Food Chemistry 157 (2014) 148–156

Contents lists available at ScienceDirect

Food Chemistry

journal homepage: www.elsevier .com/locate / foodchem

Relation of sensory perception with chemical compositionof bioprocessed lingonberry

http://dx.doi.org/10.1016/j.foodchem.2014.02.0300308-8146/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Tel.: +358 20 722 4457; fax: +358 20 722 7071.E-mail address: Riitta.puupponen-pimia@vtt.fi (R. Puupponen-Pimiä).

Kaarina Viljanen, Raija-Liisa Heiniö, Riikka Juvonen, Tuija Kössö, Riitta Puupponen-Pimiä ⇑VTT Technical Research Centre of Finland, P.O. Box 1000, FI-02044 VTT, Finland

a r t i c l e i n f o

Article history:Received 10 October 2013Received in revised form 23 January 2014Accepted 5 February 2014Available online 14 February 2014

Keywords:LingonberrySensoryFlavourSPME–GC/MSFermentationEnzyme

a b s t r a c t

The impact of bioprocessing on lingonberry flavour was studied by sensory evaluation and chemical anal-ysis (organic acids, mannitol, phenolic compounds, sugars and volatile compounds). Bioprocessing of lin-gonberries with enzymes, lactic acid bacteria (LAB) or yeast, or their combination (excluding pure LABfermentation) affected their perceived flavour and chemical composition. Sweetness was associated espe-cially with enzyme treatment but also with enzyme + LAB treatment. Yeast fermentation caused signifi-cant changes in volatile aroma compounds and perceived flavour, whereas minor changes were detectedin LAB or enzyme-treated berries. Increased concentration of organic acids, ethanol and some phenolicacids correlated with perceived fermented odour/flavour in yeast fermentations, in which increase inbenzoic acid level was significant. In enzymatic treatment decreasing anthocyanins correlated well withdecreased perceived colour intensity. Enzyme treatment is a potential tool to decrease naturally acidicflavour of lingonberry. Fermentation, especially with yeast, could be an interesting new approach toincrease the content of natural preservatives, such as antimicrobial benzoic acid.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Lingonberry (Vaccinium vitis-idaea) is one of the most importantharvested wild berries in Finland. However, only a small amount ofberries is picked and used by the food industry (Turtiainen, Salo, &Saastamoinen, 2011). One of the hindrances for wider use oflingonberries is the intense, sour or even bitter flavour. The flavourof lingonberries is typically very acidic and sour due to highamount of organic acids, especially citric acid (Mikulic-Petkovsek,Schmitzer, Slatnar, Stampar, & Vebeic, 2012; Viljakainen, Visti, &Laakso, 2002). The high amount of organic acids leads a pH of be-low 4. Lingonberries also contain relatively high amount of sugarsbut the sweetness is masked by acids. Besides organic acids, theconcentration of benzoic acid is high in lingonberries (Viljakainenet al., 2002). Benzoic acid is one of the most used chemical preser-vatives in foods. It has been shown to inhibit both the growth ofyeast and bacteria at low pH (Brul & Coote, 1999). The concentra-tion of flavour-active compounds, such as sugars, organic acids,phenolic compounds, volatile compounds, and hence the flavourof lingonberries is also affected by season, maturity and geographicorigin.

Neither the sensory quality nor chemical composition of lingon-berries has been widely studied. Concerning the chemical composi-tion, most studies have focused on identification of phenoliccompounds (e.g., Ek, Kartimo, Mattila, & Tolonen, 2006; Kylliet al., 2011), organic acids and sugars (e.g., Mikulic-Petkovseket al., 2012; Viljakainen et al., 2002). The composition of volatilecompounds has only been studied from the extracted essential oil(Anjou & von Sydow, 1967). In contrast to the chemical composi-tion, the antioxidant activity of lingonberries has received muchattention (e.g., Kähkönen, Heinämäki, Ollilainen, & Heinonen,2003; Viljanen, Kylli, Hubbermann, Schwarz, & Heinonen, 2005;Viljanen, Kylli, Kivikari, & Heinonen, 2004). Lingonberries havebeen shown to inhibit very effectively lipid and LDL oxidation in dif-ferent model systems. They are also good free radical scavengers.

Bioprocessing with fermentation provides a potential tool toprepare food products with special sensory and chemical proper-ties (Bourdichon et al., 2012). Food fermentations can be carriedout spontaneously or by using microbial starter cultures. Howeverlingonberries are a challenging material for fermentation due toacidity and high concentration of antimicrobial phenolic constitu-ents, such as benzoic acid, which is active at low pH values(Viljakainen & Laakso, 2002). As a result attempts to modify theacidic mouthfeel of lingonberry juice using malolactic fermenta-tion with Oenococcus oeni (malic acid converted to less acidic lacticacid) was not successful (Viljakainen & Laakso, 2002). The

http://crossmark.crossref.org/dialog/?doi=10.1016/j.foodchem.2014.02.030&domain=pdfhttp://dx.doi.org/10.1016/j.foodchem.2014.02.030mailto:Riitta.puupponen-pimia@vtt.fihttp://dx.doi.org/10.1016/j.foodchem.2014.02.030http://www.sciencedirect.com/science/journal/03088146http://www.elsevier.com/locate/foodchem

K. Viljanen et al. / Food Chemistry 157 (2014) 148–156 149

fermentation of lingonberries usually requires the reduction of theantimicrobial constituents and the addition of sugar (Visti,Viljakainen, & Laakso, 2003).

The aim of the study was to investigate how bioprocessing withdifferent microbes and enzymes affect the sensory characteristicsof lingonberries and to correlate the findings with chemical com-position of the material. Our approach was not to add any additivesto fermentation, but rather use microbial strains, which were ableto utilise the berry raw-material as such. In addition to chemicalanalysis of non-volatile and volatile flavour-active compounds,descriptive sensory analysis was used to gain understanding onhow different microbes and enzymes affect perceived flavour.

2. Materials and methods

2.1. Berry material

Frozen, ripe lingonberries (Vaccinium vitis-idaea) of Finnishorigin were obtained from a local distributor, Pakkasmarja Ltd.(Suonenjoki, Finland). The lingonberries were frozen and storedat �20 �C until use.

2.2. Bioprocessing of lingonberries

Six types of lingonberry samples were studied: (1) referencematerial without bioprocessing, (2) yeast fermented material, (3)LAB fermented material, (4) yeast and LAB fermented material,(5) enzyme treated material, and (6) enzyme-treated and LAB fer-mented material.

For microbial fermentations, frozen lingonberries and ultra-pure water were mixed together (1:1) and heated at 80 �C for5 min. Subsequently the mixture was cooled in an ice bath and ber-ries were crushed by a sterile potato masher. The pH of the mixturewas adjusted to pH 5.0 with 5 N sodium hydroxide. The lingon-berry fermentations were started either with Lactobacillus planta-rum VTT E-78076 (LAB) or Hanseniaspora uvarum VTT C-11885(yeast), or with both microbes. Our preliminary experimentsindicated that these strains are able to grow in the lingonberrymaterial (data not shown). The microbes were pre-grown infood-grade media. For LAB general edible medium (Saarela et al.,2004) and 24 h incubation at 30 �C in anaerobic conditions wasused. Yeast was grown in 0.5 l wort sucrose broth (yeast extract1.0 g, sucrose solution (10% w/v) and 0.5 l pretreated wort (12%;www.culturecollection.fi) for 24 h at 25 �C with 100 rpm shaking.The cells were collected from the cultures and washed with Ring-eŕs solution (Merck, Darmstadt, Germany) before use. The lingon-berry material was inoculated with approximately 106 cfu/g ofLAB and/or with approximately 107 cfu/g of yeast. The fermenta-tions (8 kg each) were carried out in a 15-l capacity bioreactorfor 3 days at 30 �C (LAB) or 7 days at 25 �C (other) under constantmixing (130 rpm). LAB fermentations were purged with sterilenitrogen gas to create anaerobic conditions. The viable counts ofLAB and yeasts were measured before and after the fermentationsusing plate count technique as previously described and expressedas colony-forming units (CFU) per gram of wet weight (w.w.)(Katina et al., 2007). The pH value was measured with a standardpH meter. All analyses were performed from triplicate samples.

