at SciVerse ScienceDirect

Food Microbiology 36 (2013) 267e274

Contents lists available

Food Microbiology

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

Genetic diversity of Oenoccoccus oeni isolated fromwines treated withphenolic extracts as antimicrobial agents

Almudena García-Ruiz a, Raquel Tabasco a, Teresa Requena a, Olivier Claisse b,c,Aline Lonvaud-Funel b, Begoña Bartolomé a, M. Victoria Moreno-Arribas a,*

a Instituto de Investigación en Ciencias de la Alimentación (CIAL), CSIC-UAM, Nicolás Cabrera 9, CEI UAMþCSIC, Campus de Cantoblanco,Universidad Autónoma de Madrid, 28049 Madrid, SpainbUniversité de Bordeaux, ISVV, EA 4577 œnologie, 210 Chemin de Leysotte, F-33140 Villenave d’Ornon, Francec INRA, ISVV, USC 1366 Œnologie, 210 Chemin de Leysotte, F-33140 Villenave d’Ornon, France

a r t i c l e i n f o

Article history:Received 28 January 2013Received in revised form24 May 2013Accepted 24 June 2013Available online 2 July 2013

Keywords:Antimicrobial phenolic extractsMalolactic fermentationOenococcus oeniGenetic characterization

* Corresponding author.E-mail address: victoria.moreno@csic.es (M.V. Mo

0740-0020/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.fm.2013.06.012

a b s t r a c t

Molecular techniques have been applied to study the evolution of wine-associated lactic acid bacteriafrom red wines produced in the absence and presence of antimicrobial phenolic extracts, eucalyptusleaves and almond skins, and to genetically characterize representative Oenococcus oeni strains. Moni-toring microbial populations by PCR-DGGE targeting the rpoB gene revealed that O. oeniwas, as expected,the species responsible for malolactic fermentation (MLF). Representative strains from both extract-treated and not-treated wines were isolated and all were identified as O. oeni species, by 16S rRNAsequencing. Typing of isolated O. oeni strains based on the mutation of the rpoB gene suggested a morefavorable adaptation of L strains (n ¼ 63) than H strains (n ¼ 3) to MLF. Moreover, PFGE analysis of theisolated O. oeni strains revealed 27 different genetic profiles, which indicates a rich biodiversity ofindigenous O. oeni species in the winery. Finally, a higher number of genetic markers were shown in thegenome of strains from control wines than strains from wines elaborated with phenolic extracts. Theseresults provide a basis for further investigation of the molecular and evolutionary mechanisms leading tothe prevalence of O. oeni in wines treated with polyphenols as inhibitor compounds.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Malolactic fermentation (MLF) is a biological process that usu-ally occurs once alcoholic fermentation (AF) by yeast is completed.MLF is commonly performed by the indigenous lactic acid bacteria(LAB) existing in grapes andwineries, although sometimes it can beinduced by starter cultures. These bacteria are responsible for thedegradation of malic acid into lactic acid and carbon dioxide, pro-ducing a reduction in total acidity of the wine. This biologicaldeacidification is always accompanied by the provision of addi-tional flavors and microbiological stability for wines (Lonvaud-Funel, 1999; Moreno-Arribas and Polo, 2005). In the majority ofcases, Oenococcus oeni is the most tolerant species of unfavorablewine conditions (low pH and high ethanol levels), being the mainspecies conducting MLF in wine (Davis et al., 1985; van Vuuren andDicks, 1993).

Once malic acid is fully transformed, microbial populations arecontrolled by addition of sulfur dioxide in order to avoid any post-

reno-Arribas).

All rights reserved.

fermentation microbial metabolism that could alter the organo-leptic quality of wines. Most of the bacteria and possible remainingyeasts are sensitive to sulfur dioxide, although the effectiveness ofSO2 may be limited by wine pH and other wine components. Thus,in certain conditions, Lactobacillus and Pediococcus may be pre-dominant and induce wine spoilage. Nowadays, there is a world-wide trend to reduce SO2 levels in wine, so totally or partiallynatural alternatives to the traditional use of SO2 in winemaking,such as plant polyphenols are claimed (García-Ruiz et al., 2008).

