lable at ScienceDirect

Food Microbiology 28 (2011) 1483e1491

Contents lists avai

Food Microbiology

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

Analysis and direct quantification of Saccharomyces cerevisiae and Hanseniasporaguilliermondii populations during alcoholic fermentation by fluorescence in situhybridization, flow cytometry and quantitative PCR

Imma Andorra a,b, Margarida Monteiro a, Braulio Esteve-Zarzoso b, Helena Albergaria a,*, Albert Mas b

a LNEG, Unidade Bioenergia, Estrada do Paço do Lumiar 22, 1649-038 Lisboa, PortugalbUniversidad Rovira & Virgili, Dept Bioquim & Biotecnol, Fac Enologia, Tarragona 43007, Spain

a r t i c l e i n f o

Article history:Received 30 November 2010Received in revised form29 July 2011Accepted 7 August 2011Available online 12 August 2011

Keywords:WineCulture-independent methodsRNA stabilityYeast population dynamics

* Corresponding author. Tel.: þ351 210924721; fax:E-mail address: helena.albergaria@ineti.pt (H. Alb

0740-0020/$ e see front matter � 2011 Elsevier Ltd.doi:10.1016/j.fm.2011.08.009

a b s t r a c t

Traditionally, it was assumed that non-Saccharomyces (NS) yeasts could only survive in the early stages ofalcoholic fermentations. However, recent studies applying culture-independent methods have shownthat NS populations persist throughout the fermentation process. The aim of the present work was toanalyze and quantify Saccharomyces cerevisiae (Sc) and Hanseniaspora guilliermondii (Hg) populationsduring alcoholic fermentations by plating and culture-independent methods, such as fluorescence in situhybridization (FISH) and quantitative PCR (QPCR). Species-specific FISH probes labeled with fluorescein(FITC) were used to directly hybridize Sc and Hg cells from single and mixed cultures that wereenumerated by epifluorescence microscopy and flow cytometry. Static and agitated fermentations wereperformed in synthetic grape juice and cell density as well as sugar consumption and ethanol productionwere determined throughout fermentations. Cell density values obtained by FISH and QPCR revealed thepresence of high populations (107e108 cells/ml) of Sc and Hg throughout fermentations. Plate counts ofboth species did not show significant differences with culture-independent results in pure cultures.However, during mixed fermentations Hg lost its culturability after 4e6 days, while Sc remained cul-turable (about 108 cells/ml) throughout the entire fermentation (up to 10 days). The rRNA content of cellsduring mixed fermentations was also analyzed by flow cytometry in combination with FISH probes. Thefluorescence intensity conferred by the species-specific FISH probes was considerably lower for Hg thanfor Sc. Moreover, the rRNA content of Hg cells, conversely to Sc cells, remained almost unchanged afterboiling, which showed that rRNA stability is species-dependent.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

The transformation of grape must into wine involves the coex-istence and succession of different yeast species. The microorgan-isms present on the berry surfaces are composed mainly ofnon-Saccharomyces (NS) yeasts which predominate during theearly stages of the alcoholic fermentation. These are soon overtakenby the growth of Saccharomyces cerevisiae (Sc) which dominates themid to final stages of the fermentation (Fleet and Heard,1993; Fleet,2003). This typical growth pattern has long been accepted andestablishedmainly by platingmethods. Indeed, more recent studieshave questioned this pattern after using molecular methods thatreveal the persistence of NS populations throughout the fermen-tation process (Fernández et al., 1999; Cocolin et al., 2000; Andorràet al., 2010; Zott et al., 2010). The causes of the early displacement

þ351 217163636.ergaria).

All rights reserved.

of NS wine yeasts are still controversial. Previously, it was thoughtthat this was mainly due to the lower tolerance of NS speciestoward the increasingly adverse conditions (low pH values, highlevels of ethanol and organic acids, nutrient depletion, etc.)established in the medium as the fermentation progresses (Fleetand Heard, 1993). More recently, the dominance of Sc has beenattributed to other factors such as growth arrest mediated bycell-to-cell contact mechanisms (Nissen et al., 2003) and thesecretion of toxic compounds (Pérez-Nevado et al., 2006;Albergaria et al., 2010). However, most of these studies have beencarried out by using classic plating methods which are laborious,time-consuming, somewhat unreliable (Giraffa, 2004) and whichonly detect culturable populations.

Molecular techniques have been developed and used to controlmicrobial growth and to characterize the microflora of differentprocesses and environments. These methods are generally faster,more specific, more sensitive, and more accurate and allow theprecise study of microbial populations and their diversity (Justé

I. Andorra et al. / Food Microbiology 28 (2011) 1483e14911484

et al., 2008). Molecular techniques can be used to identify or geno-type microorganisms previously grown on a culture media(commonly known as culture-dependent techniques), or they canbe applied directly to a sample (known as a culture-independenttechniques) (Rantsiou et al., 2005). The application of culture-independent methods gives a better understanding of the truemicrobial diversity, avoiding the biases that growth and isolation byenrichment platting might introduce. Furthermore, culture-dependent techniques can underestimate the size and diversity ofa given population because they do not account for non-culturablepopulations, such as sublethally injured and/or viable butnon-culturable (VBNC) cells, which may fail to grow on plates andare common in wine (Millet and Lonvaud-Funel, 2000). Under-estimating VBNC and/or sublethally injured populations can beextremely important in some situations (e.g. food contaminantsand/or pathogens) since these cells are still metabolically active(Oliver, 2005; Mills et al., 2008) and can recover growth undercertain conditions. The difference between the total population ofa microorganism and its culturable subpopulation within a givenenvironment corresponds to its VBNC subpopulation and indicatesthe stress effect induced by that environment on the physiologicalstate of the cells. For this reason, to better analyze a microbialpopulation both molecular and plating methods should be used.

