International Journal of Food Microbiology 159 (2012) 144–151

Contents lists available at SciVerse ScienceDirect

International Journal of Food Microbiology

j ourna l homepage: www.e lsev ie r .com/ locate / i j foodmicro

Biodiversity and probiotic potential of yeasts isolated from Fura, a West Africanspontaneously fermented cereal

Line Lindegaard Pedersen a,⁎, James Owusu-Kwarteng b, Line Thorsen a, Lene Jespersen a

a Department of Food Science, Food Microbiology, Faculty of Sciences, University of Copenhagen, Denmarkb Department of Applied Biology, Faculty of Applied Sciences, University for Development Studies, P. O. Box 24, Navrongo, Ghana

⁎ Corresponding author at: Department of Food Sciensity of Copenhagen, Rolighedsvej 30, 1958 Frederiksber3158.

E-mail address: llp@life.ku.dk (L.L. Pedersen).

0168-1605/$ – see front matter © 2012 Elsevier B.V. Alhttp://dx.doi.org/10.1016/j.ijfoodmicro.2012.08.016

a b s t r a c t

a r t i c l e i n f o

Article history:Received 3 May 2012Received in revised form 10 July 2012Accepted 22 August 2012Available online 29 August 2012

Keywords:YeastFuraTransepithelial electrical resistanceCaco-2IPEC-J2Probiotic yeast

Fura is a spontaneously fermented pearl millet product consumed in West Africa. The yeast species involved inthe fermentation were identified by pheno- and genotypic methods to be Candida krusei, Kluyveromycesmarxianus, Candida tropicalis, Candida rugosa, Candida fabianii, Candida norvegensis and Trichosporon asahii.C. krusei and K. marxianus were found to be the dominant species. Survival in pH 2.5 or in the presence of bilesalts (0.3% (w/v) oxgall) and growth at 37 °Cwere independently determined as indicators of the survival poten-tial of the isolates during passage through the human gastrointestinal tract. Selected yeast species isolates wereassessed for their probiotic potential. All of the examined yeast isolates survived and grew at human gastrointes-tinal conditions in pH 2.5, 0.3% (w/v) oxgall at 37 °C. The effect on the transepithelial electrical resistance (TEER)across polarized monolayers of intestinal epithelial cells of human (Caco-2) and porcine (IPEC-J2) origin, weredetermined. The Caco-2 cells and IPEC-J2 cells displayed clearly different relative TEER results. The strainsof C. krusei, K. marxianus, C. rugosa and T. asahiiwere able to increase the relative TEER of Caco-2monolayersafter 48 h. In comparison, the relative TEER of IPEC-J2 monolayers decreased when exposed to the sameyeasts, even though T. asahii did not differ significantly from Saccharomyces cerevisiae var. boulardii whichis used as a human probiotic. C. tropicalis resulted in the largest relative TEER decrease for both thehuman and the porcine cell model assays. Hyphal growth was observed for C. albicans and C. tropicalisafter 48 h of incubation with polarized Caco-2 monolayers, whereas this was not the case for the remainingyeast species. In the present study new yeast strains with potential probiotic properties have been isolatedto be used potentially as starter cultures for fura production.

© 2012 Elsevier B.V. All rights reserved.

1. Introduction

Fura is a Ghanaian spontaneously fermented pearl millet basedproduct eaten together with spontaneously fermented milk (nono).Wet millet pearl millet is fermented for 12 h moreover moulded andcooked. The main processing steps include washing and dehulling,wet milling with aromatic ingredients, dough fermentation, mouldingand cooking, pounding to a sticky mass and a final moulding. The spon-taneous fermentation has in preliminarywork been found to be effectedby lactic acid bacteria and yeasts (Owusu-Kwarteng et al., 2010) as alsoseen in other acidic African fermented cereals (Padonou et al., 2009;Vieira-Dalode et al., 2007). The fermentation process is particularly im-portant in developing countries where refrigeration is not an option.Generally, fermentation prolongs the shelf-life, enhances the nutritionalvalue and reduces the risk of foodborne illnesses. Besides, both pro- andeukaryotic microorganisms involved in the fermentation might have

ce, Faculty of Sciences, Univer-g C, Denmark. Tel.: +45 35 33

l rights reserved.

beneficial effects on human health andwell-being. It is therefore obviousto explore whether the Africanmicrobial niches of fermented foods offeryeast cultures with improved impact on human health. Currently, theyeastmicroflora of fura has not previously been identified and the probi-otic potential has not been characterized.

Strains of Saccharomyces cerevisiae have beenwidely tested for probi-otic properties such as the protection of bacterial translocation and pres-ervation of gut barrier integrity (Generoso et al., 2010; Klingberg et al.,2008;Martins et al., 2007). Different non-Saccharomyces yeast species es-pecially of the genera Debaryomyces, Torulaspora, Kluyveromyces, Pichia,and Candida (Silva et al., 2011) have been shown to have probiotic poten-tial as judged by their ability to survive and colonize the gastrointes-tinal tract in different mammalian cell model assays. Additionally,probiotic yeasts may have inhibitory activity against pathogenic bac-teria (Kumura et al., 2004; Psani and Kotzekidou, 2006; Tiago et al.,2009). Nevertheless, S. cerevisiae var. boulardii is the only yeast withproven clinical effects and the only yeast with proven probiotic efficien-cy in double blind studies (Sazawal et al., 2006). It is used for preventionand treatment of many different types of human gastrointestinal dis-eases (Born et al., 1993; Buts et al., 2006; McFarland, 2007; Surawicz,2003; Zanello et al., 2011).

145L.L. Pedersen et al. / International Journal of Food Microbiology 159 (2012) 144–151

C. albicans is an opportunistic pathogenic yeast, but a benign mem-ber of the microflora of healthy individuals (Kleinegger et al., 1996).However, uncontrolled growth of C. albicans can occur and lead to pen-etration of the gastrointestinal mucosa (Mavor et al., 2005; Pittet et al.,1994). Hyphael growth appears to be the invasive form of the microor-ganism (Ray and Payne, 1988; Scherwitz, 1982). Pathogenic yeasts areobliged to survive the conditions in the gastrointestinal tract i.e. beingable to survive at 37 °C and in the presence of acidic gastric juice, alka-line pancreatic juice, lysozymes and bile salts (Mavor et al., 2005). De-pendent on the mode of action, this also accounts for probiotic yeasts(Tiago et al., 2009).

Human colon tumorigenic cell lines such as Caco-2, T84 and HT-29are recognized as models for host-beneficial bacteria/yeast-pathogeninteractions studies (Czerucka et al., 2000; Frank and Hostetter, 2007;Klingberg et al., 2008, 2005; Satish Kumar et al., 2011). The IntestinalPig Epithelial Cell Jejunum (IPEC-J2) cell line has been suggested as apossible improved in vitro model for studying pathogen/beneficial in-teractions of microorganisms, due to its similarity with the human gut(Langerholc et al., 2011; Lunney, 2011) and due to the fact that it isnon-transformed, non-tumorigenic and mucin producing (Cencic andLangerholc, 2010; Langerholc et al., 2011; Liu et al., 2010). The trans-epithelial electrical resistance (TEER) assay is a non-destructivemethodto monitor confluency, viability of tissue culture monolayer and thepresence of tight junctions (Geens and Niewold, 2011), and it hasbeen used for the investigation of pathogenic and probiotic effects(Klingberg et al., 2008, 2005). To our knowledge, comparable TEERmeasurements between a tumorigenic and a non-tumorigenic epitheli-al cell model line have never been obtained for a higher number of yeastspecies.