The advances in molecular tools, usually based on polymerasechain reaction (PCR) techniques, have allowed a fast and sensitivecharacterization of the majority of wine LAB. The intraspecific di-versity of O. oeni and strain typing is also studied by enzymaticrestriction coupled with restriction endonuclease analysis bypulsed-field gel electrophoresis (REA-PFGE) (Gindreau et al., 1997).By PCR followed by Denaturing Gradient Gel Electrophoresis(DGGE), the visualization of the microbial population diversity in acomplex community is possible (Pozo-Bayón et al., 2009). More-over, it includes the detection of the non-cultivable microbiota.DGGE is based on the separation of PCR amplicons of different se-quences and the same size. For wine bacteria, the gene coding for

A. García-Ruiz et al. / Food Microbiology 36 (2013) 267e274268

the beta subunit RNA polymerase (rpoB gene), which is present as aunique copy in the genome, is the most reliable target for thisanalysis. It provides more phylogenetic resolution than the 16SrRNA gene which is repeated, with differences between the copies,leading sometimes to ambiguous profiles (Renouf et al., 2006).Unexpectedly, the rpoB analysis shows for O. oeni two close butdifferent bands in the DGGE gels: the L band as the lower-migratedband, and the H band as the higher-migrated band in the gel. Thesetwo rpoB sequences differed by only one nucleotide: a guanidine forL was substituted by an adenine for H (Renouf et al., 2006). Inanother study, Renouf et al. (2008) suggested that 16 geneticmarkers may possibly be linked to enological properties of O. oenistrains, such as survival, multiplication in wine and the ability toperform MLF. This genetic characterization is important for un-derstanding the selection mechanism during the first stages ofwinemaking.

In a previous study, and after screening a great number of plantextracts for antimicrobial properties against O. oeni and spoilagespecies of LAB in pure cultures, we tested the technological appli-cability, as an alternative to SO2, of an extract from eucalyptusleaves during the MLF of a red wine (García-Ruiz et al., 2012). Incomparison with the control wines, the malic acid consumption inthe wines treated with eucalyptus extract (2 g/L) was lower in bothinoculated and spontaneous MLF (García-Ruiz et al., 2012), sug-gesting the potential application of natural phenolic extracts in thecontrol of MLF. However, it is unknown if the addition of thisantimicrobial phenolic extract may affect the evolution of the LABpopulation, especially O. oeni, duringMLF. Therefore, the aim of thiswork was to genetically type wine-associated LAB from red winesobtained in the absence and presence of antimicrobial phenolicextracts and to genetically characterize representative O. oenistrains by (i) targeting the rpoB gene, (ii) comparing the PFGEprofiles and (iii) analyzing the presence/absence of enological ge-netic markers that seem related to the adaptation of LAB to thewine environment. As well as the eucalyptus extract previouslytested in MLF experiments, a second extract from almond skins e

also active against the growth of enological LAB strains (García-Ruizet al., 2012) ewas also selected for the study. Overall, our final goalis to expand the knowledge of the effect of antimicrobial agents(i.e., phenolic extracts) on the microbial and genetic diversity ofwine microbiota, in particular of O. oeni.

2. Materials and methods

2.1. Malolactic fermentation assays in wine

A redwine (var.Merlot) (vintage 2009)was elaborated at BodegasMiguel Torres S.A. (Catalonia, Spain), following their own wine-making procedures (García-Ruiz et al., 2012). The AF was carried outin a controlled form in stainless steel (10,000 L) at 25� 2 �C. The endof AF was established by measuring the alcohol degree (13.9% v/v)and the residual sugar amount (<3.5 g/L); the wine pH at the end ofAF was 3.22. MLF experiments were conducted in laboratory-scale,sterile conditions, in 250 mL flasks. Parallel inoculated and sponta-neous MLF experiments were carried out; O. oeni population at thebeginning of these fermentations was <103 CFU/mL and >104 CFU/mL, respectively. The plant extracts (from eucalyptus leaves andalmond skins) were dissolved (2 g/L) in 200 mL of previously inoc-ulated or non-inoculated wine. Commercial phenolic extracts fromeucalyptus leaves and almond skins were kindly provided by theirproducer, Biosearch Life S. A. (Granada, Spain). Phenolic content was89 and 165 mg of gallic acid equivalents/g for the eucalyptus andalmond extracts, respectively (García-Ruiz et al., 2012).

The malolactic starter comprised a mix of three O. oeni strainspreviously isolated by the winery, and was inoculated inwine at 3%

(v/v). A control containing no extract was also prepared for bothinoculated and spontaneous MLF experiments. An extra positivecontrol containing K2S2O5 (30 mg/L) as an antimicrobial agent wasalso prepared for the inoculated MLF assay. Control wines andwines containing phenolic extracts or sulfites, all in duplicate, wereincubated at 25 �C in the dark. During the incubation, the winecontent of L-malic acid was monitored using an enzymatic kit(Megazyme International Ireland Ltd., Bray, CO. Wicklow, Ireland),with determinations being carried out in duplicate. Initial contentof L-malic acid was 0.90 g/L. MLF was considered finished when thecontent of L-malic acid was �0.05 g/L. 50 mL of each type of redwine were aseptically collected (start, middle and end of MLF) andcentrifuged (10 min, 10,000 g, 4 �C). The pellets were kept in acommercial freezer (�20 �C) until the molecular analysis.