Several culture-independent techniques have been developedand used for detecting and quantifying wine yeast species (Cocolinet al., 2000; Mills et al., 2002; Phister and Mills, 2003; Hierro et al.,2006, 2007; Andorrà et al., 2008, 2010). One of the most promisingof these methods, due to its simplicity and rapidity, is the fluores-cence in situ hybridization (FISH). This technique combines thesimplicity of microscopic observation with the specificity of DNA/RNA analysis. Furthermore, most of the molecular techniques donot provide information regarding the morphology, number andspatial distribution of cells within a given environment. In theory,the FISH technique can detect single cells but in practice thedetection limit is often 104 cells/ml since enumeration is usuallycarried out by hemocytometry. This limitation can be overcome byconcentrating the samples prior to hybridization and counting(Blasco et al., 2003). Another limitation of using hemocytometry isthat it does not have sufficient automation for high throughputsample analysis (Amann et al., 2001). This can be solved by usingflow cytometry (FC) in combination with FISH for the selectiveenumeration of mixed microbial populations, which allows a highresolution and highly automated analysis (Amann et al., 1990). Themain advantage of this technique is its sensitivity since it can detectone cell in a million. FC in combination with FISH probes has beenused to identify and analyze mixed microbial populations (Amannet al., 1990; Wallner et al., 1993; Rigottier-Gois et al., 2003).

Xufre et al. (2006) developed 26S rRNA gene probes for identi-fication of numerous wine-related yeast species, including Sc,Candida stellata, Hanseniaspora uvarum, Hanseniaspora guillier-mondii (Hg), Kluyveromyces thermotolerans, Kluyveromyces marx-ianus, Torulaspora delbrueckii, Pichia membranifaciens and Pichiaanomala. Stender et al. (2001) detected the slow growing yeastDekkera bruxellensis. These authors, however, used cultivation andisolation steps prior to applying the FISH technique and thus thoseresults did not account the overall cell population but rather justthe culturable cells.

The aim of the present work was to use plating and culture-independent methods based on RNA (FISH) and DNA (QPCR) toanalyze Sc and Hg populations. These methodologies were used forthe direct quantification and identification of Sc and Hg pop-ulations in single cultures and mixed fermentations performed onsynthetic media and on a simulated wine fermentation. FC incombination with FISH probes was used to quantify the relativefluorescence intensity of hybridized cells of Sc and Hg during the

course of mixed fermentations and also to determine the stabilityof the rRNA content of the cells.

2. Materials and methods

2.1. Yeasts strains, inoculum cultures and growth media

The yeasts strains used were Sc CCMI 885 (Culture Collection ofIndustrial Microorganisms, LNEG, Lisbon, Portugal) and Hg NCYC2380 (National Collection of Yeast Cultures, Norwich, UK). Bothstrains were first isolated from Portuguese wines, Sc from Alentejoand Hg from Douro wine producing regions. Yeasts were main-tained on YMPD-agar slants (1% dextrose, 0.5% peptone, 0.3% maltextract, 0.3% yeast extract, 2% agar, w/v) and stored at 4 �C.

Inocula of Hg and Sc were prepared by transferring the biomassof one YMPD-agar slant (pre-grown for 48 h at 30 �C) into 50 ml ofYMPD medium in 100 ml flasks which were incubated for 16 h at30 �C and 150 rpm.

Single cultures of Sc and Hg were performed on YMPD mediaand mixed fermentations were performed on synthetic grape juice(SGJ) (D-glucose, 110 g/L; D-fructose, 110 g/L; L-(1)-tartaric acid,6.0 g/L; L-(2)-malic acid, 3.0 g/L; citric acid, 0.5 g/L; YNB (yeastnitrogen base), 1.7 g/L; CAA (Casamino Acids), 2.0 g/L; CaCl2, 0.2 g/L;arginine-HCl, 0.8 g/L; L-(2)-proline, 1.0 g/L; L-(2)-tryptophan, 0.1g/L; pH 3.5), prepared as described by Pérez-Nevado et al. (2006).

2.2. Single cultures on YMPD medium

Single cultures of Sc and Hg were carried out in 500 ml flasksfilled with 250ml of YMPDmedium. Each flaskwas inoculatedwith1�104 cells/ml of the respective yeast strain and incubated withoutagitation at a constant temperature of 20 �C. Both fermentationswere performed in duplicate and monitored by plate counting andby direct application of the FISH technique, using species-specificFITC-labeled probes and DAPI staining. Hybridized cells werecounted in a Neubauer chamber using an epifluorescence micro-scope (Olympus BX-60, Tokyo, Japan).

2.3. Alcoholic fermentations

Two mixed fermentations (shaken and static) were performedin duplicate in 500 ml flasks filled with 250 ml of SGJ and inocu-lated with 1 �105 cells/ml of each yeast strain (Hg and Sc). Shakenfermentations were conducted under constant agitation of 100 rpmin an orbital shaker (Unitron, Infors, Switzerland). All the experi-ments were conducted at a controlled temperature of 20 �C. Thefermentations were monitored by taking daily samples to quantifycellular density and determine sugar consumption and ethanolproduction. The yeast population was determined by the classicplating method and also by FISH and QPCR.

To simulate the second part of the alcoholic fermentation,a commercial red wine was diluted with sterile distilled water untilthe ethanol concentration reached 30 g/l. The medium was sup-plemented with glucose and fructose to attain a final concentrationof 55 g/l for each sugar and 3 g/l of yeast extract and thus simulatethe average sugar concentration of a mid-point of a wine fermen-tation. This medium was then inoculated with 106 cells/ml of eachstrain (Sc and Hg) with inocula previously grown (for 16 h at 30 �C)on YMPD broth. The fermentations were conducted in duplicatewith constant agitation (100 rpm) at a controlled temperature of20 �C. Sugar consumption was monitored daily by both densitymeasures and by HPLC. Samples were taken every 24 h until theend of fermentation (sugars below 2 g/l). Yeasts were monitoredusing plating counts, and by direct application of the FISH (withspecies-specific probes and DAPI) and QPCR techniques.

