Production of folate in oat bran fermentation by yeasts isolated from barley and diverse foods

Authors


Abstract

Aims

The focus of the research was to identify yeasts from barley kernels in order to study their folate production capability while maintaining high viscosity caused by soluble fibres in oat bran fermentation.

Methods and Results

The 65 isolated yeasts were characterized by API carbohydrate utilization tests, and assays for extracellular enzyme activities were the following: amylase, beta-glucanase, cellulase or CMCase, lipase, protease and xylanase. Yeasts were identified by partial DNA sequencing of the 25S D1/D2 and ITS1-5.8S-ITS2 regions. They belonged to the genera Aureobasidium, Cryptococcus, Pseudozyma and Rhodotorula. Folate production was determined from supernatant and cells grown in a rich laboratory medium or directly from oat bran solution inoculated with the appropriate yeast. Food yeasts, Saccharomyces cerevisiae, Candida milleri, Kluyveromyces marxianus and Galactomyces geotrichum, were used for comparison. Most of the yeasts isolated from barley destroyed the solid, viscous structure of the oat bran solution, indicating that they degraded the viscosity-generating soluble fibres, considered to be nutritionally advantageous. The best folate producers were S. cerevisiae, followed by Pseudozyma sp., Rhodotorula glutinis and K. marxianus. The yeasts maintaining high viscosity were used together with lactic acid bacteria (LAB) Streptococcus thermophilus or Lactobacillus rhamnosus to ferment oat bran solution. None of the yeasts isolated from barley, contrary to S. cerevisiae and C. milleri, produced together with LAB significant amounts of folate.

Conclusions

Fermentative yeasts together with LAB are potential for use in developing novel high folate content healthy foods and snacks from oat bran.

Significance and Impact of the Study

High soluble fibre content and high natural folate content but low energy content food and snack products with pleasant fermentation aroma provide possibilities for new developments in the food industry.

Introduction

Folate (Vitamin B9) is a generic name for a number of derivatives of pteroylglutamic acid (folic acid) and is necessary for methylation reactions in cell metabolism and for neural development of foetus during pregnancy. Natural dietary folates are mostly reduced folates, i.e. derivatives of tetrahydrofolate.

Folate is obtained especially from cereal foods (Kariluoto et al. 2004), fruits, dairy products and vegetables (Jägerstad et al. 2005). Yeasts are a rich source of folate (Hjortmo et al. 2005; Patring et al. 2006). However, the diversity of yeasts studied was rather limited: mostly Saccharomyces spp., some representatives of Candida, Debaryomyces, Kodamea, Metchnikowia, Wickerhamiella and a few unidentified yeasts. Physiological growth conditions and culture medium composition affect the folate contents of Saccharomyces cerevisiae—minimal medium and respiro-fermentative growth at high growth rate gave the highest folate contents (Hjortmo et al. 2007). Clearly, there is a need to expand the realm of studies on yeasts for folate production—almost 150 genera and 1500 species of yeasts are currently known (Kurtzman et al. 2011).

Recently, the European Food and Safety Authority has stated that ‘Regular consumption of beta-glucans contributes to maintenance of normal blood cholesterol concentrations’ with daily consumption of at least 3 g of β-glucan from nonprocessed or minimally processed products (EFSA 2009), and specifically, ‘Oat beta-glucan has been shown to lower/reduce blood cholesterol. Blood cholesterol lowering may reduce the risk of (coronary) heart disease' when consumed at least 3 g of β-glucan per day in a balanced diet (EFSA 2010).

Yeasts have previously been isolated from barley grains, barley malt or the malting process representing 13 different genera (Noots et al. 1998 and references therein) or 16 different genera (Laitila et al. 2006). The most commonly found genera in barley grains included Aureobasidium, Bulleromyces, Candida, Cryptococcus, Filobasidium, Galactomyces (Geotrichum), Rhodotorula and Sporobolomyces. Many of these and other yeasts can make extracellular enzymes, such as β-glucanase and amylases (De Mot 1990; Strauss et al. 2001).