2.2. LAB isolation

Wine samples were diluted and plated on MRSAgar medium(Pronadisa, Madrid, Spain), supplemented with 5 g/L fructose(Panreac Química SAU, Barcelona, Spain); 1 g/L DeL malic acid(Panreac Química SAU, Barcelona, Spain), 1 mL Tween 80 (SigmaeAldrich, St. Louis, MO, USA) and 100 mg/L cycloheximide (SigmaeAldrich) were also added to the medium to suppress yeast growth.The pH of the medium was adjusted to 4.8 with 37% HCl (PanreacQuímica SAU). Plates were incubated anaerobically (WhitehouseStation, New Jersey, USA) at 28 �C for seven days. At each day’sanalysis, ten isolated colonies were randomly chosen from a plateof convenient sample dilutions, ensuring that all different colonymorphologies were considered. Isolates were subcultured onto thesame medium until purification. Each pure colony was cultured inliquid medium, with a similar composition to that of the plates butwithout agar, and was stored at �80 �C with 50% (v/v) glycerol(Panreac Química SAU).

2.3. Bacteria strains and culture conditions

The reference strains Lactobacillus plantarum CECT 4645,Lactobacillus casei CECT 4045, Pediococcus parvulus CECT 4693 andO. oeni CECT 217 from the Spanish Type Culture Collection (CECT)and the LAB isolated from wines were used in this study.

Following CECT recommendations, the Lactobacillus and Ped-iococcus species were grown in MRS broth (Pronadisa). O. oeni andLAB isolated from wines were grown in MRS broth, supplementedwith 5 g/L fructose (Panreac Química SAU) and 1 g/L DeL malic acid(Panreac Química SAU), pH 4.8 (37% HCl).

2.4. DNA extraction

For 16S rRNA gene sequencing and PCR-DGGE targeting the rpoBgene, the DNA from isolated strains was extracted using theQIAamp DNA kit (Qiagen, Hilden, Germany), according to the pro-tocol described by the manufacturer. The isolated DNA was storedat�20 �C until the analyses. DNA concentrationswere standardized(100 ng/mL) by measuring optical density at 260 nm with aSmartSpec (þ) spectrophotometer (Bio-Rad, Hercules, CA, USA).

For REA-PFGE analysis, strains were cultivated on MRS brothduring 3e4 days and cells were collected by centrifugation (5 min,10,000 g, 4 �C). The supernatant was discarded and the pelletresuspended in 600 mL of 50 mM EDTA, pH 8, with 10 mg/mL oflysozyme (SigmaeAldrich) and incubated for 1 h at 37 �C as Bilhèreet al. (2009). After a second centrifugation (2 min, 10,000 g, 4 �C),the supernatant was newly discarded and the pellet resuspended in600 mL of nucleic lysis solution (Promega, Madison,WI, USA), mixedsoftly with the pipette and incubated for 5 min at 80 �C. Then,200 mL of protein precipitation solution (Promega) were added and

A. García-Ruiz et al. / Food Microbiology 36 (2013) 267e274 269

mixed for 20 s. Cellular fragments were precipitated on ice for5 min. After another centrifugation (3 min, 10,000 g, 4 �C), thesupernatant containing the DNA was transferred to a new micro-centrifuge tube containing 600 mL of isopropanol and gently mixedby inversion. After centrifugation (2 min, 10,000 g, 4 �C), 600 mL of aroom temperature 70% ethanol solution were added to the pelletbefore a final centrifugation (2 min, 10,000 g, 4 �C). Ethanol wascarefully removed and the tube dried. Fifty microliters of waterwith 3 mL of RNase (Promega) were used to resuspend DNA over-night at 4 �C.

2.5. Identification of isolated colonies: 16S rRNA gene sequencing

LAB colonies isolated from experimental wines were identifiedby sequencing the 16S rRNA gene. The 16S rRNA genewas amplifiedby PCR using primers POmod and PC5rev (Table 1) and the PCRconditions described by Rodtong and Tannock (1993). Additionalprimers used to assist in sequencing were 16Smidfor and P3rev(Table 1). Sequencing of PCR fragments was carried out at the DNAsequence service of the Centro de Investigaciones Biológicas-CSIC(Madrid, Spain). The resulting sequences were used to search se-quences deposited in a database using the BLAST algorithm. Theidentity of the strains was determined on the basis of the highestscore.