For the enzymatic treatment frozen lingonberries were thawedin cold-storage room (+4 �C) over night and crushed with a potatomasher. The berry mash was heated for 10 min in a mixing reactor(45 �C), after which the enzyme mixture (Biopectinase Super 8�,100 nkat/g and Acid Protease A, 0.1% w/w) was added. The berrymash was incubated by continuous mixing at 45 �C for 4 h (=enzy-matic treatment). For LAB fermentation, the enzyme-treated mate-rials were mixed with water, heat-treated and pH adjusted as for

previous fermentations. Enzyme-treated material inhibited thegrowth of the yeast and was therefore not studied. Bioprocessedlingonberry materials were stored frozen prior to use in sensoryevaluation and chemical analyses.

2.3. Sensory analysis

Frozen bioprocessed lingonberry samples were thawed beforesensory analysis. Enzyme-treated sample and ultra-pure waterwere mixed together (1:1) and heated at 80 �C for 5 min. After thatthe mixture was cooled in an ice bath and the pH of the mixturewas adjusted to 5.0 with 5 N sodium hydroxide.

Descriptive sensory analysis (Lawless & Heymann, 2010) wascarried out at the sensory laboratory of VTT, which fulfils therequirements of the ISO standards (ISO 1985; ISO 1988). The sen-sory panel consisted of 11 trained assessors skilled in the sensoryassessment of diverse plant-based samples. The vocabulary of thesensory attributes was developed by describing differences be-tween various bioprocessed lingonberry samples. The evaluatedattributes of the lingonberry samples were: fresh, lingonberryintensity, sour and fermented odour and flavour, sweetness, bitter-ness, colour redness and clarity, thickness, amount of berry pieces,and intensity of possible off-taste. The attribute intensities wererated on continuous unstructured, graphical intensity scales. Thescales were 10 cm in length and verbally anchored at each end,the left side of the scale corresponding to the lowest intensity (va-lue 0) and the right side to the highest intensity (value 10) of theattribute. The panel members practiced the evaluation of attributedefinitions and their intensities during training sessions. The defi-nitions of the descriptors served as a guide for the subjects duringtesting to minimise confusion over the meaning of each attribute.Panellists had also the possibility to describe verbally the charac-teristics of the samples.

The berry samples were blind-coded by using 3-digit numbersand presented to the trained assessors in random order from 50-mL disposable plastic containers covered by lids. The lingonberrysamples were evaluated crushed. Water was served to the asses-sors for cleansing the palate between the samples. The sampleswere judged in two replicate sessions. The scores were recordedand collected using a computerized data system (Compusense Five,Ver 5.4.15, CSA, Computerized Sensory Analysis System; Compu-sense Inc., Guelph, ON, Canada).

2.4. Analysis of non-volatile flavour-active compounds

The non-volatile flavour-active compounds, such as organicacids, mannitol, phenolic compounds and sugars were analysedby HPLC. Non-volatile chemical compounds were extracted fromfreeze-dried fermented lingonberries with methanol (100 mg offreeze-dried material/2 mL methanol; HPLC grade, RathburnChemicals, Ltd., Walkerburn, UK).

2.4.1. Phenolic compoundsPhenolic compounds were determined by using an analytical

HPLC method modified from methods described by Aaby, Ekeberg,and Skrede (2007), and Määttä-Riihinen, Kamal-Eldin, andTörrönen (2004). The HPLC system consisted of a Waters 600S sys-tem controller pump, Waters 717plus autosampler and Waters2996 series photodiode array detector (Waters Corporation,Milford, MA). Analytes were separated using a Hypersil BDC C-18(150 � 4.6 mm, 5 lm; Agilent, Santa Clara, CA) reverse-phase col-umn. The data was processed by Waters Empower Pro chromatog-raphy software. The solvents used for analyses of lingonberryphenolic compounds in the gradient program were 5% formic acidin water and 100% acetonitrile (Rathburn Chemicals Ltd.). Totalanalysis time was 45 min. Selected phenolic compounds were

http://www.culturecollection.fi

150 K. Viljanen et al. / Food Chemistry 157 (2014) 148–156

identified and quantified on the basis of corresponding standards(mg/g of dry weight).

2.4.2. Glucose, fructose and sucroseThe concentration of sugars (glucose, fructose and sucrose)

were analysed by high-performance anion exchange chromatogra-phy (HPAEC) (Dionex ICS-3000) with pulse amperometric detec-tion (PAD) (Dionex Corporation, Sunnyvale, CA). The pre- andseparation columns (Dionex CarboPac PA-1) were employed at30 �C with a flow rate of 1 mL/min using the following eluents:milli-Q water, 100 mM NaOH, 300 mM Na acetate/100 mM NaOH,and 300 mM NaOH. The eluant gradients were adapted fromTenkanen and Siika-aho (2000).

2.4.3. Organic acids and mannitolThe separation of organic acids and mannitol was carried out on

an Aminex HPX-87H column (300 � 7.8 mm; BioRad, Hercules, CA)with a Waters 2690 HPLC separation module and Alliance auto-sampler, Waters 410 refractive index and Waters 2487 UV detec-tors, and Waters Empower Pro chromatography software for dataprocessing (Waters Corporation). The solvent used in the isocraticelution was 2.5 mM sulphuric acid. On the basis of the correspond-ing standard compounds, lactic acid, citric acid, acetic acid andmannitol were quantified (mg/g of dry weight).

2.5. Analysis of volatile odour compounds

For analysis of volatile compounds, 2 g of lingonberry sampleswere weighed into 20-mL headspace vials containing 100 lL of sat-urated NaCl solution. Toluene was used as internal standard(87 ng/sample). Samples were analysed by SPME–GC/MS accordingto Viljanen, Lille, Heiniö, and Buchert (2011). Samples were pre-incubated at 35 �C for 30 min in closed headspace vials. Extractionof volatile compounds was done at 35 �C for 60 min with a precon-ditioned (300 �C, 1 h) 75 lm Carboxen/PDMS SPME-fibre (Supelco,Bellefonte, PA). After extraction the analytes were desorbed for5 min at 260 �C in the splitless injector (split flow 19.4 mL/min)of the gas chromatograph (Agilent 6890 Series, USA) combinedwith an MS detector (Agilent, 5973Network MSD, USA) and SPMEautosampler (Combipal; CTC, Zwingen, Switzerland). Analyteswere separated on an Ultra-2 capillary column (60 m � 0.25mm � 1 lm; Agilent Technologies, USA), with a constant flow of1.5 mL/min, using helium as carrier gas. The temperature pro-gramme started at 45 �C with 3 min holding time, then increased10 �C/min up to 100 �C, followed by 5 �C/min increase up to150 �C and finally a 10 �C/min increase up to 300 �C, where thetemperature was kept for 9 min. MSD was operated in electron-im-pact mode at 70 eV, in full scan m/z 40–550. The ion source tem-perature was 230 �C and the interface was 280 �C. Compoundswere tentatively identified by comparing their mass spectra withthose in the Palisade Complete 600 K Mass Spectral Library (Pali-sade Corporation, Newfield, NY). Each sample was analysed fourtimes and normalised peak areas were expressed with respect tointernal standard (compound area/ISTD area).

For ethanol measurement, 500 mg of sample were weighed intoa 20-mL headspace vial. 1-Butanol was used as internal standard.Ethanol was extracted from the headspace using a Tekmar 7000autosampler (Teledyne Tekmar, Mason, OH) at 80 �C for 20 min.The autosampler was combined with an Agilent 6890 gas chro-matograph and analytes were separated on a Solgel-WAX column(30 m � 0.33 mm; 0.5 lm) (SGE, Ringwood, Australia). Ethanolwas detected with a flame ionisation detector at 250 �C. The inlettemperature was 220 �C and split ratio was 0.6:1. The gas chro-matograph was operated in constant flow mode with helium ascarrier gas at 1.7 mL/min. The oven was programmed from 60 �Cto 200 �C with total run time of 12 min. Amount of ethanol (mg/g

of sample) in samples was calculated by using ChemStationsoftware (Agilent) with standard curve prepared from AAS gradeethanol (Alko Oy, Helsinki, Finland). Each sample was analysed intriplicate and the mean of these values was used in furthercalculations.