We have previously shown that certain bacteria isolated from oat bran or rye flakes (Herranen et al. 2010; Kariluoto et al. 2010) or found in fermenting rye sour dough (Kariluoto et al. 2006) are able to synthesize significant amounts of folate in rich medium. In the current study, we isolated yeasts from barley kernels and investigated their ability to synthesize folate in pure culture alone in rich laboratory medium and in oat bran solution or in combination with lactic acid bacteria (LAB) in oat bran solution. One of the LABs was Streptococcus thermophilus known to produce folate and the other Lactobacillus rhamnosus unable to synthesize folate (Crittenden et al. 2003; Sybesma et al. 2003; for a review on LABs and folate, see Rossi et al. 2011). The objective of this study was to test whether growth or fermentation by pure cultures of yeasts alone or together with LABs in oat bran solution is a feasible approach for increasing food folate level while maintaining the high viscosity of the product.

Materials and methods

Isolation of yeasts from barley grains and from diverse foods

The barley Hordeum vulgare, variety Minttu, kernels were ground by scarification for 20 s periods with an abrasive mill to five fractions. Whole kernels were soaked for 3 h in 0·9% NaCl, and each ground fraction was homogenized in stomacher blender for 30 min. Yeasts were isolated by plating direct and diluted samples on rich medium (YPD or malt extract agar), with antibiotics chlortetracycline and chloramphenicol (at 0·01% each) preventing bacterial growth and with added Triton X-100 (at 0·02%) preventing fungal growth. The plates were incubated at 18 or 28°C for 4–20 days. Yeast colonies were picked, purified by restreaking, observed microscopically and tested for carbohydrate utilization using API 32C test strips (BioMerieux Inc., Marcy l'Etoile, France).

The food yeasts were isolated by direct plating on YPD agar plates. ABM4949 and ABM5103 were isolated from commercial rye sour dough starter, ABM5031 from spontaneously fermented apple cider, ABM5032 from spontaneously fermented lingonberry jam, ABM5099 from fermenting soya feed, ABM5102 from food laboratory air, ABM5130 and ABM5131 from domestic kefir grains, ABM5136 from fermented milk product viili and ABM5147 from fermented oat product. Saccharomyces cerevisiae ALKO743 was used as reference yeast—originally a commercial baker's yeast (Codón et al. 1998).

DNA extraction, PCR amplification and sequencing of rDNA coding for 25S D1/D2 and ITS1-5.8S-ITS2 regions

Representative and the control food yeasts were identified by partial rDNA sequence analysis by utilizing the PCR primers NL1-NL4 and ITS4-ITS5 described in Kurtzman and Robnett (2003). Genomic DNA was isolated from 2 days grown YPD shake-flask cultures using Wizard Genomic DNA Purification Kit (Promega Ltd., Essex, UK) according to manufacturer's instructions. The cells were broken by lyticase (Sigma L2524, Sigma Chemical Co., St. Louis, MO) treatment or cultures which were resistant to enzymatic hydrolysis by vigorously shaking for 3 min with glass beads. Universal PCR primers NL-1 (5′-ATATCAATAAGCGGAGGAAAAG-3′) and NL-4 (5′-GGTCCGTGTTTCAAGACGG-3′) were used to amplify a 0·6 kb fragment of 25S D1/D2 region; ITS-4 (5′-TCCTCCGCTTATTGATATGC-3′) and ITS-5 (5′-GGAAGTAAAAGTCGTAACAAGG-3′) were used to amplify a 0·4–0·8 kb fragment of ITS1-5.8S-ITS2 rDNA region (Kurtzman and Robnett 2003). Each 50-μl PCR reaction contained 0·5 μmol l−1 of each primer, 200 μmol l−1 dNTPs, 1 U DyNAzyme II DNA polymerase (Finnzymes Oy, Espoo, Finland), 1 × PCR buffer (10 mmol l−1 Tris-HCl, pH 8·8, 1·5 mmol l−1 MgCl2, 50 mmol l−1 KCl, 0·1% Triton X-100) and 1 μl of genomic DNA as a template. The amplifications were performed in GeneAmp PCR System 2700 thermocycler (Applied Biosystems, Foster City, CA) with the following parameters: an initial denaturation step at 94°C for 5 min, followed by 36 cycles at 94°C for 1 min, 52°C for 1 min and 72°C for 2 min and then a final extension step at 72°C for 10 min. The PCR products were then partially sequenced using primers NL-1, NL-4, ITS-4 and ITS-5 in combination with BigDye Terminator Cycle Sequencing Kit and ABI3130XL Genetic Analyser (Applied Biosystems). The nucleotide sequences were checked and edited with chromas lite software (ver. 2.01; Technelysium Pty Ltd., South Brisbane, Australia) and compared against the sequences in the National Centre for Biotechnology Information (NCBI) nr-database using the blastn programme. The closest strain level match (% identity) was considered as the identification. The sequence data has been deposited in the embl nucleotide sequence database (www.ebi.ac.uk) under accession numbers HG532066HG532115.