The aim of present work has been to identify the yeast species oc-curring during fermentation of fura and to characterize new yeaststrains with probiotic potential. The yeasts were tested for their abil-ity to survive gastrointestinal conditions such as high bile salt concen-tration and low pH. Furthermore their influence on the TEER of twodifferent cell lines, the mammalian epithelial colorectal adenocarci-noma cell line Caco-2 and a porcine jejunal mucin producing cellline IPEC-J2 were investigated and the two mammalian cell linescompared. A probiotic yeast strain of S. cerevisiae var. boulardii anda pathogenic yeast strain of C. albicans were included as internalcontrols.

2. Materials and methods

2.1. Yeast strains and growth conditions

Samples were taken during 12 h fermentation of fura and from thefinal product. In total, 330 yeast isolates were isolated. The samplingof furawas carried out in three different towns in the Upper East regionof Ghana in which fura is produced: Bolgatanga, Paga and Navrongo.Fura fermentationwas conducted by local producers at household scale.

Ten g of each sample was homogenized in 90 ml sterile diluent(1% peptone (Difco, Detroit, Michigan, USA), 0.85% NaCl, pH 7.0)using a stomacher (Stomacher-Bagmixer, Buch and Holm). Serial dilu-tions (10−1 to 10−9) were made for each sample and 0.1 ml of the ap-propriate dilutionwas inoculated ontoMalt Extract agar (MEA) (Merck,Darmstadt, Germany) containing 100 mg/l chloramphenicol (Fisherchemicals, C/4322/47, Loughborough, Leics, U.K) and 50 mg/l chlortet-racycline (SigmaAldrich, Schnelldorf, Germany); plateswere incubatedfor 72 h at 25 °C. After incubation, isolates were taken from a sector ofthe highest dilution and were further purified by successive streakingon MEA. All isolates were maintained at −80 °C in 20% (v/v) glycerol(Merck).

Isolates were propagated in YPD broth (10 g/l glucose (Merck),10 g/l Bacto Peptone (Becton, Dickinson and Company (BD), Le Pontde Claix, France), 5 g/l Bacto yeast extract (BD), pH 5.6±0.1) for 48 hat 30 °C. For the examination of the probiotic properties, C. albicans

CA-1 was included as a pathogenic reference strain. The strain was iso-lated from a clinical specimen and kindly provided by Statens SerumInstitut, Copenhagen, Denmark. S. cerevisiae var. boulardii 259–1 wasisolated from the commercial product Precosa, batch 259 (Logic aps,Lynge, Denmark), was used as probiotic reference strain. (Kuhle et al.,2005).

2.2. Identification of fura yeast isolates to species level by genotypic andphenotypic tests

2.2.1. Genotypic testsTotal DNA was extracted using the InstaGene Matrix DNA extrac-

tion kit following the instructions of the manufacturer (Bio-Rad,Hercules, CA, USA). The rep-PCR reaction was carried as describedby Nielsen et al. (2007). In brief, the rep-PCR reaction was carriedout in a 24 μl volume containing 1 U dreamTaq DNA polymerase(Fermentas GmbH, St. Leon-Rot, Germany), 2.5 μl 10× dreamTaqbuffer (Fermentas), 200 μM of each dNTP (Fermentas), 1.5 mMMgCl2 (Fermentas), 0.8 μM (GTG)5 (5′-GTG GTG GTG GTG GTG-3′,DNA Technologies, Aarhus, Denmark), 1.5 μL of DNA template, andsterile MilliQ water for adjustment of the volume to 25 μl. The reac-tion was performed in an automatic thermal cycler on a BiometraTrio-Thermoblock (Biotron, Göttingen, Germany) under the follow-ing thermocycling program: 5 min of initial denaturation at 94 °C,30 cycles of 94 °C for 30 s, 45 °C for 60 s, 65 °C for 8 min followedby a final elongation step of 65 °C for 16 min. The rep-PCR-profileswere normalized and cluster analysis performed using Bionumerics2.50 (Applied Maths, Sint-Martens-Latem, Belgium). The dendrogramswere calculated on the basis of the Dice's Coefficient of similarity withthe unweighted pair groupmethod with arithmetic averages clusteringalgorithm (UPGMA) (Nielsen et al., 2007). Based on this grouping rep-resentative isolates were selected for sequencing of the D1/D2 regionof the 26S rRNA gene using the primers: NL-1 (5´-GCA TAT CAA TAAGCG GAG GAA AAG-3´) and NL-4 (5´-GGT CCG TGT TTC AAG ACGG-3´) (Jespersen et al., 2005). The PCR products were sent for purifica-tion and sequencing to a commercial sequencing facility (MacrogenInc., Korea). Sequences were manually corrected and aligned using theprogram CLC Main Workbench 5.6.1 (CLC BIO, Aarhus, Denmark). Sub-sequently, the corrected sequences were aligned to 26S rRNA gene se-quences in the Genbank database using the BLAST algorithm (Altschulet al., 1997).

2.2.2. Phenotypic testsInitially, all yeast isolates were micro- and macro-morphological

characterized according to Kurtzman et al. (2011b). For the assimila-tion and fermentation test yeast colonies grown on MYGP agar plates(3 g/l Bacto malt extract (BD), 3 g/l Bacto yeast extract (BD), 10 g/l Glu-cose (Merck), 5 g/L Bacto peptone (BD), 20 g/L Bacto agar (BD), pH5.6±0.1) were diluted corresponding to the value 2+ on Wickerhamscard (Kurtzman et al., 2011b) in sterile MilliQ water and a volume of0.1 ml was used for inoculation. For the test of assimilation of nitrate(117.0 g/L Difco Carbon Base (BD)+7.8 g/l KNO3 (Merck) pH 5.6±), atube without KNO3 was used as a control, and the tubes were incubatedat 25 °C for 7 days. For testing the assimilation of carbohydrates (67 g/lDifco yeast nitrogen base+50 g/l carbohydrates (glucose (Merck), ga-lactose (Sigma Aldrich, ST. Louis, USA), maltose (Sigma Aldrich), sucrose(Sigma Aldrich), lactose (Merck) or raffinose (Sigma Aldrich) pH 5.6±),a tube without carbohydrate was used as negative control; the tubeswere incubated at 25 °C for 3–4 weeks.