2.6. PCR-DGGE

The PCR-DGGE protocol using rpoB1, rpoB1o and rpoB2 primers(Table 1) and described by Renouf et al. (2006) was used withmodifications in band staining. The PCR program began with aninitial touchdown step in which the annealing temperature waslowered from 59 to 45 �C in intervals of 1 �C every cycle. Further-more, 20 additional cycles were carried out with an annealingtemperature of 45 �C. DGGE was performed with a DCode system(Bio-Rad), using a 9% polyacrylamide gel with a 30e60% gradient of

Table 1Primers used in this study.

Genes/Markers Forward primer (50 / 30)

16S rRNAPOmod/P3rev CAGAGTTTGATCCTGGCTCAG16midfor/PC5rev GGCCGTTACTGACGCTGAGBeta subunit RNA polymeraserpoB1, rpoB1o/rpoB2 rev ATTGACCACTTGGGTAACCGTCG

ATCGATCACTTAGGCAATCGTCGMarkerCadmium transporting P-type GAAGCTCAAGATACCATCCATPase-M1Dps ferritine-M2 TTGGTTAATTCAGCGCCGTTGTPolysaccharide biosynthesis export CTCGTAAGCATGGTTCTCTCProtein-M3Maltose phosphorylase-M4 ACGCATGATTCCTCATTATTATCTranscriptional regulator-M5 TGGCAAACGTCTCAATCAACHypothetical protein-M6 TACTGTTCGTCAGCCGATGTPredicted transcriptional CAATCAAGCCGGAATAGTTRegulators-M7Hypothetical protein-M8 ATGACGCCATTCTATATCCASugar-alcohol dehydrogenase-M9 GGAAACAATTTACGCTTGCCopper chaperone-M10 CCTCCTACTTAACCTTGACGArabinose efflux protein MFS-M11 TGGCTTAATCCCATCAGAAAThioredoxin-M12 GTTTCTGAAGACCCGCTTAGlycerol uptake facilitator CTAACGCATTCCTGAAGAACProtein-M13Arabinose efflux permease-M14 TTTATCTGTCCAAGCAGGTGlycosyl transferases involved in TGTTAACGATACGAAGCGCGCell wall biogenesis-M15Hypothetical protein lp_3433-M16 AAATAACGCAGGCCAATC

7 M urea (SigmaeAldrich) and 40% formamide (SigmaeAldrich)that increased in the electrophoresis running direction. Electro-phoresis was run in a 1x TAE buffer (20 mM Tris, 10 mM acetic acidand 0.5 mM EDTA) at constant temperature (60 �C) for 10 min at20 V and subsequently for 16 h at 85 V. After migration, gels werestained with AgNO3 as described by Sanguinetti et al. (1994).

2.7. REA-PFGE

Strains were grown and the cultures centrifuged as describedabove. The pelleted cells were washed twice with 1x TE (10 mMTriseHCl, 1 mM EDTA, pH 8) and finally resuspended in 50 mL T100E(10 mM TriseHCl, 100 mM EDTA, pH 8). The cell suspensions wereheated at 50 �C and mixed with an equal volume of 1% (w/v)agarose (Chromosomal Grade Agarose (Bio-Rad, Hercules, CA, USA),which was pre-melted and kept at 60 �C. Aliquots were made intomoulds to prepare plugs and were kept for 15 min at 4 �C. Theagarose plugs were removed and placed in 1 mL lysis buffer (T100E,10 mg lysozyme (SigmaeAldrich) for 3 h at 37 �C. The lysis bufferwas replaced with a 1 mL pronase buffer (T100E, 2 mg of Pronase Efrom Streptomyces griseus (SigmaeAldrich), 1.5% N-lauryl sarcosyl(SigmaeAldrich) and incubated for 16 h at 37 �C. Afterward theplugs were washed four times in 1x TE with gentle shaking for30 min per wash. A third of a plug of each strain was digested withNotI restriction endonuclease (New England BioLabs, Ipswich, MA,USA) in a volume of 100 mL for 16 h at 25 �C. The plugs were rinsedwith 1x TE at 4 �C before electrophoresis. The digested DNA frag-ments were separated by electrophoresis in a 1% agarose gel (PulseField Certified Agarose, Bio-Rad) in 0.5x TBE buffer (0.1 M Tris,0.09 M boric acid, 0.01 M EDTA, pH 8) with a CHEF-DRIII apparatus(Bio-Rad). Electrophoresis was performed at 15 �C at 6 V/cm;interpolation pulse time of 25 s for 22 h. Gels were stained withethidium bromide (0.5 mg/mL) and photographed under UV light.The Midrange II PFG Marker (New England BioLabs) was used as asize marker and normalization reference. The DNA fingerprint