Screening for extracellular hydrolytic enzymes

The ability of yeasts to excrete hydrolytic enzymes was studied by substrate hydrolysis plate assay method as described in Herranen et al. (2010). Yeast strains were first grown on YPD plates (1% yeast extract, 2% tryptone, 2% glucose, 2% agar) for 2 days at 28°C. The yeasts were then streaked on enzyme assay plates and incubated at 28°C for 3 days. The plates for amylolytic, cellulolytic, xylanolytic, beta-glucanase or protease activities were glucose-free PCA plates (PCA-G; 0·5% casein peptone, 0·25% yeast extract, 1·5% agar) supplemented with 0·5% (w/v) soluble starch (Merck), 0·5% carboxymethylcellulose (Sigma), 1% xylan (from oat spelt, Sigma), 0·2% oat beta-glucan or 30% (v/v) skim milk as a substrate, respectively. The plates for amylolytic activity were stained with Lugol solution, while those for cellulose, xylanase or beta-glucanase were stained with 0·2% Congo Red. Lipolytic activity was determined with modified Sierra lipolysis agar containing 10 g tryptone, 5 g NaCl, 0·1 g CaCl2, 3 g meat extract, 0·2 g ferric citrate, 15 g agar and 10 ml of Tween 80 l−1. The distance from the margin of the colony to the rim of the hydrolysis zone was measured, and the enzyme activity was expressed as a function of the distance as follows: + <0·5 cm, ++ 0·5 to 1 cm, +++ 1 to 1·5 cm and ++++ >1·5 cm.

Analysis of total folate

Yeast strains were precultured in 10 ml of YPD broth overnight at 28°C with agitation at 180 rev min−1. The overnight cultures were used to inoculate fresh YPD medium to 5–10 Klett60 units, and the cultures were grown with agitation at 28°C. 30-ml samples were withdrawn at the stationary (24–41 h) growth phase. The cell wet weight (approx. 20% dry weight) yield for different yeasts varied in the range of 0·4–2·5 g. Cells were harvested by centrifugation at 4000 × g for 15 min at room temperature and washed once with sterile phosphate-buffered saline (pH 7·1). The supernatants were filtered through 0·45-μm membrane filters (Sarstedt, Nümbrecht, Germany). Both the cells and supernatants were flushed with nitrogen gas and stored at 20°C for further analysis.

Total folate contents were determined by a microbiological assay on microtiter plates using L. rhamnosus ATCC 7469 as the growth indicator organism (Kariluoto et al. 2004). The sample preparation procedure included heat extraction followed by deconjugation of folate polyglutamates by hog kidney conjugase and treatments with amylase and protease to liberate folate from the matrix. Method performance was confirmed by analysing a blank sample as well as certified reference material CRM 121 (wholemeal flour) or in-house reference in each set of samples.

Oat bran fermentations

Commercial oat bran product OatWell 14% (Swedish Oat Fiber AB, Bua, Sweden; Herranen et al. 2010) at 3·5% concentration with or without 2% glucose in water was boiled for 10 min, sterilized by autoclaving at 120°C, divided into 50-ml aliquots in plastic minigrip bags or Sarstedt tubes and after cooling inoculated with 1 ml of LAB or yeast overnight culture in YPD. In cases where LAB (Streptococcus thermophilus ABM5097 or Lactobacillus rhamnosus LC-705) and yeast were inoculated together, the volume of each culture was 0·5 ml. To mix the contents, the plastic bags were kneaded by hand, and the tubes were capped, mixed by inverting 10 times, then the caps were loosened and incubated at 28°C for 1 or more days. Samples for viable count determinations on YPD spread plates and for viscosity measurements were taken after finishing the experiment and mixing the bag or tube contents as above.