For the fermentation of carbohydrates Durham tubes including thefollowing medium (4.5 g/l Bacto Yeast Extract (BD)+7.5 g/l Bacto Pep-tone (BD)+12 ml/l Bromthymol blue solution (0.4 g Bromthymolblue(Merck)+10 ml 0.1 M NaOH (Merck)+190 ml MilliQ water pH 6.0±0.2)) were autoclaved (121 °C for 15 min) with 5 ml and inoculated asdescribed above. Subsequently 2.5 ml of a sterilefiltered carbohydrate so-lution (glucose 60 g/l (Merck), galactose 60 g/l (Sigma Aldrich), maltose

146 L.L. Pedersen et al. / International Journal of Food Microbiology 159 (2012) 144–151

60 g/l (Sigma Aldrich), sucrose 60 g/l (Sigma Aldrich), lactose 60 g/l(Merck), or raffinose 120 g/l (Sigma Aldrich)) were added after cooling.The inoculated tubes were incubated at 25 °C for 1–2 weeks. Reading ofthe assimilation and fermentation was done according to Kurtzman etal. (2011b) and compared to the literature (Kurtzman et al., 2011a).

2.2.3. Nucleotide accession numbersThe nucleotide sequences determined in this study have been

assigned GenBank Accession numbers JN110345–JN110429@@.

2.3. Probiotic properties of fura yeasts

2.3.1. Tolerance to low pH and bile saltsSurvival and growth under gastrointestinal conditions were in-

vestigated by incubation at 37 °C in pH 2.5 and 0.3% (w/v) bilesalts, respectively. The experiment was performed in a 200 μl volume in100-microwell plates (Isotron, Ede, The Netherlands). The wells wereinoculated in triplicates with 106 yeast cells pregrown for 48 h at 30 °Cin yeast nitrogen base (YNB, 6.7 g/l yeast nitrogen base (BD), and 10 g/lglucose (Merck), pH 5.4). Cells were harvested by centrifugation(10 min at 3000 g at room temperature). YNB (pH2.5) or YNB containing0.3% (w/v) Oxgall (Difco)was added to the cells. YNBwithout inoculationwas used as a negative control, and inoculated YNBwas used as a positivecontrol. Changes in optical density were measured (Bioscreen C,Labsystem, Helsinki, Finland) every hour from 0 to 24 h of incuba-tion at 37 °C. The yeast growth was determined as the area underthe growth curve (OD600×h) obtained from the bioscreen data. Thegrowth rate for the yeasts in pH 2.5 and 0.3% (w/v) oxgall was calcu-lated relative to their growth rate in YNB, which was used as a posi-tive control of growth.

2.3.2. Growth and maintenance of mammalian cell culturesThe human colon adenocarcinoma cell line Caco-2 was purchased

from the Deutsche Sammlung von Mikroorganismen und Zellkulturen(DSMZ, Braunschweig, Germany) and grown inminimumessentialme-dium (Minimum Essential medium, GlutaMAX™, Invitrogen, Paisley,United Kingdom) supplemented with 10% (v/v) foetal bovine serum(FBS, Lonza, Basel, Switzerland), 1% (v/v) nonessential amino acids(Invitrogen, Gibco) and 50 μg ml−1 gentamicin (Invitrogen, Gibco).The cells were routinely grown in 75 cm2 culture flasks (BD, Falkon)at 37 °C in a humidified atmosphere of 5% CO2 and 95% air until a con-fluent monolayer was obtained. The culture medium was changedevery second day and the cells were passaged every 5–7 day bytrypsination with 1% trypsin-EDTA 10× (Invitrogen, Gibco). Cell pas-sages used were P55–57.

The porcine jejunum epithelial cell line IPEC-J2 was a kind gift fromProfessor Anthony Blikslager, North Carolina State University, USA. Thecells were grown in Dulbecco's Modified Eagle Medium: NutrientMixture F-12 (DMEM/F-12) (Invitrogen, Gibco) supplemented with100 mg/l streptomycin (Fluka Chemie GMbH, Steinheim, Switzerland),100 mg/L penicillin (Sigma Aldrich), 2 mM L-glutamine (Sigma Al-drich), and 10% (v/v) FBS (Lonza). The cells were grown under sameconditions as Caco-2 cells. Change of media and the trypsination waspreformed as described above for the Caco-2 cells. Cell passages usedwere P26–28.

2.3.3. Comparison of in vitro models and the effect of fura yeast strains ontransepithelial electrical resistance (TEER)

The probiotic ability to increase the TEER was studied for 5 furayeasts, and C. albicans ca-1 and S. cerevisiae var. boulardii 259–1were in-cluded as pathogenic and probiotic controls respectively. To obtain po-larizedmonolayers, Caco-2 and IPEC-J-2 cells were seeded on Corning®Transwell® polycarbonate membrane inserts (8.0 μm pore size, mem-brane diam. 6.5 mm, Corning Incorporated, Sigma Aldrich, Schnelldorf,Germany) at a density of 1×105 cells/ml. A volume of 100 μL cellgrowthmediumwas added to the inner chamber (apical compartment)

and 600 μl to the outer chamber (basolateral compartment). Caco-2cells were cultured on the membranes for 21 days after confluency wasobserved, in order to obtain cell differentiation (Briske-Andersen et al.,1997). IPEC-J2 cells were cultured for 16 days after confluency was ob-served in order to obtain a proper mucus layer as observed by the com-bined alcian blue-PAS technique (Moncada et al., 2003; Mowry, 1956).Yeast strains were added (106 cells/ml) to the apical compartment andincubated at 37 °C in a humidified atmosphere of 5% CO2 and 95% air.The effects of the yeast strains on cell polarity were evaluated by mea-surement of the TEER using theMillicell-ERS Electrical Resistance System(Millipore, Bedford, MA). The net value of the TEER (Ωcm−2) wascorrected for background resistance by subtracting the contributionof the cell free filter and the medium (140 Ωcm−2). The TEER wasmeasured before the addition of the yeast (t=0) and then at varioustime intervals (1, 8, 24 and 48 h) and expressed as the ratio of TEERat time t in relation to the initial value (at t=0) for each series. TheTEER of monolayers with Caco-2 or IPEC-J2 cell lines without addedyeast were used as controls. Each assay was conducted twice (twodifferent passages) with triplicate determination.

Images of yeast cells on the Caco-2 cell layer in the TEER trayswere recorded using a microscope (Zeiss Axiovert 25) equippedwith a Zeiss 40× objective 0.5 combined with a Canon Power ShotS90.

2.4. Statistics

If nothing else is stated, values are presented as themean+standarderror mean (SEM) of replicates. Differences in mean values were testedby ANOVA using the Tukey–Kramer test for significant differences be-tween groups using the program JMP 8.0.1 statistical discovery bySAS, and a p value of b0.05 was considered significant.

3. Results

3.1. Identification of fura yeasts by genotypic and phenotypic tests to spe-cies level

Fig. 1 shows the rep-PCR profiles and clusters of the yeast speciesoccurring during 12 h of fura fermentation and in the final product.The isolates clustered mainly in two large clusters, 1 and 2, and onesmall, cluster 3. Isolates of cluster 1 were identified by sequencingof the D1/D2 domain of the large subunit (26S) rDNA as C. kruseiwith 100.0% homology to GenBank sequences, being supported bythe assimilation and fermentation profile listed in Table 1. Cluster 2was identified by sequencing of the D1/D2 domain as K. marxianuswith 100.0% homology to GenBank sequences, though one isolateshowed 99.6% homology. The sequencing results were verified bythe assimilation and fermentation profiles, though all isolates ofK. marxianus had faint positive assimilation of maltose, which doesnot correspond to the literature (Kurtzman et al., 2011a). Cluster 3was identified by sequencing of the D1/D2 domain as T. asahii with100.0% homology to GenBank sequences,whichwere verified by the as-similation and fermentation profiles, though T. asahii had faint positiveassimilation of raffinose, which does not correspond to the literature(Kurtzman et al., 2011a).