Reverse primer (50 / 30) Ampliconlength (bp)

GGCCGTTACTGACGCTGAG 792e825CTCACTATAGGGATACCTTGT-TACGACTT 767e771

CGCCCGCCGCGCGCGGGGCGG- 250GGGC ACGATCACGGGTCAAAC- C ACC

CGACTTGCACAGATTCC 650

ATTGATCACGATGTCCCAAC 500ATTGGTTTGATGAAAAATGG 565

GGTCTTTCAAAATACCATCG 600AGCTTACGGCTGATGCTTT 380CTCCCGACAAACTGCTAATG 400TGACCAGTTCGAATGAATTC 462

ATTTGCCTCGATAGTTTCTG 605CGGCCTGTTTGATAAAGAA 471AGTCCCACCTCCTGAATAAA 420CCAAATTGTCCAGAATACCG 600TGATGCCCCCTTCGTAAT 300CCCAACTATATTCCCAGTGA 602

AATTAGAAGAACGCTGATAGCC 330GAATCACTCCATTCCGTCACC 600

CCATGATTCCTGGTTTACTGAG 569

Fig. 1. DGGE profiles of wine samples elaborated in the presence/absence of antimi-crobial phenolic extracts during MLF. Lanes 1e3: wine elaborated in the absence ofphenolic extract 1: start MLF, 2: middle MLF, 3: end MLF; 4e6: wine added withalmond skins 4: start MLF, 5: middle MLF, 6: end MLF; 7e9: wine elaborated witheucalyptus leaves extract 7: start MLF, 8: middle MLF, 9: end MLF. The four last lanescorrespond to pure strains: lane A, Lactobacillus casei, lane B, Oenococcus oeni (type L),lane C, Pediococcus parvulus, lane D, Lactobacillus plantarum.

A. García-Ruiz et al. / Food Microbiology 36 (2013) 267e274270

patterns were analyzed using Bionumerics 5.1 software (AppliedMaths, Kortrijk, Belgium) with 1% optimization and 1% band posi-tion tolerance settings. The comparison of profiles obtained wasperformed with the Dice coefficient and the Unweighted-PairGroup Method with Arithmetic means (UPGMA).

2.8. Genetic characterization: presence of gene markers

The presence of 16 geneticmarkers (Table 1) was determined forO. oeni strains isolated during the MLF process. The genetic char-acterization protocol was performed using the method of Renoufet al. (2008). Each 25 mL amplification reaction mixture containeda 2 ng DNA template, 12.5 mL custom-made PCR Master Mix(Finnzymes, Espoo, Finland) and 5 pmol of each primer. The reac-tion mixture was preheated for 5 min at 95 �C and subjected to 30cycles, each consisting of denaturing (30 s, 95 �C), annealing (30 s,55 �C) and an extension step (30 s, 72 �C), in an iCycler IQ (Bio-Rad).In addition to the conventional negative PCR control run withoutDNA, a positive control with the DNA of O. oeni strains (Table 2) wasused. These strains belong to the bacterial culture collection fromthe Biological Resource Center Enology (http://www.crboeno.univ-bordeauxsegalen.fr/). Amplified products were resolved on Mul-tiNA MCE 202 capillary electrophoresis system (Shimadzu Biotech.,Kyoto, Japan) using the kit DNA1000 with an internal size calibratorand the separation buffer (Shimadzu Biotech.) with SyBr Gold(Invitrogen, Paris, France). The phi-X-174 HaeIII DNA ladder (NewEngland BioLabs) was also used for calculation of the fragmentsizes.

3. Results

3.1. Monitoring the LAB population in wines

PCR-DGGE has been used to study the evolution of the LABpopulation from red wines elaborated in the absence and presenceof antimicrobial phenolic extracts (eucalyptus leaves and almondskins). For this analysis, the PCR-rpoB amplicons obtained fromL. plantarum CECT 4645, L. casei CECT 4045, P. parvulus CECT 4693and O. oeni CECT 217 (type L) were used as reference markers.