Viscosity determinations of oat bran solution samples

Viscosity properties of the oat bran samples were characterized using a ThermoHaake RheoStress 600 rheometer (Thermo Electron GmbH, Dreieich, Germany). A flow curve was obtained using a cone and plate geometry (35 mm, 2°) over a shear rate range of 0·3–300–0·3 s−1. All the rheological experiments were performed at 20°C.

Results

Identification of yeasts

We have isolated from different fractions of barley kernels altogether 65 pure cultures of yeasts. The viable count of yeasts from whole kernels was 4 × 104 per gram when incubated at 18°C and 6 × 103 per gram at 28°C. In the different kernel fractions, yeast viable count when incubated at 18°C varied from 3 × 103 to 4 × 105 g−1 flour and when incubated at 28°C from zero to 2 × 103 g−1. Most of the isolates were first characterized by observations on colony and cell morphology. API 32C carbohydrate utilization tests were made for selected strains isolated from barley grains and for the food yeasts, and a tentative identification based on API reference database comparison was made (data not shown). The barley isolates were assigned to various species in the genera Candida, Cryptococcus and Rhodotorula and food yeasts to Candida, Geotrichum and Saccharomyces.

A portion of the isolates was identified by partial DNA sequencing of the 25S D1/D2 and ITS1-5.8S-ITS2 regions. The identified yeasts (Table 1) isolated from barley grains belonged to the genera Aureobasidium, Cryptococcus, Pseudozyma and Rhodotorula, all of which are assimilative, nonfermentative yeasts. The food yeasts belonged to the genera Candida, Clavispora, Galactomyces, Kluyveromyces, Pichia, Rhodotorula and Saccharomyces (Table 1). Strain R59 was not sequenced for the D1/D2 region. ALKO 743 produced overlapping sequences for the ITS1-5.8S-ITS2 region, indicating that the region is heterozygous (data not shown). The tentative overall identification (Table 1) is based on the fact that ITS1-5.8S-ITS2 region is more discriminatory than D1/D2 region at species level (Kurtzman and Robnett 2003).

Table 1. Identification of yeast strains by DNA sequencing. More than one species is listed if the scores were (nearly) identical
Strain25S D1/D2 identitySpecies in NCBI databaseITS1-5.8S-ITS2 identitySpecies in NCBI databaseTentative overall identification
  1. ND, not determined due to mixed peaks in electropherograms.

  2. a

    Suc+ phenotype was taken into account.

Barley yeasts
R38556/556Aureobasidium pullulans/Kabatiella microsticta562/562 Aureobasidium pullulans Aureobasidium pullulans
R124555/555 Kabatiella microsticta 607/607 Aureobasidium pullulans Aureobasidium pullulans
R43564/564 Cryptococcus adeliensis 542/542 Cryptococcus adeliensis Cryptococcus adeliensis
R76582/582 Cryptococcus adeliensis 557/557 Cryptococcus adeliensis Cryptococcus adeliensis
R78582/582 Cryptococcus adeliensis 565/565 Cryptococcus adeliensis Cryptococcus adeliensis
R133607/609Cryptococcus sp.586/586Cryptococcus sp.Cryptococcus sp.
R46564/564Cryptococcus sp.566/570Cryptococcus sp.Cryptococcus sp.
R134619/619 Cryptococcus laurentii 530/530 Cryptococcus laurentii Cryptococcus laurentii
R59ND 567/567 Cryptococcus magnus Cryptococcus magnus
R47638/638Pseudozyma sp./Moesziomyces bullatus683/686Pseudozyma sp./Moesziomyces bullatusPseudozyma sp.
R45559/559 Rhodotorula glutinis/graminis/Rhodosporidium babjevae 574/574 Rhodotorula glutinis Rhodotorula glutinis
R48559/559 Rhodotorula glutinis/graminis/Rhodosporidium babjevae 571/571 Rhodotorula glutinis Rhodotorula glutinis
R63559/559 Rhodotorula glutinis/graminis/Rhodosporidium babjevae 597/597 Rhodotorula glutinis Rhodotorula glutinis
R132564/564 Rhodotorula graminis 585/585 Rhodotorula glutinis Rhodotorula glutinis
R106583/583 Rhodotorula laryngis 571/571 Rhodotorula laryngis Rhodotorula laryngis
Food yeasts
ABM4949603/604 Candida humilis 593/599 Candida humilis Candida milleri a
ABM5099603/604 Candida humilis 621/633 Candida humilis Candida milleri a
ABM5147556/561 Clavispora lusitaniae 373/373 Clavispora lusitaniae Clavispora lusitaniae
ABM5136538/540 Galactomyces geotrichum 347/347Galactomyces sp. Galactomyces geotrichum
ABM5032539/539 Kluyveromyces marxianus 678/679 Kluyveromyces marxianus Kluyveromyces marxianus
ABM5130511/511 Kluyveromyces marxianus 584/585 Kluyveromyces marxianus Kluyveromyces marxianus
ABM5031560/562 Saccharomyces cerevisiae 441/442 Pichia membranifaciens Pichia membranifaciens
ABM5102582/582 Rhodotorula pinicola 517/517 Rhodotorula pinicola Rhodotorula pinicola
ABM5103566/566 Saccharomyces cerevisiae 761/766 Saccharomyces cerevisiae Saccharomyces cerevisiae
ABM5131591/595 Saccharomyces cerevisiae 826/831 Saccharomyces cerevisiae Saccharomyces cerevisiae
ALKO743566/566 Saccharomyces cerevisiae ND  Saccharomyces cerevisiae