The remaining isolates that did not cluster in the three rep-PCRgroups (1–3) were identified by sequencing of the D1/D2 domain asC. norvegensis, C. fabianii, C. rugosa and C. tropicalis showing 100.0%,99.8%, 99.8% and 100.0% homology to GenBank sequences, respective-ly. Relatively few isolates were identified as these species. The assim-ilation and fermentation profiles of these species corresponded to theliterature (Kurtzman et al., 2011a), though C. rugosa had faint positiveassimilation of maltose. Within these clusters a representative num-ber of isolates were analysed as shown in Supplementary materialTable S1. In total 85 yeast isolates were sequenced. For further detailson the sequence analyses see Supplementary material, Table S1.

Strain Group Identity*

1

2

Candida norvegensis

100

90807060

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

f2 fp21

f2 0h13

f1 10h5

f2 0h1

f2 10h6

f2 10h9

f2 0h16

f2 10h5

f2 12h6

f2 10h15

f2 10h28

f1 12h19

f112h23

f1 0h16

f1 0h2

f1 0h3

f1 0h4

f1 0h30

f1 fp22

f1 4h9

f1 4h16

f1 fp8

f1 2h14

f1 2h17

f1 6h2

f1 fp26

f2 10h13

f1 8h19

f2 fp12

f2 fp13

f2 fp4

3

Candida fabianii

Candida tropicalis

Candida krusei

Kluyveromyces marxianus

Candida rugosa

Trichosporon asahii

*

*

*

*

*

*

*

Fig. 1. Dendogram obtained by cluster analysis of rep-PCR (GTG5) fingerprints for yeasts isolated during traditional fura processing in Ghana. The dendogram is based on Dice'scoefficient of similarity with the unweighted pair group method with arithmetic average clustering algorithm (UPGMA). Only a sub-sample of representative sequenced isolatesis shown. Identity is by sequencing of the 26S rRNA gene and additional tests as described in text.

147L.L. Pedersen et al. / International Journal of Food Microbiology 159 (2012) 144–151

At all production sites i.e. in Navrongo, Paga and Bolgatanga, thepredominant yeast species were C. krusei and K. marxianus, account-ing in average for respectively 60% and 38% of the yeast population

Table 1Assimilation of carbohydrates, nitrate and fermentation of carbohydrates of all se-quenced yeast strains isolated from fura fermentations and final product.

Yeast species

C.krusei

K.marxianus

C.rugosa

C.tropicalis

C.norvegensis

C.fabianii

T.asahii

AssimilationGlucose + + + + + + +Galactose – + + + – – +Maltose – ± ± + – + +Sucrose – + – + – + +Lactose – + – – – – +Raffinose – + – – – + ±+Nitrate – – – – – + –

–Nitrate – – – – – – +

FermentationGlucose + + – + + + –

Galactose – + – + – – –

Maltose – – – + – + –

Sucrose – + – + – + –

Lactose – – – – – – –

Raffinose – + – – – + –

+: growth, ±: faint growth, –: no growth.

during the fermentation and in average for respectively 73% and 15%of the yeast population in the final product (Table 2). C. krusei andK. marxianus were the only species found in Navrongo. C. fabianii andC. rugosa occurred only once during the fermentation in Bolgatanga.C. norvegensis and T. asahii were isolated from the final product inBolgatanga and T. asahii also in Paga. C. tropicalis were isolated dur-ing fermentation and from the final product in Paga and Bolgatanga.

3.2. Probiotic properties of fura yeasts

3.2.1. Tolerance to acid and bile at 37 °CThe ability to survive the passage through the gastrointestinal

tract was investigated for randomly selected isolates of the followingspecies C. krusei, K. marxianus, T. asahii, C. tropicalis, C. norvegensis,C. rugosa and C. fabianii. Results for the tolerance to acid and bilesalt at 37 °C for the strains are listed in Table 3. The probiotic strainS. cerevisiae var. boulardii and the pathogenic C. albicanswere includedas references. All the tested strains were tolerant to low pH, as growthwas observed at pH 2.5, though C. fabianii and T. asahii grew ratherslowly. Further all of the tested strains were tolerant to bile as growthwas observed in 0.30% (w/v) oxgall.

3.2.2. Influence of fura yeast strains on transepithelial electrical resis-tance (TEER) of Caco-2 and IPEC-J2 epithelial cell lines

The ability to increase the relative TEER of a monolayer of polar-ized intestinal epithelial cells was investigated, for the same selected

Table 2Occurrence of yeast species in Bolgatanga, Navrongo and Paga isolated from the fermentation and in the final product, fura.

Production sites Bolgatanga Navrongo Paga No of sequenced isolates

Fermentation Final product Fermentation Final product Fermentation Final product

Yeast species (ratio %)C. krusei 51 56 77 85 51 78 54K. marxianus 44 20 23 15 47 11 20C. rugosa 1 1C. tropicalis 3 4 3 3C. fabianii 2 6 1C. norvegensis 8 2T. asahii 12 6 4

–: species not isolated.

148 L.L. Pedersen et al. / International Journal of Food Microbiology 159 (2012) 144–151

fura yeast strains as tested for the tolerance to acid and bile at 37 °C,using a human cell line (Caco-2) and a porcine cell line (IPEC-J2).C. fabianii and C. norvegensis was though not included in this assay asthey were found in low counts during the fermentation and showedslow growth in the cell line media (results not shown). Results areshown in Figs. 2 and 3 respectively.

As seen in Fig. 2, after 1 h, 8 h and 24 h no significant differencesin the relative TEER of the polarized Caco-2 monolayer were observedbetween the yeast strains and the Caco-2 control (Caco-2 cells with-out yeast). After 48 h the relative TEER of the polarized Caco-2 mono-layers inoculated with C. albicans and C. tropicalis was significantlydecreased compared to the Caco-2 control. For the remaining yeaststrains no significant differences were observed when compared tothe Caco-2 control, however, a slight but non-significant increase inthe relative TEER was observed for Caco-2 cells inoculated withK. marxianus, T. asahii, C. krusei and C. rugosa, respectively. For the latter,the relative TEER was even significantly higher than the relative TEERobserved for Caco-2 cells inoculated with S. cerevisiae var. boulardii.

As seen in Fig. 3, for the IPEC-J2 cells no significant differences inrelative TEER were observed after 1 h for any of the yeast strains ascompared to IPEC-J2 control (IPEC-J2 cells without yeast). Significantdifferences were observed after 8 h where a slight but significant in-crease in the relative TEER could be observed for the IPEC-J2 cells in-oculated with either S. cerevisiae var. boulardii or T. asahii. After 24 h,the relative TEER was significantly lower for the IPEC-J2 cells inocu-lated with all other yeast strains than S. cerevisiae var. boulardii andT. asahii, respectively. After 48 h S. cerevisiae var. boulardii was theonly yeast species able to maintain a relative TEER similar to theIPEC-J2 control. The largest decrease in TEER was observed for IPEC-J2cells inoculated with C. tropicalis and C. albicans, with TEER valuesbeing very low i.e. 0.006 and 0.007 respectively (Fig. 3). All investigated

Table 3Growth at pH 2.5 adjusted with HCl and bile salt (0.3% (w/v) oxgall) of nine yeaststrains; seven isolated from fura; the probiotic strain S. cerevisiae var. boulardii 259and the pathogenic C. albicans Ca-1. The growth rate of the yeasts at pH 2.5 and in0.3% (w/v) oxgall was calculated relative to their growth rate in YNB, which wasused as a positive control for growth.