I. Andorra et al. / Food Microbiology 28 (2011) 1483e1491 1485

2.4. Fluorescence in situ hybridization (FISH)

2.4.1. Oligonucleotide FISH probesOligonucleotides ranging from 15 to 20 nucleotides and targeted

at the D1/D2 domain of 26S rRNA of the yeasts species Sc and Hgwere synthesized and labeled with the fluorochrome FluoresceinIsoThioCyanate (FITC) at the 50-end. The species-specific FISHprobes used in this study were those previously designed andvalidated by Xufre et al. (2006) with the following sequences for Hgand Sc, respectively: 50-CAATCCCAGCTAGCAGTAT-30 and 50-TGACTTACGTCGCAGTCC-30. Furthermore, a universal eukaryoteFISH probe (EUK 516 50-ACCAGACTTGCCCTCC-30) was also used asa positive permeabilization control.

2.4.2. FISH procedureThe FISH procedure used the protocol described in Xufre et al.

(2006) but it was adapted to be used directly in the fermentationsamples. The detailed protocol for direct analysis was as follows:One ml of the fermentation samples was centrifuged for 5 min at5000 g, the cells were washed once with 1 � phosphate-bufferedsaline (PBS e 130 mM sodium chloride, 10 mM sodium phosphatebuffer, pH 7.2) and incubated with 4% (v/v) of paraformaldehyde forat least 3 h at 4 �C under strong agitation. Fixed cells were there-after centrifuged for 2 min at 10,000 g and resuspended in 800 ml ofa solution containing 400 ml of 1 � PBS buffer and 400 ml ethanol(98%) and kept at �20 �C until required. Approximately 106 of fixedcells were hybridized in 50 ml of hybridization buffer (0.9 M sodiumchloride, 0.01% w/v sodium dodecyl sulfate, 20 mM TriseHCl and0.025% v/v formamide) with 10 ng of FITC-labeled probe permicroliter and incubated at 46 �C for at least 3 h. After incubation,cells were pelleted by centrifugation, and the supernatant wasdiscarded. The pellet was resuspended in 100 ml of washing buffer(25 mM Tris/HCl and 0.5 M NaCl) and incubated for 30 min at 48 �C.Before enumeration, the suspension of washed cells was pelletedagain, resuspended in 100 ml of 1� PBS, stainedwith DAPI (1 mg/ml)(4,6-diamidino-2-phenylindole), a fluorescent dye that stains thedouble stranded-DNA, and incubated for 1 min at room tempera-ture. Approximately 5 ml of this cell suspensionwas mixed with 5 mlof Vecta Shield (an anti-fading agent, Vector Laboratories, USA),spotted onto a Neubauer chamber and enumerated by epifluor-escence microscopy (Olympus BX-60, Tokyo, Japan). Total cellswere enumerated using an NU-MWB filter (wavelength for DAPI)and Hg and Sc hybridized cells were enumerated using an U-MWBfilter (wavelength for FITC).

2.5. DNA extraction and QPCR

DNA was extracted according to Hierro et al. (2006). QPCR wasperformed in an Applied Biosystems 7300 Fast Real-Time PCRSystem (Applied Biosystems, USA). Power SyberGreen master mixwas used according to the manufacturer’s instructions (AppliedBiosystems, USA). An ABI Prism 96-well optical plate (AppliedBiosystems, USA) was used for the reaction. The instrument auto-matically determined the Ct. Yeast was quantified using the primersYESTF/YESTR (Hierro et al., 2006) for total yeasts, CESPF/SCERR forSaccharomyces and CESPF/HUVR for Hanseniaspora (Hierro et al.,2007). Standard curves were made for each type of microor-ganism using triplicate determinations.

2.6. Flow cytometry

Cells from single and shaken mixed fermentations of Hg and Scwere hybridized with the respective FISH probes and analyzed byFC. The fluorescence of the hybridized cells was quantified with anFAC-Scan FC (Becton Dickinson Instruments, Erembodegem,

Belgium). The argon ion laser was tuned to an output of 15 mW at488 nm. Forward-angle light scatter (FSC) and side-angle lightscatter were used and fluorescence (FL1) was detected with a 530(�30 nm) nm band pass filter. FITC-probes were measured withdeionized filtered water (0.2 mm Millipore membranes) as thesheath fluid. The FC automatically stabilized the optical alignmentand standardized the fluorescence intensities of the hybridizedcells.

2.6.1. Data acquisition and processingThe parameters FSC, SSC and FL1 were recorded and 20,000

events for each measure were stored in list modes. Samples wererun such that the event rate was below 800e1000 events s�1 andthe analyses were performed at a low flow rate (12 l min�1). FSCand SSC were plotted in a logarithmic scale and data analyzed withCellQuest software (BD Bioscience) and WinMDI software (version2.8). Average FSC and SSC signals were obtained from the samesoftware. Rainbow calibration particles (3.0e3.4 mm) were used forroutine calibration of the flow cytometer.

Probe-conferred fluorescence was determined as the(geometric) mean of the fluorescence values emitted by singlehybridized cells (of Sc and Hg) that were measured in the FL1detector and recorded on a gate that was first defined in anFSC-versus-SSC density plot. This mean value was then divided bythemean value of the fluorescence emitted by non-hybridized cells.Samples from mixed fermentation were first analyzed withouthybridization in an FSC-versus-SSC density plot. Due to theconsiderably different size of Sc and Hg cells it was possible toobtain separated gates containing the entire cell population of eachyeast species. Mixed samples were hybridized separately with eachspecies-specific probe and analyzed with and without the probe.FISH experiments were performed in triplicate for each sample inindependent hybridization assays (6 replicates for mixed samples),and fluorescence values were measured on the FC at various timeintervals (30 min, 1 h, 2 h and 4 h) after applying the FISH proce-dure. No significant loss of fluorescence was detected for at least4 h. The stability of the cell rRNA was analyzed by FC, whichmeasured the fluorescence intensity emitted by hybridized cellsafter boiling (10 min at 100 �C) and also after RNase treatment(10 ml of RNase solution (0.2 mg/ml) was added to 100 ml of sample).