2.6. Statistical data analysis

The chemical composition data were analysed using analysis ofvariance (ANOVA) and Tukey’s Honestly Significant Difference(HSD) test (significance of differences at p < 0.05) (SPSS 19.0 forWindows, SPSS Inc., Chicago, IL).

Means of the sensory raw data (i.e. scores given by the asses-sors) obtained from both sensory sessions were calculated. The sig-nificance of each descriptive attribute in discriminating betweenthe samples was analysed using ANOVA and Tukey’s HSD test (sig-nificance of differences at p < 0.05; SPSS 19.0 for Windows). Whenthe difference in ANOVA among the samples was statistically sig-nificant, pairwise comparisons of these samples were performedusing Tukey’s test.

The intensities of the perceived attributes of the sensory profileand the non-volatile and volatile chemical compounds of the ling-onberry samples were related statistically by PLS (partial leastsquares) regression using Unscrambler software (Unscrambler9.8, CAMO Software AS, Oslo, Norway). In the PLS regression onlyvolatile compounds that were found to be significantly differentfrom reference (untreated lingonberries) sample were used. Thesensory data used for the multivariate analysis was means calcu-lated over all panellists, and the means of the levels of the chemicalcompounds were used for the PLS regression. Both the sensory andchemical data were standardised, and the results of the chemicalcompounds were normalised before the multivariate analysis inUnscrambler. The model was validated by cross-validation. PLSregression is specifically designed to determine relationships exist-ing between blocks of dependent (Y sensory) and independent(X instrumental) variables by seeking underlying factors commonto both sets of variables (Martens & Naes, 1998).

3. Results and discussion

3.1. Viable cell counts and pH

After the 7-day yeast fermentation (H. uvarum), the viable yeastcounts in the lingonberry material were slightly lower than at thebeginning (1.1 � 107 ± 1.7 � 104 and 6.7 � 106 ± 4.6 � 105, respec-tively). During the 3-day fermentation of the lingonberry materialwith LAB, an approximately 10-fold increase in viable counts ofLAB was detected (from 9.9 � 105 ± 1.7 � 104 to 6.7 � 106 ±4.6 � 105 cfu/g). The lingonberry material that was co-fermentedwith LAB and yeast contained approximately 2 � 107 cfu/g of bothLAB and yeast at the end of the 7-day fermentation. In the enzyme-treated lingonberry material the viable counts of LAB showedapproximately 10-fold decrease during the 3-day fermentation.Thus, the enzyme-treated lingonberry material appeared to begrowth-inhibitory to this strain.

The final pH values in all lingonberry materials varied from 4.6to 4.9. Hence the changes in pH values during the fermentationswere negligible, ranging from 0.05 to 0.1 pH units (data notshown).

3.2. Sensory evaluation of lingonberry

The sensory profile of lingonberry samples showed that pro-cessing caused significant changes in all other assessed attributesexcept in sourness odour (Fig. 1). The enzyme treatment alone

Fig. 1. Sensory profile of lingonberry samples evaluated by a trained panel (n = 2 � 11).

K. Viljanen et al. / Food Chemistry 157 (2014) 148–156 151

and together with LAB fermentation decreased the redness of ling-onberry, colour clarity and thickness. In another study enzymatictreatment has been shown to influence sensory profile and chem-ical composition of blackcurrant juice by significantly improvingjuice yield, but simultaneously increasing fermented and astrin-gent notes (Laaksonen et al., 2012). The largest and most abundantberry pieces were found in reference and LAB-fermented lingon-berries. In addition, the freshness (odour/flavour), and lingonberryflavour/odour was most intense in the reference and LAB-fer-mented lingonberries. As expected, fermented odour and flavour,as well as bitter and sour taste, were most noticeable in yeastand LAB + yeast fermented lingonberries. Bitterness increased afterenzymatic treatment but not as much as after fermentation. On theother hand, the sweet taste was least intense in these samples.Moreover, yeast and LAB + yeast fermentation caused most intenseoff-taste for the samples.

3.3. Non-volatile flavour-active compounds in lingonberries

Selected non-volatile and volatile chemical compounds wereanalysed with the aim to identify chemical flavour-active com-pounds responsible for the changes in perceived flavour. The

Table 1Amount (mg/g of dry weight; average ± SD) of non-volatile flavour-active compounds in f

Ref Enzymatic Yeast LAB

Organic acidsCitric acid 99.6 ± 7.53b 140 ± 3.63c 103 ± 3.45b 109 ± 7Lactic acid nda nda 2.91 ± 2.57a 18.05 ± 1Acetic acid nda nda 5.18 ± 3.50b nda

Sugars and sugar alcoholsSucrose 23.0 ± 0.01c 0.80 ± 0.00a 15.0 ± 0.00b 20.9 ± 0Glucose 248 ± 0.03b 253 ± 0.01b 12.8 ± 0.00a 246 ± 0Fructose 260 ± 0.01b 267 ± 0.04b 11.8 ± 0.01a 258 ± 0Mannitol nda nda 171 ± 15.7b nda

Phenolic compoundsp-Coumaric acid 0.13 ± 0.00a 0.02 ± 0.00a 0.37 ± 0.00b 0.11 ± 0Caffeic acid 0.01 ± 0.00a 0.03 ± 0.00a 0.10 ± 0.00a 0.01 ± 0p-Hydroxybenzoic acid 0.71 ± 0.00a 0.96 ± 0.00a 4.41 ± 0.00b 0.88 ± 0Benzoic acid 6.69 ± 0.01a 10.1 ± 0.00b 14.2 ± 0.01c 7.24 ± 0Quercetin 0.39 ± 0.01a 3.67 ± 0.00b 0.87 ± 0.00a 0.42 ± 0Anthocyanins 4.34 ± 0.01c 1.91 ± 0.00b 4.09 ± 0.01c 3.26 ± 0

nd, not detected.a–cValues in the same row followed by different letters are significantly different (p < 0.

A Flavour descriptions adapted from Acree and Arn, 2010.

flavour of lingonberries is due to relative amounts and interactionsof many different compounds (Ek et al., 2006; Kylli et al., 2011; Lee& Finn, 2012; Mikulic-Petkovsek et al., 2012; Viljakainen et al.,2002). Microbial fermentations consumed simple sugars with con-comitant formation of ethanol and/or organic acids. The concentra-tion of both glucose and fructose decreased in samples fermentedwith yeast and yeast + LAB by 95% (Table 1). This correlated withformation of ethanol (Table 3). It has been previously reported thatH. uvarum ferments glucose and fructose to ethanol at the samespeed under oxygen limiting conditions (Moreira, Mendes, Guedesde Pinho, Hogg, & Vasconcelos, 2008). Glucose concentration de-creased by 45% in enzyme + LAB treated lingonberries. In addition,the concentration of sucrose decreased by more than 96% in sam-ples treated with enzymes and enzymes + LAB, probably due toacid hydrolysis of sucrose at the high temperature used in the en-zyme treatment.