Isolates Growth in pH 2.5* Growth in 0.3% oxgall

C. krusei f1 0h4 +++ ++++K. marxianus f1 0h30 ++ ++++C. tropicalis f110h5 ++ +++C. rugosa f1 8h19 + ++++C. norvegensis f2 fp21 ++++ +++C. fabianii f2 0h13 ± ++T. asahii f2 fp4 ± ++++C. albicans Ca-1 +++ +++S. cerevisiae var. boulardii 259 + +++

The tested strains are marked with a * in Fig. 1.*Survival was observed for all strains after 4 h.++++: 140% or more of the growth rate in YNB, +++: 100–139% of the growthrate YNB, ++: 60–99% of the growth rate YNB, +: 20–59% of the growth rate YNB,±: 1–19% of the growth rate YNB, –: 0% or less of the growth rate YNB.

yeast strains were able to grow during both cell line assays, althoughtheirmorphology varied. Thus as determined by use of lightmicroscopysome of the yeast cultures i.e. C. tropicalis and C. albicans did even showhyphal growth,whereas e.g. S. cerevisiae var. boulardii showed only uni-cellular growth (results not shown).

4. Discussion

In the present study on fura an African indigenous fermented cerealfood product C. krusei and K. marxianuswere found to be the dominantyeast species during fermentation and in the final product. C. krusei haspreviously been reported to be one of the predominant yeast species inAfrican indigenous fermented cereals (Jespersen et al., 1994; Jespersen,2003; Mugula et al., 2003). These fermentations are often initiated bylactic acid bacteria and various yeast species including Candida spp.,S. cerevisiae and others (Jespersen, 2003). One of these products iskenkey an indigenous fermented maize dough product (Jespersenet al., 1994). At the end of the fermentation C. krusei dominatesover S. cerevisiae as C. krusei has been shown to be more tolerant tohigh lactic acid concentrations at the at the end of the fermentationwhere pH it rather low (3.70) (Halm et al., 2004). S. cerevisiae was notisolated from fura even thoughmilled pearl millet should be supportiveof the growth of S. cerevisiae (Khetarpaul and Chauhan, 1991).In fermented togwa produced by using various cereals in Tanzania,C. krusei occurred in highest numbers, followed by S. cerevisiae,C. pelliculosa and C. tropicalis (Mugula et al., 2003). As already men-tioned, C. krusei was found to be the most predominant yeast speciesfollowed by K. marxianus during fura fermentation. K. marxianus hasbeen found to be of great importance in other African fermentedproducts such as gowé (Vieira-Dalode et al., 2007) and fermentedmilk (Narvhus and Gadaga, 2003). The isolation of the species ofC. tropicalis, C. norvegensis, C. rugosa and T. asahii in low numbersduring fura fermentation corresponds to what previously havebeen recorded for other African indigenous fermented foods(Gadaga et al., 2000; Hellstrom et al., 2010; Mugula et al., 2003;Padonou et al., 2009). Pichia fabianii (C. fabianii) has earlier beenisolated from an Indian traditional solid state starter called‘hamei’ used for production of rice wine (Jeyaram et al., 2008). Itis usually presumed that unknown yeasts can be identified atspecies level, when the percentage of homology of the sequenceis more than or similar to 99% (Kurtzman and Robnett, 1998). Allsequenced strains in this study exhibited a homology percentagemore than 99% to other yeasts species indicating that no newspecies were isolated from these fura fermentations.

All yeast strains survived and were even able to grow under con-ditions simulating passage through the human gastrointestinal tract.The growth rate did though differ between the species. In generalyeasts associated with food, have been found by other authors tohave a fairly good survival under conditions simulating passage throughthe human gastrointestinal tract (Kuhle et al., 2005; Kumura et al.,2004; Psomas et al., 2001; Trotta et al., 2012).

AA

BA

BA

A A

BA

B

A

A

BA

BA

A

A

BA

BA

AA

A

BA

AA

A

A

AA

BA

C

AA

B

C

0

0.5

1

1.5

2

2.5

482481

Rel

ativ

e T

EE

R (

TE

ER

t/T

EE

Rt 0

)

Time (hours)

Control (Caco-2 without yeast) Saccharomyces cerevisiae var. boulardii 259Kluyveromyces marxianus f1 0h 30 Trichosporon asahii f2 fp 4Candida krusei f1 0h 4 Candida rugosa f1 8h 19Candida tropicalis f2 10h 5 Candida albicans Ca-1

Fig. 2. Relative TEER of polarized Caco-2 monolayers; Control (Caco-2 without yeast) and Caco-2 cells added the following yeast strains: Saccharomyces cerevisiae var. boulardii 259,Kluyveromyces marxianus f1 0 h 30, Trichosporon asahii f2 fp 4, Candida krusei f1 0 h 4, Candida rugosa f1 8 h 19, Candida tropicalis f1 10 h 5 and Candida albicans ca-1. TEER isexpressed as the ratio TEER (at time (t) in relation to the initial value at time 0 h (t0)). Values are given as the mean+standard error of two independent experiments (two differentpassages) in triplicate. The Tukey–Kramer test was used for post hoc comparison between isolates. TEER values that are not significantly different have the same capital letters inthe figure, within the time point. A value of Pb0.05 was considered to indicate statistical significance.

149L.L. Pedersen et al. / International Journal of Food Microbiology 159 (2012) 144–151

As seen from the current results rather big differences wereobserved between the results obtained for the Caco-2 cells and theIPEC-J2 cell lines. However, conclusive results were obtained for

A

DC

A

A

A

CB

A

BA

A

D

A

CB

A

DC

A

D

Time (

Control (IPEC-J2) without yeast)Kluyveromyces marxianus f1 0h 30Candida krusei f1 0h 4Candida tropicalis f1 10h 5

2.5

2

1.5

1

0.5

0

1 8

Rel

ativ

e T

EE

R (

TE

ER

t/T

EE

Rt 0

)

Fig. 3. Relative TEER of polarized IPEC-J2 monolayers; Control (IPEC-J2 without yeast) andKluyveromyces marxianus f1 0 h 30, Trichosporon asahii f2 fp 4, Candida krusei f1 0 h 4, Caexpressed as the ratio TEER (at time (t) in relation to the initial value at the time 0 h (t0))different passages) in triplicate. The Tukey–Kramer test was used for post hoc comparisonin the figure, within the time point. A value of Pb0.05 was considered to indicate statistica

C. albicans and C. tropicalis as both yeast species lead to significantdecreases in the relative TEER for both intestinal epithelial cellmodels. T. asahii, C. rugosa, K. marxianus and C. krusei increased

AA

A BA

CB

D

BA

BDC

D

CB

CD

DD

D

hours)

Saccharomyces cerevisiae var. boulardii 259Trichosporon asahii f2 fp 4Candida rugosa f1 8h 19Candida albicans Ca-1

24 48

IPEC-J2 added the following yeast strains: Saccharomyces cerevisiae var. boulardii 259,ndida rugosa f1 8 h 19, Candida tropicalis f1 10 h 5 and Candida albicans ca-1. TEER is. Values are given as the mean+standard error of two independent experiments (twobetween isolates. TEER values not significantly different have the same capital lettersl significance.