Fig. 1 shows a representative rpoB PCR-DGGE gel correspondingto wines subjected to spontaneous MLF in the absence and pres-ence of antimicrobial phenolic extracts (eucalyptus leaves andalmond skins). Amaximum of five different bands per sample couldbe revealed on DGGE gel during MLF. By reference to themarkers, itwas only possible to identify the rpoB band corresponding toO. oeni. In the control wine, the five bandswere detected at the startof MLF. Then, the O. oeni band was the only one detected in thefollowing days. In the presence of antimicrobial phenolic extracts,five bands were seen for the samples collected at the start andmiddle of MLF. At the end of MLF, two bands were respectivelyrevealed in the red wine added with almond skins and witheucalyptus leaves extracts. The band corresponding to O. oeni wasobserved in both profiles. With regard to the wines inoculated withthe O. oeni starter, in the absence and presence of antimicrobialphenolic extracts, and with addition of SO2, the PCR-DGGE revealed

Table 2Oenococcus oeni strains positive control to genetic characterization.

Strains Markers

O. oeni 7.147 All excepted M8 and M9O. oeni 7.135 M4,8,9,11,12,14,15O. oeni 7.125 All excepted M4O. oeni 10.13 All excepted M9O. oeni 10.10 M3,9,11,14,15

few bands (results not shown). During MLF and at the end of MLF,the band corresponding with O. oeni was consistently identified.

3.2. Identification of isolated colonies

A total of 66 colonies isolated from the wines undergoingspontaneous or inoculated MLF in absence or presence of antimi-crobial phenolic extracts and SO2 were identified by 16S rRNAsequencing as O. oeni. Moreover, the 66 isolated colonies were alsoanalyzed by rpoB PCR-DGGE assay. As expected, we obtained twodifferent profiles (L and H) corresponding to the two rpoB ampliconsequences. Within all the 66 strains collected there were 3 strainstyped as H and 63 L strains (results not shown). The 3 H strainswere isolated from the control wine inoculated with the malolacticstarter, once the MLF was performed. This result reflected differ-ences between the strains isolated from inoculated wines inabsence (H and L strains) or presence (L strains) of phenolic extractsat the end of MLF. In contrast, the strains isolated from spontaneouswines did not show differences among them, all the strains weretype as L. On the other hand, the analysis of the starters showedstrains characterized by L and H bands.

3.3. Genotypic characterization of O. oeni strains

A total of 43 O. oeni colonies isolated from spontaneous fer-mentations (n¼ 23, being approximately 3e4 strains from differenttype of wine and sampling time), inoculated fermentations (n ¼ 16,being approximately 2e4 strains from different type of wine andsampling time), and from the malolactic starter (n ¼ 4) werecharacterized genotypically by REA-PFGE (Table 3). The number ofO. oeni isolates was higher in the wines subjected to spontaneousMLF than in the wines inoculated with malolactic starter, assuminga greater microbial biodiversity in the spontaneous MLF red wine.

O. oeni genomic DNA digested with NotI yielded 5 to 11 bands inthe 30 kbe450 kb size range. Cluster analysis of the PFGE profiles ofthe 43 O. oeni isolates revealed 27 genotypes with specific profiles(Fig. 2). The percentage of similarity between unrelated profiles

Table 3Oenococcus oeni strains isolated from spontaneous and inoculated malolactic fermentation red wines elaborated in the absence/presence of antimicrobial phenolic extracts:almond skins and eucalyptus leaves, and sulfur dioxide (SO2).

Red wine Treatment Samplingtimea

No. O. oeniisolates

RepresentativeO. oeni isolates

DGGEprofiles

PFGEprofiles

Spontaneous MLF Control 0 10 SCtW.00 L 12SCtW.03 L 5SCtW.06 L 7SCtW.09 L 7

1 10 SCtW.11 L 7SCtW.14 L 7SCtW.17 L 22

2 10 SCtW.22 L 1SCtW.23 L 4SCtW.27 L 4SCtW.28 L 3

Almond skins 1 10 SWA.13 L 21SWA.14 L 6SWA.15 L 2SWA.16 L 8

2 10 SWA.20 L 9SWA.23 L 11SWA.25 L 4SWA.28 L 11

Eucalyptus leaves extract 1 10 SWE.10 L 10SWE.12 L 3SWE.13 L 4SWE.14 L 4

Inoculated MLF Control 0 10 ICtW.01 L 25ICtW.08 L 27

2 10 ICtW.22 L 18ICtW.23 L 18ICtW.24 H 19ICtW.25 H 13

SO2 0 10 IS02.00 L 13IS02.01 L 15

1 10 IS02.10 L 17IS02.13 L 13

2 10 IS02.23 L 23IS02.24 L 16

Almond skins 2 10 IWA.24 L 15IWA.26 L 18

Eucalyptus leaves extract 1 10 IWE.12 L 15IWE.14 L 14

Starter 10 St.2 H 26St.3 L 5St.5 H 20St.6 L 24

a Sampling time: 0 (start MLF), 1 (middle MLF), 2 (end MLF).