The fermentation of lingonberries with LAB and yeast alone aswell as together produced small amounts of lactic acid. Possiblycontaminating LAB were responsible for lactic acid production inthe yeast fermentations although LAB could not be detected(

Table 2The identification of flavour-active volatile compounds found in lingonberries with and without fermentation with their odour descriptions.c

Retention time/min (peak number)a Compound Fragments (abundance)b Odour descriptionc

Aldehydes7.35 (4) (E)-2-butenal 41 (100%), 42 (32%), 69 (69%), 70 (85%) Green, fruit8.36 (6) Pentanal 41 (32%), 44 (100%), 57 (31%), 58 (49%) Almond, malt, pungent10.91 (12) Hexanal 41 (62%), 43 (69%), 44 (100%), 56 (80%) Grass, tallow, fat12.09 (15) 2-Furaldehyde 42 (20%), 67 (20%), 95 (100%), 96 (105%) Bread, almond, sweet15.44 (20) (E)-2-Heptenal 41 (100%), 55 (92%), 70 (56%), 83 (92%) Soap, fat, almond16.14 (23) Benzaldehyde 51 (63%), 77 (98%), 105 (94%), 106 (100%) Almond, burnt sugar16.67 (24) Octanal 41 (90%), 43 (100%), 57 (63%), 84 (44%) Fat, soap, lemon, green19.66 (33) Nonanal 41 (100%), 57 (90%), 70 (33%), 82 (16%) Fat, citrus, green

Ketones5.84 (1) Diacetyl 42 (10%), 43 (100%), 86 (21%) Butter8.05 (5) 2-Pentanone 41 (17%), 43 (100%), 58 (10%), 86 (15%) Ether, fruit8.70 (8) 2-Hydroxy-2-butanone 43 (60%), 45 (100%), 88 (8%) Butter, cream15.88 (21) 4-Octen-3-one 44 (100%), 55 (65%), 70 (48%), 97 (10%) Coconut, fruity16.07 (22) 6-Methyl-5-hepten-2-one 43 (100%), 55 (35%), 67 (35%), 108 (30%) Pepper, mushroom, rubber19.25 (31) Acetophenone 51 (38%), 77 (88%), 105 (100%), 120 (32%) Must, flower, almond

Alcohols7.00 (3) Ethanol nd Alcohol9.17 (10) 3-Methyl-1-butanol 42 (91%), 43 (76%), 55 (100%), 70 (62%) Whiskey, malt, burn9.26 (11) 2-Methyl-1-butanol 41 (96%), 56 (88%), 57 (100%), 70 (42%) Malt17.38 (26) 2-Ethyl-1-hexanol 55 (25%), 56 (20%), 57 (100%), 70 (20%) Rose, green18.20 (30) Benzyl alcohol 77 (63%), 79 (100%), 107 (53%), 108 (74%) Sweet, flower19.80 (34) (Z)-3-Hexen-1-ol 41 (76%), 55 (41%), 67 (100%), 82 (36%) Grass20.554 (36) Phenylethyl alcohol 65 (23%), 91 (100%), 92 (54%), 122 (23%) Honey, spice, rose, lilac

Terpenes14.75 (19) a-Pinene 77 (24%), 77 (24%), 91 (23%), 93 (100%) Pine, turpentine17.03 (25) b-Elemene 69 (78%), 91 (28%), 93 (91%), 105 (54%) Herb, wax, fresh17.56 (27) p-Cymene 91 (18%), 117 (10%), 119 (100%), 134 (28%) Solvent, gasoline, citrus17.65 (28) Limonene 68 (100%), 79 (45%), 93 (88%), 121 (26%) Citrus, mint17.92 (29) 1,8-Cineole 43 (100%), 71 (49%), 81 (60%), 108 (42%) Mint, sweet19.48 (32) Linalool 55 (63%), 71 (100%), 80 (32%), 93 (67%) Flower, lavender27.28 (39) b-Selinene 77 (60%), 80 (64%), 93 (100%), 105 (87%) Herb

Esters6.25 (2) Ethyl acetate 43 (100%), 45 (22%), 61 (20%), 70 (20%) Pineapple8.50 (7) Ethyl propionate 45 (14%), 57 (100%), 74 (14%), 75 (15%) Fruit12.76 (17) Isoamyl acetate 42 (17%), 43 (100%), 55 (40%), 70 (58%) Banana19.91 (35) Methyl benzoate 51 (27%), 77 (62%), 105 (100%), 136 (36%) Brune, lettuce, herb, sweet21.75 (38) Ethyl benzoate 51 (19%), 77 (43%), 105 (100%), 122 (31%) Camomile, flower, celery, fruit

Acids11.40 (13) 3-Methylbutanoic acid 41 (26%), 43 (44%), 60 (100%), 87 (12%) Sweat, rancid, acid11.66 (14) 2-Methylbutanoic acid 41 (55%), 57 (49%), 74 (100%), 87 (30%) Cheese, sweatOther8.79 (9) Diethyl acetal 45 (100%), 47 (20%), 73 (55%), 103 (15%) Fruit, cream12.07 (16) 2,5-Dimethylfuran 53 (44%), 81 (37%), 95 (85%), 96 (100%) Chemical, ethereal, meaty, gravy, roast beef, bacon13.73 (18) Styrene 51 (16%), 78 (29%), 103 (42%), 104 (100%) Balsamic, gasoline

a Peak numbers refer to Fig. 4 (not all compounds are found in reference lingonberry sample).b Ions used for semi-quantification.c Odour descriptions adapted from Acree and Arn, 2010.

152 K. Viljanen et al. / Food Chemistry 157 (2014) 148–156

produced some acetic acid. However, the level of these two organicacids remained low, under 20 mg/g dry matter. Acetic acid is pro-duced during yeast fermentation from enzymatic oxidation of acet-aldehyde (Díaz-Montaño & de Jesús Ramírez Córdova, 2009). H.uvarum has a low level of acetyl CoA synthase activity which al-lows accumulation of acetate in the medium. The fermentationswith yeast and yeast + LAB produced mannitol. Since mannitol isformed during reduction of fructose, this result correlated wellwith the decreasing fructose concentration in the same fermenta-tion samples.

Selected phenolic compounds, benzoic acid, p-coumaric acid,caffeic acid, quercetin and anthocyanins were quantitated by HPLC(Table 1). The distribution of these selected phenolic compounds isin accordance with the literature (Mattila, Hellström, & Törrönen,2006). The concentration of benzoic acid increased during fermen-tation in almost all lingonberry samples compared to referencelingonberry sample without any fermentation. Benzoic acid levelswere highest in lingonberry samples fermented with yeast and

yeast + LAB. In addition, the concentration of p-hydroxybenzoicacid increased in yeast and yeast + LAB fermented samples, com-pared with reference lingonberry sample and other fermentations.Slightly increase was also seen in enzymatically treated samples.The concentration of both p-coumaric acid and caffeic acid was rel-atively low in all samples (under 0.4 mg/g dry matter). The highestconcentrations of these phenolic acids were found in samples fer-mented with yeast and yeast + LAB. It is known that many aromaticcompounds, such as terpenes, anthocyanins and phenolics arestored in wine grapes in a glycosidically-bound form. Added,endogenous and microbially produced glycosidases can releasethe more volatile compounds and have a major impact on sensoryproperties (Claus, 2009).

Highest quercetin concentrations were found in yeast + LAB fer-mented lingonberries as well as enzyme treated lingonberries withand without LAB fermentation. Quercetin and kaempferol, whichare the main flavonols in lingonberry, exist mainly as various sugarconjugates in the berry fruit, in which the amount of free quercetin

Table 3Normalised peak areas (compound area/ISTD area; average ± SD) of volatile flavour compounds of lingonberries.

Ref Enzymatic Yeast LAB Yeast + LAB Enzymatic + LAB

AldehydesE-2-Butenal nda nda 0.006 ± 0.001a nda 0.089 ± 0.017b nda

Pentanal 0.034 ± 0.005d 0.023 ± 0.003bc 0.017 ± 0.002b 0.020 ± 0.002b 0.010 ± 0.002a 0.029 ± 0.005cd

Hexanal 0.053 ± 0.005c 0.018 ± 0.002b 0.015 ± 0.002b 0.020 ± 0.003b 0.009 ± 0.001a 0.006 ± 0.001a

2-Furaldehyde 0.006 ± 0.001ab 0.008 ± 0.001b 0.007 ± 0.000ab 0.006 ± 0.001a 0.007 ± 0.001ab 0.008 ± 0.001ab

(E)-2-Heptenal 0.005 ± 0.001a 0.005 ± 0.001b 0.003 ± 0.001a 0.003 ± 0.001a 0.005 ± 0.001b 0.005 ± 0.000b

Benzaldehyde 0.060 ± 0.009ab 0.060 ± 0.012ab 0.102 ± 0.015c 0.046 ± 0.003a 0.071 ± 0.013b 0.062 ± 0.008ab

Octanal 0.003 ± 0.001b 0.003 ± 0.000b nda nda nda nda

Nonanal 0.005 ± 0.001b 0.005 ± 0.000b 0.002 ± 0.000a 0.003 ± 0.000a 0.003 ± 0.000a 0.003 ± 0.000a

KetonesDiacetyl 0.070 ± 0.009a 0.045 ± 0.005a 0.546 ± 0.081b 0.053 ± 0.008a 0.899 ± 0.107c 0.010 ± 0.001a