150 L.L. Pedersen et al. / International Journal of Food Microbiology 159 (2012) 144–151

the relative TEER of the polarized Caco-2monolayers but decreased therelative TEER of polarized IPEC-J2monolayers. The probiotic S. cerevisiaevar. boulardii produced the lowest increase in relative TEER of the polar-ized Caco-2 monolayer after 48 h, but the highest relative TEER after48 h of the polarized IPEC-J2 monolayer. A previous study has demon-strated the functional resemblance of IPEC-J2 cells to the Caco-2 cellline, as they both are covered by typical brush border microvilli atconfluency, exhibiting a structural and functional differentiation pat-tern characteristic of mature enterocytes (Geens and Niewold, 2011).The differences observed in the present study can be due tomanydiffer-ent factors. The two cell lines were cultured in different lengths of timebefore adding the yeast strains. Caco-2 cells for 21 days in order to ob-tain differentiated cells (Briske-Andersen et al., 1997) and IPEC-J2 cellsfor 16 days in order to obtain a mucin layer (Schierack et al., 2006).However, preliminary studies with different culture time and passagesof the two cell lines indicated that the results obtained were notinfluenced by the culture time (results not shown). Further, therewere no differences in growth of the yeasts in the different cell linemedia (results not shown), although differences in metabolite pro-ductions which could affect the status of the cell lines were not in-vestigated. Even though, adhesion appears not to be a prerequisitefor probiotic yeasts as for e.g. lactic acid bacteria (Klingberg et al.,2008; Kuhle et al., 2005), differences in adhesion patterns couldalso be one of the reasons for the deviations between the resultsobtained for the two epithelial cell lines. Differences in adhesionhave previously been indicated for S. cerevisiae in two differentstudies conducted on IPEC-J2 and Caco-2 cell lines respectively(Klingberg et al., 2008; Kuhle et al., 2005). Also different immuno-logical responses may be generated due to the fact that IPEC-J2 is aporcine cell line whereas Caco-2 is a human cell line.

The strains of C. krusei, K. marxianus, C. rugosa and T. asahii demon-strated a protective effect on the Caco-2 cell line as observed by anincrease in relative TEER of Caco-2 monolayers after 48 h. Strains ofK. marxianus and C. krusei have earlier been isolated from kefir afermented milk product with health-promoting properties(Latorre-Garcia et al., 2007; Romanin et al., 2010), but C. rugosaand T. asahii have not been associated with any health beneficial ef-fects before. The fura yeast C. tropicalis caused the largest relative de-crease in TEER of both cell lines. The negative effect on the cell linesexerted by this yeast species was similar to the negative effect pro-duced by C. albicans. The decrease in relative TEER of polarizedCaco-2 monolayers inoculatedwith C. albicans corresponds to the lit-erature (Frank and Hostetter, 2007). Hyphal growth appears to bethe invasive form of C. albicans, as most of the intracellular organismsare found as hyphae, whereas yeast with unicellular growth typically arelocated between or on the surface of the epithelial cells (Dalle et al.,2010; Ray and Payne, 1988; Scherwitz, 1982). In this study, hyphalgrowth was observed both for C. albicans and C. tropicalis after 48 h ofco-incubationwith Caco-2monolayers. This indicates that the C. tropicalisisolated from fura exerts the same potential as C. albicans to invade epi-thelial cells. C. tropicalis has been found to be one of the three most com-monly non-albicans Candida species isolated from yeast infections(Krcmery and Barnes, 2002), but to our knowledge, previously TEERexperiments have not been reported for C. tropicalis.

Due to incidences of infections in immune-compromised patientsby Candida spp., only K. marxianus of the fura yeasts in this study areon the QPS list (Qualified Presumption of Safety) developed by theEuropean Food Safety Authority (EFSA) (EFSA, 2010). The risk factorsof the fura yeast strains should therefore carefully be investigated beforeany use as probiotics. In the present study, S. cerevisiae var. boulardiiwasincluded as a probiotic reference strain. S. cerevisiae is on the QPS list butS. cerevisiae var. boulardii is contraindicated for patients of fragile health,as well as for patients with a central venous (EFSA, 2010). In this study,the relative TEER of the two cell lines following exposure to S. cerevisiaevar. boulardii showed no differences after 48 h as compared to the celllines without added yeast. Similar TEER results were obtained by others

authors when human intestinal T84 epithelial cells were exposed to astrain of S. cerevisiae var. boulardii (Czerucka et al., 2000). Differentstrains of S. cerevisiae var. boulardiimay show different properties, as an-other strain previously was shown to increase the TEER of polarizedCaco-2 monolayers as compared to the control (Klingberg et al., 2008).Generally, S. cerevisiae strains, whether of clinical or nonclinical origin,do not seem to exert the pathogenic ability to disrupt polarized epithelialcell layers (Klingberg et al., 2008).

In conclusion, the present study identifies for the first time the yeastmicrobiota of the African spontaneously fermented cereal fura sampledat different production sites, by the use of molecular methods. C. kruseiand K. marxianuswere the predominant yeast species occurring duringfura fermentation and in the final product. Additionally 5 other yeastspecies were identified. Selected yeast strains were capable of survivaland growth under gastrointestinal like conditions. In addition, two dif-ferent cell lines were used to assess the probiotic potential of selectedisolates. Our results demonstrate that C. krusei, K. marxianus, C. rugosaand T. asahii isolated from fura were capable of increasing the TEER ofthe human Caco-2 cell line, whereas this was not the case for the por-cine IPEC-J2 cell line. In both cell line assays, C. tropicalis decreasedTEER in a manner similar to that of the pathogenic reference strainC. albicans. The probiotic reference strain S. cerevisiae var. boulardii wasable tomaintain TEER for 48 h in both cell line assays. The study demon-strates the different effects various yeast strains may have on the twoepithelial cell lines Caco-2 and IPEC-J2. Overall, the study has identifiedsome yeast candidates to be used as starter cultures with potentialprobiotic properties. However, due to the very different traits of theyeast species identified, their impact on human health and well-beingshould be thoroughly investigated before their commercial use.

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.ijfoodmicro.2012.08.016.

Acknowledgment

The grant from University of Copenhagen, Faculty of Life Scienceand the financial support from Danida (Danish International Develop-ment Agency), Ministry of Foreign Affairs, Denmark is highly appreci-ated. A special thank is directed to the production sites in Ghana andthe Faculty of Applied Sciences, University for Development Studies,Navrongo, Ghana.

References

Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W., Lipman, D.J.,1997. Gapped BLAST and PSI-BLAST: a new generation of protein database searchprograms. Nucleic Acids Research 25, 3389–3402.