2.7. Analytical methods

Glucose and fructose consumption and ethanol productionweredetermined by HPLC (Merck, Darmstadt, Germany) using a Sugar-Pak� column (Waters, Milford, USA) and a refractive indexdetector. The samples were eluted at 90 �C using a degassedaqueous mobile phase containing 50mg/l Ca-EDTA, at a flow rate of0.5 ml/min. All samples were analyzed in triplicate.

2.8. Plating counts

The Hg and Sc populations during mixed fermentations werequantified by the classical plating method. Samples were takenaseptically throughout the fermentations and inoculated ontoYMPD-agar plates after appropriate dilution in sterile water. In themixed fermentations, the CFU count of Hg was carried out onYMPD-agar plates containing 0.01% of cycloheximide, a selectivemedium in which Sc is unable to grow. The CFU count of Sc wascalculated as the difference between the total CFU counts obtainedon YMPD-agar plates (without cycloheximide) and the CFU countsof Hg. All plates were incubated at 30 �C for 2e4 days and the CFUswere enumerated after no increase of colonies was observed on theplates.

I. Andorra et al. / Food Microbiology 28 (2011) 1483e14911486

2.9. Statistical analysis

The results were statistically analyzed using the SPSS 17.0 soft-ware for Windows (SPSS Inc., Chicago, IL). Analysis of variance wascarried out by an ANOVA Tukey test to determine significantdifferences between the samples. The statistical level of signifi-cance was set at a P � value < 0.05.

3. Results

3.1. Validation of the direct application of FISH

In order to validate the direct application of FISH single culturesof Sc and Hg were performed on YMPD medium. Samples fromthose cultures were directly hybridized with the respective FISHprobes, stained with DAPI and counted in a Neubauer chamber byepifluorescence microscopy. The cell population of those fermen-tations was also followed by plating. The FISH technique did notshow significant differences with the plating counts for both yeaststhroughout cultivation (Fig. 1). Furthermore, results showed thatunder single cultivation both strains were able to grow on YMPD-agar plates long after sugar exhaustion, which occurred within24 h, although a slight decreasewas observed in the cellular densityof Hg after 72 h of cultivation (by both methods). This means thatcells kept their culturability even after long periods of starvation.

The FISH procedure was applied directly to samples frommixedfermentations and microscopic observation of the hybridized cells(Fig. 2) confirmed the species-specificity of the FISH probes. Inaddition, hybridized cells of Hg (Fig. 2A) and Sc (Fig. 2C) comparedwell with the total cell counts (Fig. 2B and D) measured by DAPIstaining, thus confirming and validating the effectiveness of theFISH procedure and the reproducibility of enumeration byhemocytometry.

Fig. 1. Yeast population evolution during single Sc (A) and Hg (B) cultivations carriedout on YMPD medium, as determined by FISH and plating ( , Sc-FISH; , Sc-CFU; , Hg-FISH; , Hg-CFU; , total cells-DAPI). (means � SD).

3.2. Monitoring of mixed fermentations on SGJ

Sugar consumption and ethanol production during shaken andstatic fermentations are represented in Fig. 3A and A0 respectively.The time required to consume the sugars present on the SGJ was 6and 10 days for shaken and static fermentation respectively. In bothcases, the consumption rates of glucose were higher than those offructose, and residual sugars present in the final wine were lessthan 2 g/l of fructose. Both fermentations produced similaramounts of ethanol with final concentrations of about 100 g/l.

Fig. 3 represents the cell density evolution during shaken andstatic fermentations for the total yeast (B, B0), the Sc (C, C0) and theHg (D, D0) populations, as determined by plating, FISH and QPCR.The results showed that in both fermentations yeasts raised theirinitial cell populations from 105 cells/ml to 107e108 cells/ml (Fig. 3Band B0) within 24e48 h. During the stationary growth phase the celldensity profiles of Hg (Fig. 3D and D0) significantly diverged whencomparing QPCR and FISH values with those of plating. Quantifi-cation of Hg population by both culture-independent techniques(FISH and QPCR) revealed the existence of about 1�108 cells/ml upto 100 h in the shaken fermentation, although a 10-fold decreasewas observed after that time by QPCR. In contrast, Sc kepta constant cellular density of about 6 � 107 cells/ml. Whenfermentations results were statistically analyzed, no statisticallysignificant differences were found between Sc cell density values(Fig. 3C and C0) obtained by QPCR, FISH and Plate methods duringshaken and static mixed fermentations. However, for Hg celldensity (D and D0) statistically significant differences were foundbetween Plate counts and either QPCR or FISH after 24 h in shaken(D) and 72 h in static (D0) fermentations, respectively. Moreover,after 96 h there was also statistically significant differencesbetween QPCR and FISH values for Hg cell density both in shakenand static fermentations (Fig. 3D and D0). Independently of whetherthe method used was culture-dependent or not, no major declinewas observed in the cell population of Sc for at least 200 h in allfermentations, whereas Hg completely lost its ability to formcolonies on agar plates after 4 and 6 days in shaken and staticfermentations, respectively.

3.3. Monitoring of a simulated middle wine fermentation

In the simulated wine fermentation yeasts showed a similarfermentation pattern to that of SGJ fermentations in that theyconsumed glucose faster than fructose (Fig. 4A). The initial ethanolcontent of the media was 30 g/l and after 96 h of fermentation theresidual sugars were lower than 2 g/l and ethanol reached a finalconcentration of 80 g/l. The total yeast and Sc population increasedto 108 cells/ml in the first 24 h, and maintained this cell densitythroughout the process when measured by both culture-independent methods (FISH and QPCR) (Fig. 4B and C). Mean-while, Hg began to lose its culturability on YMPD-agar plates after24 h and did not grow at all on plates after 96 h. When analyzedusing culture-independent methods (FISH and QPCR), the Hgpopulation did not increase and was constant throughout thefermentation at slightly over 106 cells/ml. No change was observedfor QPCR, which always enumerated a higher population than FISH.