2-Pentanone 0.002 ± 0.000a 0.006 ± 0.001ab 0.019 ± 0.003c 0.059 ± 0.010e 0.011 ± 0.002bc 0.037 ± 0.005d

3-Hydroxy-2-butanone 0.005 ± 0.001a 0.005 ± 0.001a 1.217 ± 0.220b 0.008 ± 0.001a 2.870 ± 0.480c 0.006 ± 0.001a

4-Octen-3-one 0.004 ± 0.001ab 0.003 ± 0.000a 0.003 ± 0.001a 0.003 ± 0.000a 0.004 ± 0.000b 0.004 ± 0.000ab

6-Methyl-5-hepten-2-one 0.004 ± 0.001ab 0.005 ± 0.001b 0.007 ± 0.001c 0.004 ± 0.000ab 0.007 ± 0.001c 0.004 ± 0.001a

Acetophenone 0.030 ± 0.005a 0.052 ± 0.007b 0.049 ± 0.009b 0.028 ± 0.003a 0.052 ± 0.009b 0.030 ± 0.002a

AlcoholsEthanolA 0.5 ± 0.0a 0.6 ± 0.0a 98.4 ± 2.7b 0.5 ± 0.0a 107.8 ± 3.1c 0.5 ± 0.0a

3-Methyl-1-butanol 0.004 ± 0.001a 0.003 ± 0.000a 1.690 ± 0.263b 0.002 ± 0.000a 1.569 ± 0.151b 0.002 ± 0.000a

2-Methyl-1-butanol 0.002 ± 0.000a nda 0.829 ± 0.157c nda 0.545 ± 0.041b nda

2-Ethyl-1-hexanol nda nda 0.002 ± 0.000b nda nda nda

Benzyl alcohol 0.050 ± 0.008a 0.075 ± 0.013bc 0.093 ± 0.014cd 0.057 ± 0.005ab 0.103 ± 0.020d 0.114 ± 0.009d

(Z)-3-Hexen-1-ol 0.003 ± 0.000cd 0.003 ± 0.000d 0.003 ± 0.000cd 0.002 ± 0.000bc 0.002 ± 0.000b nda

Phenylethyl alcohol nda nda 0.212 ± 0.025b nda 0.612 ± 0.103c nda

Terpenesa-Pinene 0.039 ± 0.006c 0.069 ± 0.008d 0.023 ± 0.004b 0.046 ± 0.008c 0.010 ± 0.001a 0.008 ± 0.001a

b-Elemene 0.002 ± 0.000c 0.003 ± 0.001d 0.001 ± 0.000b 0.002 ± 0.000c 0.001 ± 0.000b 0.000 ± 0.000a

p-Cymene 0.044 ± 0.009c 0.058 ± 0.009d 0.043 ± 0.007c 0.041 ± 0.004c 0.030 ± 0.006b 0.010 ± 0.001a

Limonene 0.006 ± 0.001c 0.009 ± 0.001d 0.004 ± 0.001b 0.005 ± 0.000bc 0.004 ± 0.001b 0.002 ± 0.000a

1,8-Cineole 0.020 ± 0.002cd 0.023 ± 0.001d 0.015 ± 0.003b 0.017 ± 0.002bc 0.014 ± 0.002b 0.001 ± 0.000a

Linalool 0.012 ± 0.002b 0.019 ± 0.002c 0.013 ± 0.002b 0.011 ± 0.002b 0.012 ± 0.002b 0.004 ± 0.000a

b-Selinene nda nda 0.006 ± 0.001b nda 0.005 ± 0.001b nda

EstersEthyl acetate 0.113 ± 0.022a 0.082 ± 0.005a 22.559 ± 4.325b 0.108 ± 0.004a 33.757 ± 4.054c 0.133 ± 0.024a

Ethyl propionate 0.007 ± 0.001a 0.006 ± 0.001a 0.055 ± 0.010b 0.009 ± 0.002a 0.138 ± 0.026c 0.008 ± 0.001a

Isoamyl acetate nda nda 0.179 ± 0.025b nda 0.764 ± 0.138c nda

Methyl benzoate 0.111 ± 0.019bc 0.124 ± 0.017bc 0.137 ± 0.023c 0.091 ± 0.016b 0.104 ± 0.018bc 0.056 ± 0.007a

Ethyl benzoate 0.032 ± 0.003a 0.019 ± 0.004a 0.274 ± 0.047c 0.005 ± 0.001a 0.202 ± 0.038b 0.016 ± 0.003a

Acids3-Methylbutanoic acid 0.065 ± 0.012ab 0.045 ± 0.007a 0.058 ± 0.010ab 0.072 ± 0.005b 0.057 ± 0.010ab 0.100 ± 0.016c

2-Methylbutanoic acid 0.092 ± 0.016ab 0.063 ± 0.010a 0.093 ± 0.017ab 0.102 ± 0.014bc 0.131 ± 0.018c 0.113 ± 0.021bc

OtherDiethyl acetal 0.005 ± 0.001a 0.006 ± 0.001a 0.336 ± 0.060b 0.007 ± 0.001a 0.822 ± 0.119b 0.006 ± 0.001a

2,5-Dimethylfuran 0.006 ± 0.001a 0.008 ± 0.001c 0.008 ± 0.001bc 0.006 ± 0.001ab 0.007 ± 0.001abc 0.008 ± 0.001abc

Styrene 0.019 ± 0.003d 0.015 ± 0.001c 0.012 ± 0.001c 0.011 ± 0.002c 0.007 ± 0.001b 0.002 ± 0.000a

a–dValues in the same row followed by different letters are statistically significantly different (p < 0.05).nd, not detected.

A Ethanol concentration mg of ethanol/g of sample.

K. Viljanen et al. / Food Chemistry 157 (2014) 148–156 153

is minor. Thus, bioprocessing with enzymes or microbial cells maydegrade phenolic aglycone–sugar linkages resulting in free querce-tin accumulation, and in that way affecting sensory properties,especially bitter character (Sáenz-Navajas, Ferreira, Dizy, &Fernández-Zurbano, 2010).

The concentration of total anthocyanins remained unchanged inlingonberry samples fermented with yeast and LAB alone or to-gether. On the contrary, the anthocyanin concentration decreasedin enzymatically treated lingonberry samples with and withoutLAB fermentation. The stability and the profile of anthocyaninsare greatly affected by the glycosidase side activities present inthe enzyme preparations, which are able to hydrolyse certainanthocyanins to the corresponding aglycones. The aglycones aremore unstable than the glycosidic forms, and this may result inanthocyanin decrease (Buchert et al., 2005; Koponen, Buchert,Poutanen, & Törrönen, 2008). In more detailed analyses of lingon-berry anthocyanins, we found that cyanidin-3-galactoside, themain lingonberry anthocyanin, was very sensitive to pectinasetreatments, and most of the compound (about 60%) was degraded

(data not shown). However, the two other main lingonberry antho-cyanins, cyanidin-3-glucoside and cyanidin-3-arabinoside, werenot affected by this enzyme preparation. Purified enzymes or en-zyme mixtures with known activity profiles could offer an interest-ing tool to modify berry material and also tailor sensory propertieswithout affecting the phenolic compounds, and thus antioxidantand health effects.

LAB fermentation alone was very gentle to phenolic com-pounds; the concentration remained unchanged for all of thosestudied. In general, during LAB fermentations slight decrease inall anthocyanins can be detected. However these changes aremostly caused by the fermentation conditions, not by the LAB (datanot shown).

3.4. Volatile flavour-active compounds in lingonberries

From lingonberry samples 38 volatile chemical compoundswere identified by SPME–GC/MS (Table 2); 8 aldehydes, 6 ketones,7 alcohols, 7 terpenes, 5 esters, 2 acids and 3 other compounds. A

Fig. 2. Relation of sensory perception and non-volatile compounds of processed lingonberry. The samples are marked in green (reference, yeast ferm yeast fermentation,enzyme = enzymatic treatment, LAB ferm = LAB fermentation, yeast & LAB = yeast and LAB fermentation, enzyme & LAB = enzymatic treatment and LAB fermentation),chemical compounds in blue and sensory characteristics in red (prefix O = odour, F = flavour, Col = colour).