Plate assays for hydrolytic enzyme activities amylase, beta-glucanase, cellulase or CMCase, lipase, protease, and xylanase showed that the strains exhibited relatively few activities except for Aureobasidium pullulans which had all the assayed six activities (Table 2). All the Cryptococcus as well as Rhodotorula species isolated from barley had lipase activity and Cryptococcus laurentii and Cryptococcus. magnus in addition cellulase and the latter also had protease activity. Pseudozyma sp. exhibited both amylase and protease activity. Rhodotorula minuta and R. pinicola both had good β-glucanase activity, while some of the yeasts (Table 2) made a deep red precipitate without a halo which may mean some unknown modification of the β-glucan substrate.

Table 2. Production of extracellular hydrolytic enzymes by yeast strains isolated from barley kernels and diverse foodstuffs
IsolateIdentificationAmylaseCellulaseXylanaseBeta-glucanaseProteaseLipase
  1. −, No activity; +, low activity; ++, moderate activity; +++, high activity; ++++, very high activity.

  2. Enzyme activity is expressed relatively as a function of the size of the clearing zone as described in the 'Materials and methods'.

  3. a

    Although β-glucanase negative (no clearing zone), the colonies were surrounded by a deep red halo.

Barley yeasts
R38 Aureobasidium pullulans ++++++++++++
R124 Aureobasidium pullulans +++++++++++++++++
R43 Cryptococcus albidus ++
R76 Cryptococcus albidus +++
R78 Cryptococcus albidus +++
R46 Cryptococcus laurentii +++
R59 Cryptococcus magnus ++++++
R47Pseudozyma sp.+(+)a++
R45 Rhodotorula glutinis ++
R48 Rhodotorula glutinis ++
R63 Rhodotorula glutinis +++
R106 Rhodotorula laryngis ++++
Food yeasts
ABM4949 Candida milleri (++)a
ABM5099 Candida milleri (++++)a− − 
ABM5147 Clavispora lusitaniae
ABM5136 Galactomyces geotrichum (++)a
ABM5032 Kluyveromyces marxianus
ABM5130 Kluyveromyces marxianus
ABM5031 Pichia membranifaciens (+++)a
ABM5102 Rhodotorula pinicola ++
ABM5103 Saccharomyces cerevisiae
ABM5131 Saccharomyces cerevisiae
ALKO743 Saccharomyces cerevisiae

Folate production and bran viscosity

Folate production by the yeasts and LAB were studied under conditions which might be applicable to large scale industrial production. For potential commercial product development purposes, the preservation of the oat bran matrix viscosity was viewed as one of the key characteristics.

