaken and left at room temperature for 30 min. The absorbance was read at 517 nm. The

MA

absorbance measured after 10 min was used for the calculation of the DPPH scavenged by WSE (Rizzello et al., 2010). More the absorbance was low, higher it was the DPPH scavenging activity. The scavenging activity was expressed as follows: DPPH scavenging activity (%) = [(blank

CE P

antioxidant reference.

TE

D

absorbance – sample absorbance) / blank absorbance] x 100. BHT (1 mg/ml) was used as the

2.8 Determination of dietary fiber concentration

AC

Pepsin solution was prepared dissolving 100 mg of pepsin (2000 FIP-U/g Sigma Chemical Co.) per ml of 0.08 M HCl-KCl buffer, pH 1.5. Pancreatin solution was prepared dissolving 5 mg of pancreatin (Sigma Chemical Co.) per ml of 0.1 M of phosphate buffer pH, 7.5. α-Amylase solution was prepared dissolving 120 mg of α-amylase (17.5 IU amylase/mg, Sigma Chemical Co.) in 1 ml of 0.1 M Trizma–maleate buffer, pH 6.9. The amyloglucosidase solution (14 IU/mg, Roche, Mannheim, Germany) in 0.2 M sodium acetate buffer, pH 4.75, was used. The concentration of fiber was determined as described by Goñi et al. (2009). In detail, 100 ml of beverages, concentrated 2.5 fold with a Speed-Vac centrifuge at 35°C (Thermo Scientific, Waltham, MA), were submitted to enzymatic hydrolysis. First, samples were treated with 20 ml of HCl–KCl buffer pH 1.5 (pH checked) and 0.5 ml pepsin solution. After 40 min at 40°C in water bath with constant shaking, 20 ml Trizma–maleate buffer pH 6.9 (pH checked) and 5 ml α-amylase 9

ACCEPTED MANUSCRIPT solution were added. Samples were incubated at 37°C, for 3 h. Finally, sodium acetate buffer pH 4.7 was added followed by 0.4 ml of amyloglucosidase solution, and incubation was at 60°C for 45

T

min under constant shaking. After enzyme treatments, samples were transferred into dialysis tubes

IP

(cut off 12,400 Da) and dialyzed against water for 48 h at 37°C. Retentates were freeze-dried for

SC R

gravimetric determination of dietary fibers.

β-glucans were determined with the K-BGLU kit (Megazyme International, Bray, Ireland) following the manufacturer’s instructions. Briefly, mixed-linkage β-glucans, [(1-3),(1-4)-β-D-

NU

glucans], were subjected to selective degradation of the (1-3) linkage using lichenase (a specific

MA

endo-(1-3),(1-4)-β-D-glucan 4-glucanohydrolase) β-glucosidase, and glucose oxidase (McCleary and Glennie-Holmes, 1985).

TE

D

2.9 In vitro starch hydrolysis

The analysis of starch hydrolysis mimicked the in vivo digestion of starch. The analysis was carried

CE P

out as described by Liljeberg et al. (1996). Aliquots of beverages, containing 1 g of starch, were given in randomized order to 10 volunteers. Subjects rinsed their mouths with tap water and subsequently

AC

with beverages for 15 s (approximately 15 times). Samples were then expectorated into a beaker containing 50 mg of pepsin in 6 ml of 0.05 M Na, K-phosphate buffer (containing 0.4 g/l NaCl) adjusted to pH 1.5 with 2M HCl. Finally, the subjects rinsed their mouths with 5 ml of 0.05 M Na, Kphosphate buffer (pH 6.9) for 60 s and expectorated the rinsing solution into the beaker. The contents were stirred and pH adjusted to 1.5. Each sample was incubated at 37°C for 30 min with gentle mixing three times during incubation. The pH was re-adjusted to 6.9 before incubation with porcine pancreatin α-amylase (110 U, Sigma, St Louis, U.S.A.). The sample was brought to volume (30 ml) with 0.05 M Na, K-phosphate buffer, and transferred to the dialysis tubing (cut off 12400). Each bag was incubated at 37°C for 3 h in a beaker containing 0.05 M Na, K-phosphate buffer (800 ml) and placed in a stirred water bath. Every 30 min, aliquots (2 ml) from the dialysate were removed and refrigerated. Then, 3 ml of 0.4 M sodium acetate buffer (pH 4.75) were added to each aliquot, and 80 µl of 10