Fig. 3. Relation of sensory perception and volatile compounds of processed lingonberry by PLS regression. The samples are marked in green, chemical compounds in blue andsensory attributes in red. Abbreviations are as stated in the text of Fig. 2.

154 K. Viljanen et al. / Food Chemistry 157 (2014) 148–156

chromatogram of untreated lingonberries is presented in Fig. 4. Inthe literature, only one research paper was found on volatile com-pounds in lingonberries (Anjou & Von Sydow, 1967), in whichessential oil was extracted from berries, which differs from theSPME–GC/MS analysis used in this study.

The amount and composition of volatile chemical compoundschanged during fermentations and enzymatic treatments. Theamount of diacetyl, ethyl acetate, ethyl propionate, acetoin, diethylacetal, 3-methyl-1-butanol, 2-methyl-1-butanol, isoamyl acetate,6-methyl-5-hepten-2-one, benzyl alcohol, acetophenone, phenyl-ethyl alcohol, ethyl benzoate and b-selinene increased in samplesfermented with yeast and yeast + LAB. In particular, the levels ofall measured esters increased in lingonberries fermented withyeast or with yeast + LAB. The level of ethyl acetate in yeast andyeast + LAB fermented lingonberries is more than 1000 times high-er than in reference lingonberries.

Diacetyl is small ketone that has butter flavour. It is producednormally in both LAB and yeast fermentation (Romano & Suzzi,1996; Suomalainen & Ronkainen, 1968; Walsh & Cogan, 1973). Inaddition, acetoin, which also has butter and cream odour, is pro-duced in the same route as diacetyl so it is logical that they are bothincreased in yeast and yeast + LAB fermentation. Diacetyl wasfound in yeast and yeast + LAB fermented lingonberries at morethan 7-fold higher levels than in reference untreated lingonberries.Phenylethyl alcohol is produced by reduction of acetophenone(Rogers, Hackman, Mercer, & DeLancey, 1999), and these two com-pounds are commonly found in yeast fermented foods and bever-ages such as beer and wine. Acetophenone was found in yeastand yeast + LAB fermented lingonberries around two times higherthan in reference lingonberries and it has flower and almond odour.Phenylethyl alcohol has honey, spice, rose and lilac odour and wasfound only in yeast and yeast + LAB fermented lingonberries.

Fig. 4. GC/MS chromatogram of lingonberry reference sample. The numbers of peaks refer to compounds listed in Table 2; ISTD, internal standard, toluene; A, artefact ofSPME fibre.

K. Viljanen et al. / Food Chemistry 157 (2014) 148–156 155

The level of most aldehydes remained unchanged or decreasedduring yeast and LAB fermentations and enzyme treatments. Thedecreased levels of these aldehydes may be due to their degrada-tion and oxidation. The decreased levels of aldehydes is probablyalso due to the fact that the acidic pH (4.6–4.9), used in thesetreatments is not optimal for endogenous enzymes, such aslipoxygenase and hydroperoxide lyase. Lipoxygenase has beenshown to oxidise endogenous fatty acids, and hydroperoxidelyase decomposes formed fatty acid hydroperoxides to differentaldehydes and ketones. If the activities of these enzymes areinhibited, degradation is more favoured than formation. The levelof 6-methyl-5-hepten-2-one increased in yeast fermented lingon-berries. 6-Methyl-5-hepten-2-one is a degradation product oflycopene. Lingonberries contain some carotenoids, but not in highquantities.

LAB fermentation alone and together with enzymatic treatmentincreased most the level of 2-pentanone. The level of benzyl alco-hol increased in all samples compared to the reference sample.

It seems that enzymatic treatment alone or in combination withLAB fermentation has little effect on terpenes, but in some casesterpene levels were increasing. a-Pinene levels significantlyincreased with enzymatic treatment whereas with otherbioprocesses a-pinene levels decreased or remained unchanged.Lingonberries fermented with yeast or with yeast + LAB had thehighest levels of b-selinene compared to reference lingonberriesor lingonberries with other fermentation and enzymatic treat-ments. According to these results enzymatic treatment in combi-nation with LAB fermentation decreased most the levels ofterpenes.

3.5. Comparison of sensory and chemical analysis

The chemical composition of non-volatile compounds of bio-processed lingonberries was related to sensory perception(Fig. 2). Logically anthocyanins are related to redness, glucoseand fructose to sweetness, and ethanol, phenolic and some organicacids to sour taste, bitterness and fermented odour/flavour. On theother hand, lactic acid is responsible for sour odour.

Volatile compounds were also related to sensory perception asseen in Fig. 3. Statistical multivariate analysis (PLS regression)showed excellent correlation between the sensory and chemicalresults: The two first principal components explained 88% of thevariation in chemical and 90% in sensory results. Yeast fermenta-tion alone or together with LAB fermentation resulted in fermentedodour and flavour, sour, bitter taste and off-taste, which wererelated to phenolic acids (p-coumaric, caffeic and benzoic acids),citric and acetic acid, mannitol, ethanol and pH, and diacetyl and

several other volatile compounds. Reference and LAB-fermentedlingonberries were fresh and lingonberry-like in flavour, and wererelated to octanal and nonanal. Both nonanal and octanal have aspure compounds lemon, citrus, green flavour. Enzyme treatmentalone and together with LAB fermentation was related to 3-meth-ylbutanoic acid which has sweet, rancid and acidic flavour. Mostchanges in volatile composition are related to yeast and yeast &LAB fermentations; for example the amount of diacetyl, ethylbenzoate, ethyl acetate, ethyl propionate, 3-methyl-1-butanol,2-methyl-1-butanol, phenylethyl alcohol, diethyl acetal, 2-hydro-xy-2-butanone, and (E)-2-butenal.

Enzyme and enzyme-LAB treated lingonberries were linked tosweetness, and glucose and fructose, as well as 3-methylbutanoicacid. LAB fermentation of the lingonberry material was poor anddid not seem to have impact on the sensory profile of lingonberry.Redness duly correlated with anthocyanins. Thus, yeast fermenta-tion was shown to cause a non-desired fermented odour andflavour to lingonberry, whereas enzyme treatment was shown toincrease the sweetness of lingonberry. Excessive formation of eth-anol could be controlled by oxygenation as H. uvarum producesethanol only under oxygen limitation (Díaz-Montaño & de JesúsRamírez Córdova, 2009).

4. Conclusions

Fermentation and enzymatic treatment are possible bioprocess-ing pathways to modify lingonberry flavour, and thus they couldenhance the use of lingonberries in food and beverage product for-mulation. Enzyme treatment was the most potent processingmethod to modify lingonberry taste, as it increased sweetness ofthis naturally very acidic berry. Yeast fermentation resulted in fer-mented odour and off-taste. However, optimisation of the fermen-tation process could be used to improve the flavour of the product,e.g., by reducing the amount of ethanol. Fermented berries, on theother hand, contained higher levels of phenolic compounds, whichcould give extra health advantages to the products, and also asnatural preservatives, could increase preservation of food andbeverage products.

Acknowledgements

Tekes – the Finnish Funding Agency for Technology and Innova-tion is acknowledged for financial support of the BERRYFERMEN-TATION project (No. 40437/08) during the years 2009–20011.The skilful technical assistance of Airi Hyrkäs, Liisa Änäkäinen,Pirkko Nousiainen, Heidi Eriksson, Merja Salmijärvi and UllaÖsterlund is gratefully acknowledged.

156 K. Viljanen et al. / Food Chemistry 157 (2014) 148–156

References

Aaby, K., Ekeberg, D., & Skrede, G. (2007). Characterizarion of phenolic compoundsin strawberry (Fragaria x ananassa) fruits by different HPLC detectors andcontribution of individual compounds to total antioxidant capacity. Journal ofAgricultural and Food Chemistry, 55, 4395–4406.

Acree, T., & Arn, H. (2010). Flavornet and human odor space. Gas chromatographyolfactometry (GCO) of natural products. Cornell University.

Anjou, K., & Von Sydow, E. (1967). The aroma of cranberries. I. Vaccinium vitis-idaea.Acta Chemica Scandinavica, 21, 945–952.