Folate production, assayed by a microbiological method, was determined from culture supernatant and cells grown in a laboratory medium inoculated with the appropriate yeast or LAB (Fig. 1). Folate contents in the culture supernatant after subtracting the un-inoculated control value (YPD 120 ng ml−1) varied from 65 to 230 ng ml−1 or 2–7 μg per 30 ml cultivation (Fig. 1a). The highest amounts of folate were made by Pseudozyma sp. R47, Aureobasidium pullulans R38 and Rhodotorula glutinis R48. Cell biomass from the 30-ml cultivation amounted on average to 2 g wet weight. Cell-bound folate assays showed (Fig. 1b) that baker's yeast ALKO743 made over 14 μg g−1, Pseudozyma sp. R47 over 12 μg g−1, Rhodotorula glutinis R48 about 11 μg g−1 and Kluyveromyces marxianus ABM5130 about 9 μg g−1. Aureobasidium pullulans R38 was among the poor folate producers with 4 μg g−1, while Rhodotorula laryngis R106 was the worst at 1 μg g−1. When calculated for the total amount of folate made in the 30-ml cultivations, the best overall folate producers were baker's yeast Saccharomyces cerevisiae ALKO743 at 34 μg, Pseudozyma sp. R47 at 32 μg followed by Rhodotorula glutinis R48 at 26 μg and Kluyveromyces marxianus ABM5130 at 25 μg. Folate in the biomass contributed about 78–94% of total folate. The worst folate producer was Rhodotorula laryngis R106 at 4 μg total folate. The bacterium Lactobacillus rhamnosus LC-705 did not make folate, but Streptococcus thermophilus made a significant amount of cell-bound folate (Fig. 1b).

Figure 1.

Production of folates in rich YPD medium by the yeasts in 24 h shake-flask cultures and by two lactic acid bacteria in 3 days static anaerobic cultures. (a) Folates produced by the micro-organisms into the culture supernatants (un-inoculated YPD-medium folates 120 ng ml−1 have been deducted from each value). (b) Folates found in wet weight cell mass. Error bars represent the range.

Next, the yeasts were inoculated into 50 ml of 3·5% oat bran solution and incubated for 3 days after which folate content (data not shown) and viscosity of each sample was determined. Most of the yeasts isolated from barley destroyed the solid, viscous structure of the oat bran solution (Fig. 2), indicating that they degraded the viscosity-generating soluble fibres. Control oat bran solution had a thick yoghurt-like viscosity, but at values especially below 50% of control, the yeast-containing samples were very watery. Some samples showed increased viscosity compared with the control—it remains to be determined whether that was due to the characteristics of the particular yeasts S. cerevisiae ABM5103, C. milleri ABM4949, Cryptococcus sp. R133 and C. laurentii R134.

Figure 2.

Viscosity of 3·5% oat bran solution after 72 h incubation with different yeasts compared with un-inoculated control. The values marked with an asterisk are means of two independent samples.

When some of the best folate-producing yeasts were incubated alone or together with LAB in oat bran solution, only C. milleri ABM4949, S. cerevisiae ALKO743, ABM5103 and ABM5131 made significant amounts of folates (Fig. 3). There was not much effect on folate production by the yeasts isolated from barley whether glucose was added or not, but the fermentative yeast S. cerevisiae ABM5131 responded to sugar addition (Fig. 3a). In subsequent experiments, 2% glucose was added to the oat bran solution (Fig. 3b). Yeast and L. rhamnosus LC-705 viable count values indicated that the microbes had grown in the oat bran solution samples—in the presence of glucose yeasts about 10–20-fold and L. rhamnosus LC-705 about 50-fold. Growth without glucose was poor—viable counts increased only 2–7-fold.

Figure 3.

Folate production ng per gram in 3·5% oat bran solution yeast fermentation and together with lactic acid bacteria Lactobacillus rhamnosus LC-705 and Streptococcus thermophilus ABM 5097. (a) No added glucose (image) or with 2% added glucose (image); Yeasts isolated from barley kernels: Aureobasidium pullulans R38, Pseudozyma sp. R47, Rhodotorula glutinis R48; Saccharomyces cerevisiae ABM 5131 was used as control. (b) With 2% added glucose(image); Yeasts isolated from diverse foodstuffs: Candida milleri ABM 4949, Clavispora lusitaniae ABM 5147, Galactomyces geotrichum ABM 5136, Saccharomyces cerevisiae ABM 5103, ABM 5131, ALKO 743.