3.4. FC in combination with FISH probes

The FISH procedure was used in combinationwith FC in order toevaluate changes in the rRNA content of cells during the fermen-tations and to avoid the subjectivity of operator microscopecounting. Furthermore, the RNA stability after boiling and/or RNasecell treatments was also analyzed. Cells from pure and mixedcultures were hybridized with the respective FISH probes and then

Fig. 2. Microscope visualization of fluorescence signals emitted by Sc and Hg cells of mixed fermentation samples hybridized with FITC-labeled FISH probes and stained with DAPI.Images of cells hybridized with Hg-FISH probe and captured on the filter for FITC (A) and on the filter for DAPI (B). Images of cells hybridized with Sc-FISH probe captured on thefilter for FITC (C) and on the filter for DAPI (D).

I. Andorra et al. / Food Microbiology 28 (2011) 1483e1491 1487

analyzed by FC, which acquired data for cell size (FSC), internalheterogeneity (SSC) and fluorescence (FL1). Initially, pure cellsuspensions of Hg and Sc were analyzed by FC for non-hybridizedcells and cells hybridized with the respective species-specificFISH probes (data not shown). The results showed that becauseSc cells are larger than Hg cells it is possible to use an SSC-versus-FSC density plot to design different gates for each yeast species. Theautofluorescence of the Hg and Sc yeasts was 1.21 for Hg and 3.5 forSc, as determined by the relative fluorescence intensity (i.e. the FL1signals in a LOG scale) emitted by non-hybridized cells from eachspecies. The specific fluorescence intensity conferred by each FISHprobe to the Hg and Sc yeasts cells was determined by dividing therelative fluorescence intensity of hybridized cells by their respec-tive autofluorescence values.

Similar quantification and analysis was carried out for samplesof the mixed fermentation (at 24, 48, 96, 144 and 192 h). Thisinvolved hybridizing the cells with either Hg or Sc species-specificFISH probes (in independent assays for each probe) and analyzingthe samples using FC (Fig. 5). The results showed that the Sc pop-ulation was much more heterogeneous along the fermentationthan the Hg population as can be seen in Fig. 5A and D, where twosubpopulations of Sc are present within the same sample,depending probably on whether cells are budding or not. Thenumber of Hg and Sc cells within a given sample and the respectivefluorescence intensity conferred by each FISH probe was deter-mined throughout the mixed fermentation (Table 1). The resultsshowed that the relative fluorescence intensity emitted by Hg-hybridized cells was always lower than that of Sc cells. Moreover,Sc cells exhibited 4e6 times more fluorescence during the expo-nential phase (24e48 h) than the stationary phase (96e196 h), incontrast to Hg cells, whose relative fluorescence intensity onlyslightly decreased without significant differences (Table 1). Thestability of these yeasts’ rRNA was analyzed by measuring the

fluorescence intensity emitted by dead cells that had been boiled(10 min at 100 �C) and/or treated with RNase and thereafterhybridized with the respective FISH probes. The results showedthat for both yeast species RNA disappeared immediately afterRNase treatment but not after boiling. In the case of Sc, after 24 hthere was no significant difference in the RNA content of boiledcells and RNase treated cells, whereas the RNA levels of Hg cellsremained fairly constant after boiling, being detected for at least96 h (Table 1).

4. Discussion

In the present study, the evolution of Sc and Hg populationsduring single and mixed fermentations was analyzed by using FISHin combination with either epifluorescence microscopy or FC andQPCR. The cell density profiles obtained for both yeast species withthese molecular methods were compared with those obtained byclassical plating on a YMPD-agar medium. The QPCR method usedwas previously developed to monitor the yeast evolution duringalcoholic fermentations (Hierro et al., 2007; Andorrà et al., 2008,2010). However, QPCR targeted at DNA quantifies also dead yeastsbecause of the DNA’s stability (Hierro et al., 2006). Since FISHtechnique targets rRNA, it is generally assumed that it shouldquantify only viable populations (Giraffa and Carminati, 2008).Nevertheless, in the present work a good correlation was found forboth methods. The FISH technique was first used by Xufre et al.(2006) to identify different yeast colonies obtained at variousstages of wine fermentations. In the present work we used thespecies-specific probes that were designed and validated in thatstudy to identify Sc and Hg. The FISH techniquewas optimized to bedirectly applied to samples during alcoholic fermentations, thuspreventing previous plate cultivation. The results obtainedconfirmed the species-specificity of both probes since no cross-

Fig. 3. Sugar consumption and ethanol production in shaken (A) and static fermentation (A0) ( , ethanol; , fructose; , glucose). Evolution of the total yeasts (B, B0),S. cerevisiae (C, C0) and H. guilliermondii (D, D0) species in shaken and static mixed fermentation analyzed by FISH ( ), QPCR ( ) and plating ( ). All values are means of atleast 2 replicates. Statistical analysis of data showed global significance (P > 0.05) between values obtained by QPCR, FISH and Plating for Sc cell density (C,C0) but not for Hg celldensity (D,D0). a: after 24 h all Plating values are statistically different (P < 0.05) from those of QPCR and FISH. b: after 72 h all Plating count values are statistically different(P < 0.05) from those of QPCR and FISH. c: after 96 h all QPCR values are statistically different (P < 0.05) from those of FISH.

I. Andorra et al. / Food Microbiology 28 (2011) 1483e14911488

species hybridization was detected. In addition, hybridized cells ofHg and Sc compared well with total cell counts obtained by DAPIstaining, thus demonstrating the effectiveness of the FISH proce-dure and microscopic enumeration. Therefore FISH proved to bea consistent and reproducible method for directly quantifying Scand Hg populations during mixed fermentations and singlecultures.

The Sc and Hg cell density profiles determined by plating weretotally consistent with those found by Pérez-Nevado et al. (2006)and Albergaria et al. (2010) for the same strains in mixed fermen-tations performed on SGJ. Indeed, Sc maintained its culturability forat least 10 days in both mixed fermentations, whereas Hg began tolose its culturability after the first and third day and was not able toform colonies after 4 and 6 days in the shaken and static fermen-tations, respectively. However, when measured by the two culture-independent methods, QPCR (DNA) and FISH (rRNA), the Hg pop-ulation remained at high levels (up to 107 cells/ml) during theentire fermentation process. On the other hand, this species pop-ulation declined one order of magnitude when enumerated byQPCR during the second half of the fermentation, something which

did not occur with the FISH technique. In the case of Sc, enumer-ation by the different methods always yielded similar results, evenin the simulated wine fermentation that started with a certainconcentration of ethanol (Fig. 4).