Bourdichon, F., Casaregola, S., Farrokh, C., Frisvad, J. C., Gerds, M. L., Hammes, W. p.,et al. (2012). Food fermentations: Microorganisms with technological beneficialuse. International Journal of Food Microbiology, 154, 87–97.

Brul, S., & Coote, P. (1999). Preservation agents in foods. Mode of action andmicrobial resistance mechanisms. Review. International Journal of FoodMicrobiology, 50, 1–17.

Buchert, J., Koponen, J., Suutarinen, M., Mustranta, A., Lille, M., Törrönen, R., et al.(2005). Effect of enzyme-aided pressing on anthocyanin yield and profiles inbilberry and blackcurrant juices. Journal of the Science of Food and Agriculture, 85,2548–2556.

Claus, H. (2009). Exoenzymes of wine microorganisms. In H. König, G. Unden, & J.Föchlich (Eds.), Biology of microorganisms on grapes, in must and wine(pp. 259–271). Berlin Heidelberg: Springer-Verlag.

Díaz-Montaño, D. M., & de Jesús Ramírez Córdova, J. (2009). The fermentative andaromatic ability of Kloeckera and Hanseniaspora yeasts. In T. Satyanarayana &G. Kunze (Eds.), Yeast biotechnology: Diversity and applications (pp. 297–320).Springer Science + Business Media B.V.

Ek, S., Kartimo, H., Mattila, S., & Tolonen, A. (2006). Characterization of phenoliccompounds from lingonberry (Vaccinium vitis-idaea). Journal of Agricultural andFood Chemistry, 54, 9834–9842.

Kähkönen, M. P., Heinämäki, J., Ollilainen, V., & Heinonen, M. (2003). Berryanthocyanins: Isolation, identification and antioxidant activities. Journal of theScience of Food and Agriculture, 83, 1403–1411.

Katina, K., Laitila, A., Juvonen, R., Liukkonen, K., Kariluoto, S., Piironen, V., et al.(2007). Bran fermentation as a means to enhance technological properties andbioactivity of rye. Food Microbiology, 24, 175–186.

Koponen, J., Buchert, J., Poutanen, K., & Törrönen, R. (2008). Effect of pectinolyticjuice production on the extractability and fate of bilberry and black currantanthocyanins. European Food Research and Technology, 227, 485–494.

Kylli, P., Nohynek, L., Puupponen-Pimiä, R., Westerlund-Wikström, B., Leppänen, T.,Welling, J., et al. (2011). Lingonberry (Vaccinium vitis-idaea) and EuropeanCranbery (Vaccinium microcarpon) proanthocyanidins: Isolation, identificationand bioactivities. Journal of Agricultural and Food Chemistry, 59, 3373–3384.

Laaksonen, O., Sandell, M., Nordlund, E., Heiniö, R.-L., Malinen, H.-L., Jaakkola, M.,et al. (2012). The effect of enzymatic treatment on blackcurrant (Ribes nigrum)juice flavour and its stability. Food Chemistry, 130, 31–41.

Lawless, H. T., & Heymann, H. (2010). Sensory evaluation of food principles andpractises, descriptive analysis (2nd ed.). Gaithersburg: Chapman & Hall/AspenPublishers Inc., pp. 378–441.

Lee, J., & Finn, C. E. (2012). Lingonberry (Vaccinium vitis-idaea L.) grown in the Pacificnorthwest of North America: Anthocyanin and free amino acid composition.Journal of Functional Foods, 4, 213–218.

Määttä-Riihinen, K. R., Kamal-Eldin, A., & Törrönen, A. R. (2004). Identification andquantification of phenolic compounds in berries of Fragaria and Rubus species(Family Rosaceae). Journal of Agricultural and Food Chemistry, 52, 6178–6187.

Martens, H., & Naes, T. (1998). Multivariate calibration. New York: Wiley [p. 419].Mattila, P., Hellström, J., & Törrönen, R. (2006). Phenolic acids in berries, fruits, and

beverages. Journal of Agricultural and Food Chemistry, 54, 7193–7199.Mikulic-Petkovsek, M., Schmitzer, V., Slatnar, A., Stampar, F., & Vebeic, R. (2012).

Composition of sugars, organic acids, and total phenolics in 25 wild orcultivated berry species. Journal of Food Science, 77, C1064–C1070.

Moreira, N., Mendes, F., Guedes de Pinho, P., Hogg, T., & Vasconcelos, I. (2008).Heavy sulphur compounds, higher alcohols and esters production profile ofHanseniaspora uvarum and Hanseniaspora guilliermondii grown as pure andmixed cultures in grape must. International Journal of Food Microbiology, 124,231–238.

Rogers, P. S., Hackman, J. R., Mercer, V., & DeLancey, G. B. (1999). Acetophenonetolerance, chemical adaptation, and residual bioreactive capacity of non-fermenting baker’s yeast (Saccharomyces cerevisiae) during sequential reactorcycles. Journal of Industrial Microbiology and Biotechnology, 22, 108–114.

Romano, P., & Suzzi, G. (1996). Origin and production of acetoin during wine yeastfermentation. Applied and Environmental Microbiology, 62, 309–315.

Saarela, M., Rantala, M., Hallamaa, K., Nohynek, L., Virkajarvi, I., & Matto, J. (2004).Stationary-phase acid and heat treatments for improvement of the viability ofprobiotic lactobacilli and bifidobacteria. Journal of Applied Microbiology, 96,1205–1214.

Sáenz-Navajas, M. P., Ferreira, V., Dizy, M., & Fernández-Zurbano, P. (2010).Characterization of taste-active fractions in red wine combining HPLCfractionation, sensory analysis and ultra performance liquid chromatographycoupled with mass spectrometry detection. Analytica Chimica Acta, 673,151–159.

Suomalainen, H., & Ronkainen, P. (1968). Mechanism of diacetyl formation in yeastfermentation. Nature, 220, 792–793.

Tenkanen, M., & Siika-aho, M. (2000). An a-glucuronidase of Schizophyllumcommune acting on polymeric xylans. Journal of Biotechnology, 78, 149–161.

Turtiainen, M., Salo, K., & Saastamoinen, O. (2011). Variations of yield and utilisationof bilberries (Vaccinium myrtillus L.) and cowberries (V. vitis-idaea L.) in Finland.Silva Fennica, 45, 237–251.

Viljakainen, S., & Laakso, S. (2002). Acidity reduction in northern region berry juicesby the malolactic bacterium Oenococcus oeni. European Food Research andTechnology, 214, 412–417.

Viljakainen, S., Visti, A., & Laakso, S. (2002). Concentrations of organic acids andsoluble sugars in juices from Nordic berries. Acta Agriculturae Scandinavica, B,52, 101–109.

Viljanen, K., Kylli, P., Hubbermann, E.-M., Schwarz, K., & Heinonen, M. (2005).Anthocyanin antioxidant activity and partition behaviour in whey proteinemulsion. Journal of Agricultural and Food Chemistry, 53, 2022–2027.

Viljanen, K., Kylli, P., Kivikari, R., & Heinonen, M. (2004). Inhibition of protein andlipid oxidation in liposomes by berry phenolics. Journal of Agricultural and FoodChemistry, 52, 7419–7424.

Viljanen, K., Lille, M., Heiniö, R.-L., & Buchert, J. (2011). Effect of high-pressureprocessing on volatile composition and odour of cherry tomato purée. FoodChemistry, 129, 1759–1765.

Visti, A., Viljakainen, S., & Laakso, S. (2003). Preparation of fermentable lingonberryjuice through removal of benzoic acid by Saccharomyces cerevisiae yeast. FoodResearch International, 36, 597–602.