Born, P., Lersch, C., Zimmerhackl, B., Classen, M., 1993. (The Saccharomyces boulardii ther-apy of HIV-associated diarrhea). [German] Deutsche Medizinische Wochenschrift118, 765.

Briske-Andersen, M.J., Finley, J.W., Newman, S.M., 1997. The influence of culture timeand passage number on the morphological and physiological development ofCaco-2 cells. Proceedings of the Society for Experimental Biology and Medicine214, 248–257.

Buts, J.P., Dekeyser, N., Stilmant, C., Delem, E., Smets, F., Sokal, E., 2006. Saccharomycesboulardii produces in rat small intestine a novel protein phosphatase that inhibitsEscherichia coli endotoxin by dephosphorylation. Pediatric Research 60, 24–29.

Cencic, A., Langerholc, T., 2010. Functional cell models of the gut and their applicationsin food microbiology – a review. International Journal of Food Microbiology 141,S4–S14.

Czerucka, D., Dahan, S., Mograbi, B., Rossi, B., Rampal, P., 2000. Saccharomyces boulardiipreserves the barrier function and modulates the signal transduction pathway in-duced in enteropathogenic Escherichia coli-infected T84 cells. Infection and Immu-nity 68, 5998–6004.

Dalle, F., Wachtler, B., L'Ollivier, C., Holland, G., Bannert, N., Wilson, D., Labruere, C.,Onnin, A., Ube, B., 2010. Cellular interactions of Candida albicans with human oralepithelial cells and enterocytes. Cellular Microbiology 12, 248–271.

EFSA, 2010. Scientific opinion on the maintenance of the list of QPS biological agentsintentionally added to food and feed (2010 update). EFSA Journal 8, 1–56.

Frank, C.F., Hostetter, M.K., 2007. Cleavage of E-cadherin: a mechanism for disruptionof the intestinal epithelial barrier by Candida albicans. Translational Research149, 211–222.

151L.L. Pedersen et al. / International Journal of Food Microbiology 159 (2012) 144–151

Gadaga, T.H., Mutukumira, A.N., Narvhus, J.A., 2000. Enumeration and identification ofyeasts isolated from Zimbabwean traditional fermented milk. International DairyJournal 10, 459–466.

Geens, M., Niewold, T., 2011. Optimizing culture conditions of a porcine epithelial cellline IPEC-J2 through a histological and physiological characterization. Cytotechnol-ogy 63, 415–423.

Generoso, S., Viana, M., Santos, R., Martins, F., Machado, J., Arantes, R., Nicoli, J., Correia,M., Cardoso, V., 2010. Saccharomyces cerevisiae strain UFMG 905 protects againstbacterial translocation, preserves gut barrier integrity and stimulates the immunesystem in a murine intestinal obstruction model. Archives of Microbiology 192,477–484.

Halm, M., Hornbaek, T., Arneborg, N., Sefa-Dedeh, S., Jespersen, L., 2004. Lactic acid tol-erance determined by measurement of intracellular pH of single cells of Candidakrusei and Saccharomyces cerevisiae isolated from fermented maize dough. Interna-tional Journal of Food Microbiology 94, 97–103.

Hellstrom, A.M., Vazques-Juarez, R., Svanberg, U., Andlid, T.A., 2010. Biodiversity andphytase capacity of yeasts isolated from Tanzanian togwa. International Journalof Food Microbiology 136, 352–358.

Jespersen, L., 2003. Occurrence and taxonomic characteristics of strains of Saccharomycescerevisiae predominant in African indigenous fermented foods and beverages. FEMSYeast Research 3, 191–200.

Jespersen, L., Halm, M., Kpodo, K., Jakobsen, M., 1994. Significance of yeasts and mouldsoccurring in maize dough fermentation for kenkey production. International Jour-nal of Food Microbiology 24, 239–248.

Jespersen, L., Nielsen, D.S., Honholt, S., Jakobsen, M., 2005. Occurrence and diversity ofyeasts involved in fermentation of West African cocoa beans. FEMS Yeast Research5, 441–453.

Jeyaram, K., Singh, W.M., Capece, A., Romano, P., 2008. Molecular identification of yeastspecies associated with Hamei A traditional starter used for rice wine productionin Manipur, India. International Journal of Food Microbiology 124, 115–125.

Khetarpaul, N., Chauhan, B.M., 1991. Biological utilization of pearl-millet flour fermentedwith yeasts and Lactobacilli. Plant Foods for Human Nutrition 41, 309–319.

Kleinegger, C.L., Lockhart, S.R., Vargas, K., Soll, D.R., 1996. Frequency, intensity, species,and strains of oral Candida vary as a function of host age. Journal of Clinical Micro-biology 34, 2246–2254.

Klingberg, T.D., Pedersen, M.H., Cencic, A., Budde, B.B., 2005. Application of measurementsof transepithelial electrical resistance of intestinal epithelial cell monolayers to eval-uate probiotic activity. Applied and Environmental Microbiology 71, 7528–7530.

Klingberg, T.D., Lesnik, U., Arneborg, N., Raspor, P., Jespersen, L., 2008. Comparison ofSaccharomyces cerevisiae strains of clinical and nonclinical origin by molecular typingand determination of putative virulence traits. FEMS Yeast Research 8, 631–640.

Krcmery, V., Barnes, A.J., 2002. Non-albicans Candida spp. causing fungaemia: pathoge-nicity and antifungal resistance. The Journal of Hospital Infection 50, 243–260.

Kuhle, A.V.D.A., Skovgaard, K., Jespersen, L., 2005. In vitro screening of probiotic prop-erties of Saccharomyces cerevisiae var. boulardii and food-borne Saccharomycescerevisiae strains. International Journal of Food Microbiology 101, 29–39.

Kumura, H., Tanoue, Y., Tsukahara, M., Tanaka, T., Shimazaki, K., 2004. Screening of dairyyeast strains for probiotic applications. Journal of Dairy Science 87, 4050–4056.

Kurtzman, C.P., Robnett, C.J., 1998. Identification and phylogeny of ascomycetousyeasts from analysis of nuclear large subunit (26S) ribosomal DNA partial se-quences. Antonie Van Leeuwenhoek International Journal of General and Molecu-lar Microbiology 73, 331–371.

Kurtzman, C.P., Feel, J.W., Boekhout, T., 2011a. Summary of Species Charateristics. In:Kurtzman, C.P., Fell, J.W., Boekhout, T. (Eds.), The Yeasts: a Taxonomic Study.Elsevier, London, pp. 223–278.

Kurtzman, C.P., Feel, J.W., Boekhout, T., Robert, V., 2011b. Methods for Isolation, Pheno-typic Characterization and Maintenance of yeasts. In: Kurtzman, C.P., Feel, J.W.,Boekhout, T. (Eds.), The Yeast: a Taxonomic Study. Elsevier, London, pp. 87–110.

Langerholc, T., Maragkoudakis, P.A., Wollgast, J., Gradisnik, L., Cencic, A., 2011. Novel andestablished intestinal cell line models — An indispensable tool in food science andnutrition. Trends in Food Science and Technology 22 (Supplement 1), S11–S20.