Our results are similar to those found by Cocolin and Mills(2003), which studied the effect of sulfur dioxide in the yeastpopulation during wine fermentations by applying PCR-DGGEmethods targeted to DNA and rRNA. They also found that the NSpopulation could maintain high levels of DNA or RNA after almost20 days of being unculturable, which coincides with our results inwhich Hg lost its ability to form colonies despite maintaining itsDNA/RNA for almost 240 h. Those authors concluded that NS yeastsmay persist in a non-culturable state within standard winefermentations and that plating alone is insufficient for monitoringthe various yeast populations. This is due to the existence of VBNCpopulations, injured populations and/or dead populations in whichthe cellular DNA or RNA is protected from degradation. For thisreason, Cocolin and Mills (2003) stated that future work is requiredin order to discriminate between these different populationswithin wine.

Fig. 4. Sugar consumption and ethanol production in simulated wine fermentation (A) ( , ethanol; , fructose; , glucose). Evolution of the total yeasts (B), S. cerevisiae (C)and H. guilliermondii (D) species in simulated wine fermentation when analyzed by FISH ( ), QPCR ( ) and plating ( ). Asteriscs indicate statistically significant difererences(P < 0.05) compared to QPCR values.

I. Andorra et al. / Food Microbiology 28 (2011) 1483e1491 1489

When added to a wine medium simulating the mid-point ofa wine fermentation, Hg was unable to grow and began to lose itsculturability after 48 h of fermentation. The levels of ethanol whenHg stopped growing in plates were completely different in theseveral fermentations (single, mixed and simulated wine

Fig. 5. Density plots (A, B, D, E) and histograms (C, F) obtained by FC analysis (20,000 eventthe species-specific FISH probes for Hg (A, B, C) and for Sc (D, E, F). Histograms represent theemitted by hybridized cells of Hg (C) and Sc (F) within the respective gates, as defined inintensity of the entire cell population within the respective gate.

fermentation), which means that this factor can be ruled out as themain cause of this species’ loss of culturability. Moreover, undersingle cultivation Hg was able to grow on YMPD-agar plates longafter sugar exhaustion, which occurred within 24 h. This meansthat Hg cells kept their culturability even after long periods of

s recorded) of a 144 h mixed fermentation sample in which cells were hybridized withdistribution among the entire population analyzed of the relative fluorescence intensitythe SSC-versus-FSC density plots. Mn, is the mean value of the relative fluorescence

Table 1Flow cytometry analysis of hybridized cells of Hg and Sc during the shakenmixed fermentation and also of cells that were hybridized after boiling and after RNase treatment atthe end of fermentation. All values represent the mean value (20,000 events recorded per analysis) of at least three independent experiments performed for each FISH probehybridisation. Standard deviations were in all cases lower than 10%.

Time-sampling(h)

Number of Hgcells

Number of Sccells

Fluorescence of Hgcells

Fluorescence of Sccells

24 17658 � 774 774 � 72 14.66 � 0.97a 132.41 � 8.20b

48 15460 � 620 2914 � 495 15.94 � 1.75 121.78 � 8.7196 12824 � 228 6690 � 258 14.49 � 1.09 20.74 � 1.88**144 13826 � 338 5848 � 459 12.61 � 0.54 27.55 � 2.33**196 11703 � 456 7725 � 458 9.49 � 0.52* 17.05 � 1.16**Sample 196 h Boiled cells for

10 min at 100 �C0 h 10.23 � 0.26*’ 13.08 � 0.05**’24 h 8.33 � 0.25*’ 1.25 � 0.0196 h 10.02 � 0.01*’ nd

RNase treated cells 0 h 1.14 � 0.10a’ 1.91 � 43b’

nd: not determined; Asterisks (*,*’,**,**’) indicate means statistically different (P < 0.05) from the respective control value (a, a’, b, b’).

I. Andorra et al. / Food Microbiology 28 (2011) 1483e14911490

starvation, that is, at least for 136 h. Previous research (Pérez-Nevado et al., 2006; Albergaria, 2007) carried out with the sameyeast strain showed that under single fermentation Hg was able tomaintain its culturability for long time (up to 15e20 days), whereasunder mixed fermentations Hg cells began to lose their ability togrow on YMPD-agar plates after 1e3 days of fermentationdepending on the initial Sc/Hg ratio. However, in the present studyboth the FISH and QPCRmolecular methods found high populationsof Hg (up to 108 cells/ml) throughout the mixed fermentations. Puttogether, all these findings raise the question as to how many ofthese Hg cells are metabolically active and how these in turninfluence the final wines. In addition, it remains to be determinedwhether these populations correspond to VBNC, injured and/ordead cells.

The FISH technique combined with FC can be used to analyzethe different physiological states by determining the levels of rRNAduring the different stages of winemaking. The FC quantification ofthe number of Sc and Hg cells throughout the mixed fermentationcompared well with results from the FISH microscopic enumera-tion. The fluorescence intensity of Sc cells was significantly higherduring the exponential phase than in the stationary phase, whereasonly a slight decrease of fluorescence was observed for Hg cells.