A. García-Ruiz et al. / Food Microbiology 36 (2013) 267e274 271

varied from 27% to 93%. The results showed a clear separation be-tween O. oeni isolated from wines subjected to spontaneous MLFand those isolated from wines inoculated with malolactic starter(Fig. 2).

The analysis by REA-PFGE of the O. oeni starters (Fig. 2) revealedthat one colony isolated from the CtW.3 spontaneous MLF red winesample had the same PFGE profile than the starter 3 (St3) that(Fig. 2). The rest of the starters analyzed (St.2,5 and 6) were clus-tered, as expected, together with the colonies isolated from inoc-ulated wines. However, the percentage of similarity betweenO. oeni starters and O. oeni isolated from wines subjected to inoc-ulatedMLFwas from 55% to 93%, indicating that the starters did notdominate continuously during MLF.

With respect to the O. oeni isolated from wines subjected tospontaneous MLF, the analysis by REA-PFGE yielded 5e11 bands;most of the isolated strains showed 7 bands. The 23 O. oeni isolateswere separated into 14 different PFGE profiles (Fig. 2). The strainsisolated from the control wine as well as the isolated from thewines treated with eucalyptus and almond extracts were charac-terized by showing a high number of profiles (�3) (Fig. 2). Thestrains Ct.17 and WA.13 exhibited a greater similarity with thecolonies isolated fromMLF-inoculatedwines thanwith the colonies

isolated from spontaneous MLF red wine. Profiles number 4 and 7showed the highest number of strains with five and four isolates,respectively. Profile 4 consisted of strains isolated from red wineelaborated in absence and presence of both antimicrobial phenolicextracts, whereas the strains of profile 7 were isolated from thecontrol wine (absence of phenolic extracts).

In reference to the O. oeni strains isolated fromwines inoculatedwith malolactic starter, the results by REA-PFGE revealed 7e10bands; most of the O. oeni isolated showed 8 bands. The 16 O. oenistrains were classified into 10 unrelated PFGE profiles (Fig. 2).Furthermore, the strains isolated from inoculated wines treated ornot with phenolic extracts and SO2, were characterized by showingmore than one genetic profile (Fig. 2). Profile 13 stood out as beingformed by strains isolated from control wine or sulfited wines,while profile 15 consisted of strains isolated from wine elaboratedin the presence of antimicrobial phenolic extracts (eucalyptusleaves and almond skins) or sulfited.

3.4. Presence of gene markers

Some strains isolated from both spontaneous and inoculatedMLF were characterized genetically by the presence of 16

Fig. 2. UPGMA dendrogram based on the NotI REA-PFGE profiles of the 43 Oenococcus oeni strains examined in this study, which showed 27 unrelated patterns, and four O. oenimalolactic starters.

A. García-Ruiz et al. / Food Microbiology 36 (2013) 267e274272

significant genetic markers (M1eM16, Table 1). The strains assayedrepresented profiles 3, 4, 7, 13 and 15 (Fig. 2). As shown in Table 4, 6out of the 16 markers studied were present in the profiles selected:polysaccharide biosynthesis export protein (M3), present in profiles3 and 7; predicted transcriptional regulators (M7), present in all

Table 4Presence (þ) or absence (�) of 16 enological genetic markers.

REA-PFGE profiles M1 M2 M3 M4 M5 M6 M7 M8

Control þ þ þ þ þ þ þ þ þProfile 3 e e þ e e e þ e

Profile 4 e e e e e e þ e

Profile 7 e e þ e e e þ þProfile 13 e e e e e e þ e

Profile 15 e e e e e e þ þControl - e e e e e e e e

patterns; hypothetical protein (M8), present in profiles 7 and 15;sugar-alcohol dehydrogenase (M9), present in all profiles exceptpattern 3; arabinose efflux protein MFS (M11), present only inpattern 13; and glucosyltransferase involved in cell wall biogenesis(M15), which was present in all profiles except pattern 13. This

M9 M10 M11 M12 M13 M14 M15 M16 Total

þ þ þ þ þ þ þ þe e e e e e þ e 3þ e e e e e þ e 3þ e e e e e þ e 5þ e þ e e e þ e 4þ e e e e e e e 3e e e e e e e e

A. García-Ruiz et al. / Food Microbiology 36 (2013) 267e274 273

result suggested a smaller number of markers in the genome ofstrains from MLF, spontaneous or inoculated, carried out in thepresence of antimicrobial phenolic extracts (profiles 3, 4 and 15)than in the strains from bothMLF without addition of antimicrobialphenolic extracts (profiles 7 and 13).