Viscosity determinations showed that many oat bran samples had lost their viscosity partly or completely. All the samples containing yeasts isolated from barley kernels (R38, R47, R48) were either watery or of low viscosity (data not shown), but most of the samples with yeasts isolated from diverse foods had retained their high viscosity (Fig. 4). However, viscosity of the oat bran solution was partially destroyed by the S. cerevisiae ABM5131 and completely destroyed by Galactomyces geotrichum ABM5136 (Fig. 4).

Figure 4.

Dynamic viscosity of oat bran solution after 24 (image) and 72 h (image) fermentation by yeasts and lactic acid bacteria alone and in combinations. Error bars represent the standard deviation.

Discussion

Yeasts previously isolated from barley kernels have been found to belong to ascomycetous genera Aureobasidium, Candida, Debaryomyces, Geotrichum, Hansenula, Kloeckera, Saccharomyces, Torulopsis, Williopsis and to basidio-mycetous genera Bulleromyces, Cryptococcus, Filobasidium, Rhodotorula, Sporobolomyces and Trichosporon (Noots et al. 1998; Laitila et al. 2006). We found representatives of mostly basidiomycetous yeasts in our barley kernel samples (Table 1). The total yeast colony counts 6 × 103–4 × 104 g−1 were in close agreement to those found earlier—5 × 102–4·4 × 103 g−1 (Tuomi et al. 1995), 7 × 104–2 × 105 g−1 (Laitila et al. 2006) and 1·9–4·7 × 103 g−1 (Petters et al. 1988). We did not observe any significant difference in the yeast counts between surface and deeper layers contrary to Laca et al. (2006) who found higher cell counts at the surface layers compared with deeper layers in the grain.

As a new yeast in barley, we found a representative of still another basidiomycetous genus Pseudozyma; however, identification to the species level needs confirmation by other criteria.

The yeasts isolated from diverse foods were identified by rDNA sequencing (Table 1) to genera and species often found in the corresponding matrixes. Rye sour dough is known to contain mainly Candida milleri, but Saccharomyces cerevisiae or Saccharomyces exiguus yeasts are also often present (Mäntynen et al. 1999). Candida humilis has been reported to be dominant in wheat sour dough (Gullo et al. 2002). We found that both C. milleri ABM4949 and S. cerevisiae ABM5103 were present in the same rye sour dough sample. Candida milleri ABM5099 was found in spontaneously fermenting soya bean processing feed product, which contained also LAB and thus resembled sour dough fermentation. Kefir is known to contain Brettanomyces anomalus, Candida holmii, C. inconspicua, C. krusei, C. lipolytica, C. lambica, C. maris, Cryptococcus humicolus, Geotrichum candidum (teleomorph Galactomyces geotrichum), Kluyveromyces marxianus (anamorph C. kefyr), Pichia fermentans, Saccharomyces cerevisiae, S. exiguus, S. humaticus, S. turicensis, S. unisporus, Torulaspora delbrueckii and Zygosaccharomyces sp. yeasts depending on the origin of the kefir (Wyder et al. 1997, 1999; Simova et al. 2002; Witthuhn et al. 2004, 2005; Latorre-Garcia et al. 2007; Wang et al. 2008). Our kefir sample contained both K. marxianus ABM5130 and S. cerevisiae ABM5131 as the dominant species. Finnish viili contained Galactomyces geotrichum ABM5136, but the anamorph name Geotrichum candidum only is used by the commercial producers (Merilainen 1984). Spontaneously fermented apple cider contained Pichia membranifaciens ABM5031, which species has also been found in pilot fermentations and in cider plant must in Spain (Cabranes et al. 1990) as well as in tequila fermentation (Lachance 1994). Spontaneously fermenting lingonberry jam contained Kluyveromyces marxianus ABM5032 which was unexpected, as lingonberry juice is difficult to ferment due to high content of benzoic acid (Visti et al. 2003), and Kluyveromyces yeasts are not particularly resistant to benzoic acid (Warth 1988). A commercial fermented oat product contained the yeast Clavispora lusitaniae ABM5147 which seemed to be moderately fermentative in our tests with glucose and grew in the API32C tests well on rhamnose, which latter characteristic earlier has been considered as a potential diagnostic test for this yeast (Lachance and Phaff 1998). Clavispora lusitaniae, which has been determined to be one of the dominating species on agave plants, was not significant in tequila fermentation (Lachance 1994) but was one of the three main yeasts in mescal fermentation (Escalante-Minakata et al. 2008). Clavispora lusitaniae is also involved in cheese ripening (Kaminarides and Anifantakis 1989; El-Sharoud et al. 2009) and in whey and carrot-lemon juice fermentations (Sahota et al. 2010).