Latorre-Garcia, L., del Castillo-Agudo, L., Polaina, J., 2007. Taxonomical classification ofyeasts isolated from kefir based on the sequence of their ribosomal RNA genes.World Journal of Microbiology and Biotechnology 23, 785–791.

Liu, F., Li, G., Wen, K., Bui, T., Cao, D., Zhang, Y., Yuan, L., 2010. Porcine small intestinal ep-ithelial cell line (IPEC-J2) of rotavirus infection as a newmodel for the study of innateimmune responses to rotaviruses and probiotics. Viral Immunology 23, 135–149.

Lunney, J.K., 2011. Advances in swine biomedical model genomoics. International Jour-nal of Biological Sciences 3, 179–184.

Martins, F.S., Rodrigues, A.C.P., Tiago, F.C.P., Penna, F.J., Rosa, C.A., Arantes, R.M.E., Nardi,R.M.D., Neves, M.J., Nicoli, J.R., 2007. Saccharomyces cerevisiae strain 905 reducesthe translocation of Salmonella enterica serotype Typhimurium and stimulatesthe immune system in gnotoblotic and convectional mice. Journal of Medical Mi-crobiology 56, 352–359.

Mavor, A.L., Thewes, S., Hube, B., 2005. Systemic fungal infections caused by Candidaspecies: epidemiology, infection process and virulence attributes. Current DrugTargets 6, 863–874.

McFarland, L.V., 2007. Meta-analysis of probiotics for the prevention of traveler's diar-rhea. Travel Medicine and Infectious Disease 5, 97–105.

Moncada, D.M., Kammanadiminti, S.J., Chadee, K., 2003. Mucin and Toll-like receptorsin host defense against intestinal parasites. Trends in Parasitology 19, 305–311.

Mowry, R.W., 1956. Alcian blue technics for the histochemical study of acidic carbohy-drates. The Journal of Histochemistry and Cytochemistry 4, 407.

Mugula, J.K., Nnko, S.A.M., Narvhus, J.A., Soerhaug, T., 2003. Microbiological and fer-mentation characteristics of togwa, a Tanzanian fermented food. InternationalJournal of Food Microbiology 80, 187–199.

Narvhus, J.A., Gadaga, T.H., 2003. The role of interaction between yeasts and lactic acidbacteria in African fermented milks: a review. International Journal of Food Micro-biology 86, 51–60.

Nielsen, D.S., Teniola, O.D., Ban-Koffi, L., Owusu, M., Andersson, T.S., Holzapfel, W.H.,2007. The microbiology of Ghanaian cocoa fermentations analysed using culture-dependent and culture-independent methods. International Journal of FoodMicro-biology 114, 168–186.

Owusu-Kwarteng, J., Tano-Debrah, K., Glover, R.L.K., Akabanda, F., 2010. Process char-acteristics and microbiology of fura produced in Ghana. Nature and Science 8,41–51.

Padonou, S.W., Nielsen, D.S., Hounhouigan, J.D., Thorsen, L., Nago, M.C., Jakobsen, M.,2009. The microbiota of Lafun, an African traditional cassava food product. Interna-tional Journal of Food Microbiology 133, 22–30.

Pittet, D., Monod, M., Suter, P.M., Frenk, E., Auckenthaler, R., 1994. Candida colonizationand subsequent infections in critically ill surgical patients. Annals of Surgery 220,751–758.

Psani, M., Kotzekidou, P., 2006. Technological characteristics of yeast strains and theirpotential as starter adjuncts in Greek-style black olive fermentation. World Journalof Microbiology and Biotechnology 22, 1329–1336.

Psomas, E., Andrighetto, C., Litopoulou-Tzanetaki, E., Lombardi, A., Tzanetakis, N., 2001.Some probiotic properties of yeast isolates from infant faeces and Feta cheese. In-ternational Journal of Food Microbiology 69, 125–133.

Ray, T.L., Payne, C.D., 1988. Scanning electron-microscopy of epidermal adherence andcavitation in murine candidiasis — a role for Candida acid proteinase. Infection andImmunity 56, 1942–1949.

Romanin, D., Serradell, M.a, Gonzalez Maciel, D., Lausada, N., Garrote, G.L., Rumbo, M.n,2010. Down-regulation of intestinal epithelial innate response by probiotic yeastsisolated from kefir. International Journal of Food Microbiology 140, 102–108.

Satish Kumar, R., Kanmani, P., Yuvaraj, N., Paari, K.A., Pattukumar, V., Arul, V., 2011.Lactobacillus plantarum AS1 binds to cultured human intestinal cell line HT-29and inhibits cell attachment by enterovirulent bacterium Vibrio parahaemolyticus.Letters in Applied Microbiology 53, 481–487.

Sazawal, S., Hiremath, G., Dhingra, U., Malik, P., Deb, S., Black, R.E., 2006. Efficacy ofprobiotics in prevention of acute diarrhoea: a meta-analysis of masked,randomised, placebo-controlled trials. The Lancet Infectious Diseases 6, 374–382.

Scherwitz, C., 1982. Ultrastructure of human cutaneous candidosis. The Journal of In-vestigative Dermatology 78, 200–205.

Schierack, P., Nordhoff, M., Pollmann, M., Weyrauch, K., Amasheh, S., Lodemann, U.,Jores, J., Tachu, B., Kleta, S., Blikslager, A., Tedin, K., Wieler, L., 2006. Characteriza-tion of a porcine intestinal epithelial cell line for in vitro studies of microbial path-ogenesis in swine. Histochemistry and Cell Biology 125, 293–305.

Silva, T., Reto, M., Sol, M., Peito, A., Peres, C.M., Peres, C., Malcata, F.X., 2011. Character-ization of yeasts from Portuguese brined olives, with a focus on their potentiallyprobiotic behavior. LWT- Food Science and Technology 44, 1349–1354.

Surawicz, C.M., 2003. Probiotics, antibiotic-associated diarrhoea and Clostridiumdifficile diarrhoea in humans. Best Practice & Research. Clinical Gastroenterology17, 775–783.

Tiago, F.C.P., Martins, F.S., Rosa, C.A., Nardi, R.M.D., Cara, D.C., Nicoli, J.R., 2009. Physiolog-ical characterization of non-Saccharomyces yeasts from agro-industrial and environ-mental origins with possible probiotic function. World Journal of Microbiology andBiotechnology 25, 657–666.

Trotta, F., Caldini, G., Dominici, L., Federici, E., Tofalo, R., Schirone, M., Corsetti, A., Suzzi,G., Cenci, G., 2012. Food borne yeasts as DNA-bioprotective agents against modelgenotoxins. International Journal of Food Microbiology 153, 275–280.

Vieira-Dalode, G., Jespersen, L., Hounhouigan, J., Moller, P.L., Nago, C.M., Jakobsen, M.,2007. Lactic acid bacteria and yeasts associated with gowe production from sor-ghum in Benin. Journal of Applied Microbiology 103, 342–349.

Zanello, G., Meurens, F., Berri, M., Chevaleyre, C., Melo, S., Auclair, E., Salmon, H., 2011.Saccharomyces cerevisiae decreases inflammatory responses induced by F4+ en-terotoxigenic Escherichia coli in porcine intestinal epithelial cells. Veterinary Im-munology and Immunopathology 141, 133–138.