Walsh, B., & Cogan, T. M. (1973). Diacetyl, acetoin, and acetaldehyde production bymixed-species lactic starter cultures. Applied Microbiology, 26, 820–825.

http://refhub.elsevier.com/S0308-8146(14)00204-0/h0005http://refhub.elsevier.com/S0308-8146(14)00204-0/h0005http://refhub.elsevier.com/S0308-8146(14)00204-0/h0005http://refhub.elsevier.com/S0308-8146(14)00204-0/h0005http://refhub.elsevier.com/S0308-8146(14)00204-0/h0010http://refhub.elsevier.com/S0308-8146(14)00204-0/h0010http://refhub.elsevier.com/S0308-8146(14)00204-0/h9030http://refhub.elsevier.com/S0308-8146(14)00204-0/h9030http://refhub.elsevier.com/S0308-8146(14)00204-0/h9030http://refhub.elsevier.com/S0308-8146(14)00204-0/h0015http://refhub.elsevier.com/S0308-8146(14)00204-0/h0015http://refhub.elsevier.com/S0308-8146(14)00204-0/h0015http://refhub.elsevier.com/S0308-8146(14)00204-0/h0020http://refhub.elsevier.com/S0308-8146(14)00204-0/h0020http://refhub.elsevier.com/S0308-8146(14)00204-0/h0020http://refhub.elsevier.com/S0308-8146(14)00204-0/h0020http://refhub.elsevier.com/S0308-8146(14)00204-0/h0025http://refhub.elsevier.com/S0308-8146(14)00204-0/h0025http://refhub.elsevier.com/S0308-8146(14)00204-0/h0025http://refhub.elsevier.com/S0308-8146(14)00204-0/h0030http://refhub.elsevier.com/S0308-8146(14)00204-0/h0030http://refhub.elsevier.com/S0308-8146(14)00204-0/h0030http://refhub.elsevier.com/S0308-8146(14)00204-0/h0030http://refhub.elsevier.com/S0308-8146(14)00204-0/h0030http://refhub.elsevier.com/S0308-8146(14)00204-0/h0030http://refhub.elsevier.com/S0308-8146(14)00204-0/h0035http://refhub.elsevier.com/S0308-8146(14)00204-0/h0035http://refhub.elsevier.com/S0308-8146(14)00204-0/h0035http://refhub.elsevier.com/S0308-8146(14)00204-0/h0040http://refhub.elsevier.com/S0308-8146(14)00204-0/h0040http://refhub.elsevier.com/S0308-8146(14)00204-0/h0040http://refhub.elsevier.com/S0308-8146(14)00204-0/h0045http://refhub.elsevier.com/S0308-8146(14)00204-0/h0045http://refhub.elsevier.com/S0308-8146(14)00204-0/h0045http://refhub.elsevier.com/S0308-8146(14)00204-0/h0050http://refhub.elsevier.com/S0308-8146(14)00204-0/h0050http://refhub.elsevier.com/S0308-8146(14)00204-0/h0050http://refhub.elsevier.com/S0308-8146(14)00204-0/h0055http://refhub.elsevier.com/S0308-8146(14)00204-0/h0055http://refhub.elsevier.com/S0308-8146(14)00204-0/h0055http://refhub.elsevier.com/S0308-8146(14)00204-0/h0055http://refhub.elsevier.com/S0308-8146(14)00204-0/h0060http://refhub.elsevier.com/S0308-8146(14)00204-0/h0060http://refhub.elsevier.com/S0308-8146(14)00204-0/h0060http://refhub.elsevier.com/S0308-8146(14)00204-0/h0065http://refhub.elsevier.com/S0308-8146(14)00204-0/h0065http://refhub.elsevier.com/S0308-8146(14)00204-0/h0065http://refhub.elsevier.com/S0308-8146(14)00204-0/h0070http://refhub.elsevier.com/S0308-8146(14)00204-0/h0070http://refhub.elsevier.com/S0308-8146(14)00204-0/h0070http://refhub.elsevier.com/S0308-8146(14)00204-0/h0075http://refhub.elsevier.com/S0308-8146(14)00204-0/h0075http://refhub.elsevier.com/S0308-8146(14)00204-0/h0075http://refhub.elsevier.com/S0308-8146(14)00204-0/h0080http://refhub.elsevier.com/S0308-8146(14)00204-0/h0085http://refhub.elsevier.com/S0308-8146(14)00204-0/h0085http://refhub.elsevier.com/S0308-8146(14)00204-0/h0090http://refhub.elsevier.com/S0308-8146(14)00204-0/h0090http://refhub.elsevier.com/S0308-8146(14)00204-0/h0090http://refhub.elsevier.com/S0308-8146(14)00204-0/h9010http://refhub.elsevier.com/S0308-8146(14)00204-0/h9010http://refhub.elsevier.com/S0308-8146(14)00204-0/h9010http://refhub.elsevier.com/S0308-8146(14)00204-0/h9010http://refhub.elsevier.com/S0308-8146(14)00204-0/h9010http://refhub.elsevier.com/S0308-8146(14)00204-0/h0095http://refhub.elsevier.com/S0308-8146(14)00204-0/h0095http://refhub.elsevier.com/S0308-8146(14)00204-0/h0095http://refhub.elsevier.com/S0308-8146(14)00204-0/h0095http://refhub.elsevier.com/S0308-8146(14)00204-0/h0100http://refhub.elsevier.com/S0308-8146(14)00204-0/h0100http://refhub.elsevier.com/S0308-8146(14)00204-0/h0105http://refhub.elsevier.com/S0308-8146(14)00204-0/h0105http://refhub.elsevier.com/S0308-8146(14)00204-0/h0105http://refhub.elsevier.com/S0308-8146(14)00204-0/h0105http://refhub.elsevier.com/S0308-8146(14)00204-0/h0110http://refhub.elsevier.com/S0308-8146(14)00204-0/h0110http://refhub.elsevier.com/S0308-8146(14)00204-0/h0110http://refhub.elsevier.com/S0308-8146(14)00204-0/h0110http://refhub.elsevier.com/S0308-8146(14)00204-0/h0110http://refhub.elsevier.com/S0308-8146(14)00204-0/h0115http://refhub.elsevier.com/S0308-8146(14)00204-0/h0115http://refhub.elsevier.com/S0308-8146(14)00204-0/h0120http://refhub.elsevier.com/S0308-8146(14)00204-0/h0120http://refhub.elsevier.com/S0308-8146(14)00204-0/h0125http://refhub.elsevier.com/S0308-8146(14)00204-0/h0125http://refhub.elsevier.com/S0308-8146(14)00204-0/h0125http://refhub.elsevier.com/S0308-8146(14)00204-0/h0130http://refhub.elsevier.com/S0308-8146(14)00204-0/h0130http://refhub.elsevier.com/S0308-8146(14)00204-0/h0130http://refhub.elsevier.com/S0308-8146(14)00204-0/h0135http://refhub.elsevier.com/S0308-8146(14)00204-0/h0135http://refhub.elsevier.com/S0308-8146(14)00204-0/h0135http://refhub.elsevier.com/S0308-8146(14)00204-0/h0140http://refhub.elsevier.com/S0308-8146(14)00204-0/h0140http://refhub.elsevier.com/S0308-8146(14)00204-0/h0140http://refhub.elsevier.com/S0308-8146(14)00204-0/h0145http://refhub.elsevier.com/S0308-8146(14)00204-0/h0145http://refhub.elsevier.com/S0308-8146(14)00204-0/h0145http://refhub.elsevier.com/S0308-8146(14)00204-0/h0150http://refhub.elsevier.com/S0308-8146(14)00204-0/h0150http://refhub.elsevier.com/S0308-8146(14)00204-0/h0150http://refhub.elsevier.com/S0308-8146(14)00204-0/h0155http://refhub.elsevier.com/S0308-8146(14)00204-0/h0155http://refhub.elsevier.com/S0308-8146(14)00204-0/h0155http://refhub.elsevier.com/S0308-8146(14)00204-0/h0160http://refhub.elsevier.com/S0308-8146(14)00204-0/h0160

Relation of sensory perception with chemical composition of bioprocessed lingonberry1 Introduction2 Materials and methods2.1 Berry material2.2 Bioprocessing of lingonberries2.3 Sensory analysis2.4 Analysis of non-volatile flavour-active compounds2.4.1 Phenolic compounds2.4.2 Glucose, fructose and sucrose2.4.3 Organic acids and mannitol

2.5 Analysis of volatile odour compounds2.6 Statistical data analysis

3 Results and discussion3.1 Viable cell counts and pH3.2 Sensory evaluation of lingonberry3.3 Non-volatile flavour-active compounds in lingonberries3.4 Volatile flavour-active compounds in lingonberries3.5 Comparison of sensory and chemical analysis

4 ConclusionsAcknowledgementsReferences