ACCEPTED MANUSCRIPT amyloglucosidase (140 U/ml) were used to hydrolyze the digested starch into glucose after 45 min at 60°C in a shaking water bath. The glucose content was measured with Enzy Plus D-Glucose kit

T

(Diffchamb Västra Frölunda, Sweden). Factor conversion from glucose to starch was 0.9. The rate of

IP

starch digestion was expressed as the percentage of potentially available starch hydrolyzed at different

SC R

times (30, 60, 90, 120 and 180 min). The hydrolysis curves were obtained with the equation described below, using the software STATISTICA 7.0. Hydrolysis curves follow a first order equation: C = C ∞ (1-e-kt), where C is the concentration at t time, C∞ is the equilibrium concentration, k is the kinetic

MA

NU

constant and t is the chosen time (De Angelis et al., 2007).

2.10 Sensory analyses

Two different protocols were used for sensory analysis. The first protocol (Luckow et al., 2006)

TE

D

used 10 non trained panellists. Acidic, cereal and sweet were considered as sensory attributes for flavor and taste, using a scale from 0 to10. The evaluation of sensory attributes was discussed

CE P

with the assessors during the introductory training sessions and with the aim to select the fermentation conditions for making OFB, 7.0 was considered as the optimal score. Lower or

AC

higher values than 7.0 were considered weakly or excessively pronounced, respectively. The second protocol (Lapveteläinen and Rannikko, 2000; Luckow and Delahunty, 2004) considered a vocabulary for odor and flavor attributes. References that could be used to remind panellists about the quality of each attribute were identified (Table 1). The descriptive sensory analysis was carried out once the training was completed. Beverages were served in white polystyrene cups (40 ml in a 120 ml cup), and were labeled randomly with selected codes. Beverages were served at room temperature (20°C) to better differentiate odors and flavors, and to facilitate the characterization and comparison of each sample. Each assessor received 2 samples for each beverage; 3 independent experiments were carried out.

2.11 Statistical Analysis 11

ACCEPTED MANUSCRIPT Data were subjected to one-way ANOVA; pair-comparison of treatment means was achieved by Tukey’s procedure at P0.05, using the statistical software Statistica for Windows (Statistica 7.0,

IP

T

Windows).

SC R

3. Results 3.1 Selection of starters and beverages

Preliminarily, different percentages (5 to 40%, wt/wt) of oat flakes flour (OFF) were mixed with tap

NU

water, homogenized and assessed based on sensory perception to get a yogurt-like texture. Amount

MA

of OFF lower than 20% (wt/wt) gave a too liquid matrix, allowing the fast separation of the solid particles to the bottom. Percentages higher than 30% (wt/wt) gave matrixes similar to semisolid

D

dough. The percentage of 25% (wt/wt) was the optimum and it was used for further experiments.

TE

All LAB strains were singly assayed as starters, according to the protocol described above. The screening was based on acidification and sensory properties of the beverages (Table 2). Only

CE P

lyophilized preparations were used. After 12 h of fermentation at 30°C, the value of pH decreased from 6.45 ± 0.21 to 4.23 ± 0.11 - 5.12 ± 0.15. Strains that caused the highest values of ΔpH were L.

AC

plantarum LP09, LP39, LP40 and LP48 (2.21 ± 0.04, 2.18 ± 0.03, 2.25 ± 0.02 and 2.03 ± 0.03, respectively). The rapidity of acidification was considered a discriminant technological property. Indeed, the other species and strains did not reach values of pH lower than 4.7 (corresponding to ΔpH of 1.75) and were excluded from sensory analysis. The highest scores for acidic flavour and taste were found for the beverages fermented with strains LP40 and LP48 (Table 2). These beverages also showed the lowest scores for sweet and cereal attributes. The beverage fermented with strain LP39 had the highest value for sweet taste, and suboptimal scores for all the other attributes. Considering 7.0 as the optimal score, the beverage started with L. plantarum LP09 had optimal values for all sensory attributes and the most balanced profile. This strain was selected for further analyses. The beverage started with the lyophilized preparation

12

ACCEPTED MANUSCRIPT of L. plantarum LP09 was compared to beverages that were fermented using cells pre-cultivated in mMRS broth or pre-cultured in oat flakes.

T

No significant (P>0.05) differences were found regarding the values of ΔpH. Nevertheless, cells

IP

from mMRS broth or pre-cultured in oat flakes negatively affected the sensory profile of the

SC R

beverage since the score for acidic taste and flavor significantly (P0.05) differ among the three beverages (Table 3). On the contrary, the values of aw of LP09-OFB and E-OFB were slightly higher (P