The plate tests for extracellular hydrolytic activities (Table 2) were designed to simulate degradation of the fermentation matrix components but seem not to be as sensitive as measurement of loss of viscosity in oat bran solution (Fig. 2). From control experiments (data not shown), we know that bacterial α-amylase treatment decreases the viscosity of oat bran solution by up to 50–80% and treatment with fungal β-glucanase by up to 50–90%. Thus, one would expect that amylases, cellulases and β-glucanases, if produced by the yeasts, should significantly reduce the viscosity of the oat bran solution. Only A. pullulans R38 and R124 strains had all the tested hydrolytic enzyme activities (Table 2) in agreement with recent literature (Li et al. 1993, 2007; Laitila et al. 2006; Ma et al. 2007; Liu et al. 2008), and reduced viscosity of 3·5% oat bran solution was consistently observed (Fig. 2). Also Pseudozyma sp. R47 with amylase, C. magnus R59 with cellulase and R. minuta R106 with β-glucanase activity reduced the oat bran solution viscosity. Surprisingly also C. adeliensis R43, R76 and R78, and R. glutinis R 48 and R63 reduced the oat bran solution viscosity (Fig. 2) even though the plate tests did not show any amylase, cellulase or β-glucanase activity (Table 2). However, it is possible that the physiological conditions on a solid plate are not as favourable as growth in liquid culture for the extracellular hydrolytic enzymes production. It is known that at least one strain of R. glutinis produced an extracellular endo-β-glucanase enzyme (Oikawa et al. 1998) and one C. adeliensis a xylanase (Scorzetti et al. 2000). Also the baker's yeast S. cerevisiae ALKO743 reduced the viscosity of the oat bran solution but did not show in plate assays any enzymatic activity. Later work has indicated that the viscosity-reducing activity by ALKO743 is best expressed in shake-flask cultures with 1–3·5% oat bran but without added glucose (data not shown).

The yeasts and LAB used in mixed culture fermentations were chosen in view of possible larger scale applications under industrially feasible conditions: growth at 28°C, nonaerated fermentation conditions and time of fermentation 1–3 days.

None of the assimilatory yeasts isolated from barley kernels made significant amounts of folate either alone or together with S. thermophilus or L. rhamnosus in oat bran solution (Fig. 3a). The fermentative yeast S. cerevisiae ABM5131 made folate when glucose was added to the oat bran solution. We recently reported results on oat and barley fermentations by pure cultures of the food yeasts ALKO743, ABM4949, ABM5131 and ABM5147 (Kariluoto et al. 2014). Significant increase in folate production was found with added glucose compared with plain matrix. S. thermophilus ABM5097 lowered the pH about 0·5 units less than L. rhamnosus LC-705 and consequently allowed better growth and folate production by the yeasts A. pullulans R38 and R. glutinis R48. If a daily dose of oat bran solution was 200 g similarly to yoghurt, it would mean that in the most favourable case a folate intake of 20 μg, or 10% of the recommended daily intake. The average folate content in oat bran solution with the best producers was 65 ng g−1, similar to the highest folate concentration 69 ng ml−1 obtained in Tanzanian fermented maize porridge togwa (Hjortmo et al. 2008).

Many assimilative yeasts, isolated from barley, produced considerable amounts of folate but destroyed the solid, viscous structure of the oat bran solution, indicating that they degraded the viscosity-generating soluble fibres, considered to be nutritionally advantageous. Many fermentative food yeasts also produced folate and did not reduce the viscosity or reduced it less radically (Figs 2 and 4)—they might be useful for further studies aiming at even higher folate concentrations.

Acknowledgements

This research was funded by the Academy of Finland as part of the project “FOLAFIBRE—Aqueous processing of oats and barley: In situ enhancement of folate and associated bioactive compounds while maintaining soluble dietary fibre physiologically active”.

Conflict of Interest

No conflict of interest declared.

Ancillary