These results have to be related to the kinetics and main physi-ological changes of the microbial population during alcoholicfermentation. Yeasts during winemaking are found in differentgrowth stages (i.e. the lag, exponential and stationary phases) andHierro et al. (2006) have found that the rRNA copy number is up tofour times lower in the stationary phase samples than in the lag andexponential phase (corresponding to 2 Ctwhen estimated byQPCR).Thus, the decrease in thefluorescence of hybridized Sc cells after theexponential phase indicates lower protein synthesis during this lastphase. Since the number of intact ribosomes approximately reflectsthe rates of protein synthesis, rRNA has been proposed as a morerepresentative target for assessing cell viability and as a goodgeneral marker of metabolic activity (Ward et al., 1992; Vandervlietet al., 1994; Felske et al., 2000). However, it should be taken inaccount that the fluorescence intensity emitted by the hybridizedcells is an indirect measure of the rRNA content, because the effi-ciency of the FISH procedure depends on several factors, in partic-ular accessibility of the probe to the rRNA target-sequence (Inácioet al., 2003). Consequently, these values are not indicative of abso-lute quantities of ribosomes and fluorescence intensity should not,therefore, be considered a direct measurement of protein synthesis.The fact that RNA is easily degraded by RNase enzymes is avoided inthe FISH technique because there is no need to extract and manip-ulate the RNA beforehand. The stability of the rRNA was evaluatedfor both yeasts by either boiling cells (10 min at 100 �C) or treatingcellswith RNases (10min at 65 �C). This revealed that boiledHg cellsmaintained similar amounts of rRNA for at least 96h,whereas boiled

Sc cells lost all rRNA within 24 h after boiling. RNase treatmentscompletely destroyed the rRNAcontent of cells in bothyeast species.Moreover, boiled Hg cells that had been hybridized after RNasetreatment did not show any fluorescence, which confirmed thattherewasnonon-specific hybridizationbetween the FISHprobe andother cell structures. Thus, the fluorescence emitted by the hybrid-ized cells should be attributed entirely to the hybridization of FISHprobes with the respective rRNA target-sequence for yeast species.However, it should be emphasized that the maintenance of thefluorescence signals emitted by boiled cells of Hg seems to reflecta stronger resistance of its ribosomal structure toward heat degra-dation than that of Sc cells. Thus, our results show that the degra-dation of rRNA after cell death is species-dependent. It should beemphasized that the analysis of live cell populations by culture-independent methods based on rRNA determination should takethe species-specific stability of this molecule into consideration.Apart from thework of Hierro et al. (2006), who studied the stabilityof the rRNA content of Sc cells toward heat treatment, there are nopublished data regarding the stability of the rRNA in other yeastspecies. However, in the pioneer article of Cocolin et al. (2000) therRNA of some non-Saccharomyces species was detected long afterthose populations had lost their culturability, which can only beexplainedby the stabilityof the rRNAof dead cells and/orVBNCcells.

The effect of the viable but non-culturable microorganismsduring alcoholic fermentation is not sufficiently clear. Microor-ganisms can be present in changing physiological states during thedifferent stages of winemaking. These physiological states canchange from fully alive to almost dead and can have different ratesof metabolic activity that can release metabolites and thus easilyalter the fragile status of a wine. There is a strong need for researchinto these transition states between life and dead because thiscould be really important in the later phases of thewinematuration(i.e. aging, bottling and storing) prior to consumption. The approachwe have taken in the present study has been to use differentculture-independent methodologies for estimating the viability ofthe cells. However, the concept of viability has to be framed interms of metabolic activity and this concept has to be furtherdefined and oriented according to the results of this activity. Thus,in winemaking, the metabolic activity of non-Saccharomyces yeastsshould be measured according to the production or release ofunwanted metabolites (or in other words, spoilage potential) andthe methods used should incorporate this potential so that the“metabolic activity” of non-culturable cells can be accuratelyestimated.

Acknowledgments

The present work was supported by the Bioenergy Unit of theLNEG (Portugal) and by the projects AGL2007-66417-C02-02/ALI

I. Andorra et al. / Food Microbiology 28 (2011) 1483e1491 1491

and AGL2007-65498-C02-02/ALI of the Spanish Ministry ofEducation and Science. The stage of Imma Andorrà at LNEG wassupported by a mobility bursary from the Ministry of Education(Spain). The authors would like to thank Dr Teresa Lopes da Silvafrom LNEG for her assistance with the flow cytometry work.

References

Albergaria, H. (2007). Physiological studies of non-Saccharomyces wine-relatedstrains in single and mixed cultures with S. cerevisiae. PhD Thesis, CatholicUniversity of Portugal.

Albergaria, H., Francisco, D., Gori, K., Arneborg, N., Gírio, F., 2010. Saccharomycescerevisiae CCMI 885 secretes peptides that inhibit the growth of some non-Saccharomyces wine-related strains. Appl. Microbiol. Biotechnol. 86, 965e972.

Amann, R.I., Binder, B.J., Olson, R.J., Chisholm, S.W., Devereux, R., Stahl, D.A., 1990.Combination of 16S rRNA-targeted oligonucleotide probes with flow cytometryfor analyzing mixed microbial populations. Appl. Environ. Microbiol. 56,1919e1925.

Amann, R., Fuchs, B.M., Behrens, S., 2001. The identification of microorganisms byfluorescence in situ hybridization. Curr. Opin. Biotechnol. 12, 231e236.

Andorrà, I., Landi, S., Mas, A., Guillamón, J.M., Esteve-Zarzoso, B., 2008. Effect ofoenological practices on microbial populations using culture independenttechniques. Food Microbiol. 25, 849e856.

Andorrà, I., Landi, S., Mas, A., Esteve-Zarzoso, B., Guillamón, J.M., 2010. Effect offermentation temperature on microbial population evolution using cultureindependent and dependent techniques. Food Res. Int. 43, 773e779.

Blasco, L., Ferrer, S., Pardo, I., 2003. Development of specific fluorescent oligonu-cleotide probes for in situ identification of wine lactic acid bacteria. FEMSMicrobiol. Lett. 225, 115e123.

Cocolin, L., Bisson, L.F., Mills, D.A., 2000. Direct profiling of the yeast dynamics inwine fermentations. FEMS Microbiol. Lett. 189, 81e87.

Cocolin, L., Mills, D.A., 2003. Wine yeast inhibition by sulphur dioxide: a compar-ison of culture-dependent and independent methods. Am. J. Enol. Vitic. 54,125e130.

Fleet, G.H., Heard, G.M., 1993. Yeasts: growth during fermentation. In: Fleet, G.H.(Ed.), Wine Microbiology and Biotechnology. Harwood Publishing, Switzerland,pp. 27e54.

Fleet, G.H., 2003. Yeast interactions and wine flavour. Int. J. Food Microbiol. 86,11e22.

Felske, A., de Vos, W.M., Akkermans, A.D.L., 2000. Spatial distribution of 16S rRNAlevels from uncultured acidobacteria in soil. Lett. Appl. Microbiol. 31, 118e122.

Fernández, M.T., Ubeda, J.F., Briones, A.I., 1999. Comparative study of non-Saccha-romyces microflora of musts in fermentation, by physiological and molecularmethods. FEMS Microbiol. Lett. 173, 223e229.

Giraffa, G., 2004. Studying the dynamics of microbial populations during foodfermentations. FEMS Microbiol. Rev. 28, 251e260.

Giraffa, G., Carminati, D., 2008. Molecular techniques in food fermentation: prin-ciples and applications. In: Cocolin, L., Ercolini, D. (Eds.), Molecular Techniquesin the Microbial Ecology of Fermented Food. Springer Publishing Inc, New York,pp. 1e30.

Hierro, N., Esteve-Zarzoso, B., González, A., Mas, A., Guillamón, J.M., 2006. Real-timequantitative PCR (QPCR) and reverse transcription-QPCR for detection andenumeration of total yeasts in wine. Appl. Environ. Microbiol. 72, 7148e7155.

Hierro, N., Esteve-Zarzoso, B., Mas, A., Guillamón, J.M., 2007. Monitoring ofSaccharomyces and Hanseniaspora populations during alcoholic fermentation byreal-time quantitative PCR. FEMS Yeast Res. 7, 1340e1349.

Inácio, J., Behrens, S., Fuchs, B.M., Fonseca, A., Spencer-Martins, I., Amann, R., 2003.In situ accessibility of Saccharomyces cerevisiae 26S rRNA to Cy3-labeledoligonucleotide probes comprising the D1 and D2 domains. Appl. Environ.Microbiol. 69, 2899e2905.

Justé, A., Thomma, A.B.P.H.J., Lievens, B., 2008. Recent advances in moleculartechniques to study microbial communities in food-associated matrices andprocesses. Food Microbiol. 25, 745e761.

Millet, V., Lonvaud-Funel, A., 2000. The viable but non-culturable state of winemicroorganisms during storage. Lett. Appl. Microbiol. 30, 136e141.

Mills, D.A., Johannsen, E.A., Cocolin, L., 2002. Yeast diversity and persistence inbotrytis-affected wine fermentations. Appl. Environ. Microbiol. 68, 4884e4893.

Mills, D.A., Phister, T., Neeley, E., Johannsen, E., 2008. Wine fermentation. In:Cocolin, L., Ercolini, D. (Eds.), Molecular Techniques in the Microbial Ecology ofFermented Food. Springer Publishing Inc, New York, pp. 162e192.

Nissen, P., Nielsen, D., Arneborg, N., 2003. Viable Saccharomyces cerevisiae cells athigh concentrations cause early growth arrest of non-Saccharomyces yeast inmixed cultures by a cell-cell-contact-mediated mechanism. Yeast 20, 331e341.

Oliver, J.D., 2005. The viable but non culturable state in bacteria. J. Microbiol. 43,93e100.

Pérez-Nevado, F., Albergaria, H., Hogg, T., Girio, F., 2006. Cellular death of two non-Saccharomyces wine-related yeasts during mixed fermentations with Saccha-romyces cerevisiae. Int. J. Food Microbiol. 108, 336e345.

Phister, T.G., Mills, D.A., 2003. Real-time PCR assay for detection and enumeration ofDekkera bruxellensis in wine. Appl. Environ. Microbiol. 69, 7430e7434.

Rantsiou, K., Urso, R., Lacumin, L., Cantoni, C., Cattaneo, P., Comi, G., Cocolin, L., 2005.Culture-dependent and -independent methods to investigate the microbialecology of Italian fermented sausages. Appl. Environ. Microbiol. 71, 1977e1986.

Rigottier-Gois, L., Bourhis, A.G., Gramet, G., Rochet, V., Doré, J., 2003. Fluorescenthybridisation combined with flow cytometry and hybridisation of total RNA toanalyse the composition of microbial communities in human faeces using 16SrRNA probes. FEMS Microbiol. Ecol. 43, 237e245.

Stender, H., Kurtzman, C., Hyldig-Nielsen, J.J., Sørensen, D., Broomer, A.,Oliveira, K., Perry-O’Keefe, H., Sage, A., Young, B., Coull, J., 2001. Identificationof Dekkera bruxellensis (Brettanomyces) from wine by fluorescence in situhybridization using peptide nucleic acid probes. Appl. Environ. Microbiol. 67,938e941.

Vandervliet, G.M.E., Schepers, P., Schukkink, R.A.F., Vangemen, B., Klatser, P.R., 1994.Assessment of mycobacterial viability by RNA amplification. Antimicrob. Ag.Chemother. 38, 1959e1965.

Wallner, G., Amann, R., Beisker, W., 1993. Optimizing fluorescent in situ hybrid-ization with rRNA-targeted oligonucleotide probes for flow cytometry identi-fication of microorganisms. Cytometry 14, 136e143.

Ward, D.M., Bateson, M.M., Weller, R., Ruff-Roberts, A.L., 1992. Ribosomal RNAanalysis of microorganisms as they occur in nature. In: Marshall, K.C. (Ed.),Advances in Microbial Ecology, vol. 12. Plenum Press, New York, pp. 219e286.

Xufre, A., Albergaria, H., Inacio, J., Spencer-Martins, I., Girio, F., 2006. Application offluorescence in situ hybridization (FISH) to the analysis of yeast populationdynamics in winery and laboratory grape must fermentations. Int. J. FoodMicrobiol. 108, 376e384.

Zott, K., Claisse, O., Lucas, P., Coulon, J., Lonvaud-Funel, A., Masneuf-Pomarede, I.,2010. Characterization of the yeast ecosystem in grape must and wine usingreal-time PCR. Food Microbiol. 27, 559e567.