4. Discussion

In this work, different molecular tools were applied with theaim of studying the evolution of wine-associated LAB from redwines elaborated in the absence and presence of antimicrobialphenolic extracts added after AF, and of genetically characterizingrepresentative O. oeni strains. Phenolic extracts from eucalyptusleaves and almond skins were added to wines at a concentration of2 g/L according to the IC50 data of these extracts that exhibitedantimicrobial activity against O. oeni (García-Ruiz et al., 2012). Atthat concentration, neither of these two antimicrobial extractsseemed to modify adversely the flavor properties of the winestreated with them (García-Ruiz et al., 2013). Initially, we employedPCR-DGGE to study the diversity and evolution of the wine LABpopulation. This technique has been used successfully in moni-toring the fermentation of red (Renouf et al., 2006, 2007; Spanoet al., 2007) and white (Renouf et al., 2005) wines. The resultsshowed greater microbial diversity at the beginning of MLF and adecrease as MLF progressed, with the exception of the wine treatedwith eucalyptus extract and subjected to inoculated MLF. In all thewines analyzed, amaximumof five bandswere detected at the startof MLF, but only the band corresponding to O. oeni could be iden-tified. At the end of MLF, O. oeniwas the predominant species in thewines tested. This result was as expected, since many studies hadshown before that O. oeni is the main species responsible for MLF(Dicks and Vanvuuren, 1988; Reguant and Bordons, 2003; Lópezet al. 2007; Ruiz et al. 2010).

The sequencing of the 16S rRNA gene and the PCR-DGGE tar-geting the rpoB gene enabled us to identify 66 strains isolated fromboth spontaneous and inoculated MLF fermentations at differentstages of the MLF process. By both methods, the 100% isolatedstrains were identified as O. oeni. This result again confirmed thedominance of the O. oeni species during the MLF of the winesstudied. As expected, the rpoB analysis showed two different pro-files (L and H) corresponding to themigration of two rpoB ampliconbands. DGGE gels revealed a total of 63 L and 3 H O. oeni strains,which suggested a more favorable adaptation of L strains to bothMLF. These results were in line with the results of Renouf et al.(2009) on the prevalence of L-strains over H-strains during MLF.Out of the four starters, twowere of the H type and twowere L type.

Typing of the O. oeni strains in this study was successfully ach-ieved by REA-PFGE, with NotI being the restriction enzymeemployed for this analysis. This molecular tool is considered to bethemost powerful method for strain typing (López et al., 2008). Theresulting 27 unrelated genotypes out of the total of 43O. oeni strainsisolated in this study indicated a rich biodiversity of indigenousO. oeni strains in the winery. As observed in Fig. 2, the 27 patternswere grouped clearly into two clusters corresponding to the twodifferent types of MLF, spontaneous and inoculated with malolacticstarters. This result suggested that the biodiversity of O. oeni wasmore influenced by the type ofMLF, inoculated or spontaneous, thatby the addition of phenolic extracts, eucalyptus leaves and almondskins. Some profiles were more represented than others, forexample profiles 4, 7, 13, 14 and 18. However, whatever the wine,inoculated or not, there was no dominant profile that would haveshown that some strains would be more or less tolerant to theantimicrobial phenolic extracts, eucalyptus leaves and almondskins.With regard to theO. oeni starters, the strain of starter St.3wasfound in the spontaneous fermentation in the control wine

(SCtW.03); this indicated that this strain was definitely present inthe winery environment. The profile of starter St.5 was never foundandprofiles close, but not identical, to St.2 and St.6were found in theinoculated wines. The high diversity of strains in the inoculatedwines showed how difficult it was for the starter to dominate theindigenous microbiota. This result could also be due to a concen-tration of indigenous O. oeni in wine higher (>104 CFU/mL) thanexpected at the time of inoculation of malolactic starter.

From the REA-PFGE results, some strains were characterized bythe presence of 16 enological markers (M1eM16). They repre-sented profiles 3, 4, 7, 13 and 15. Some markers may be character-ized by resistance to environmental stress (M1 and M12), othersmay be important for the transport of metabolites (M11, M13 andM14), while others may have essential cellular functions (M5, M7and M15) (Renouf et al., 2008). Six out of the 16 markers studiedwere present in the genome of selected strains (Table 4): M7 in allthe strains, M9 in all except pattern 3, M15 in all except profile 13,M8 in patterns 7 and 15, and finally M11 in profile 13. The presenceof markers M7, M9 andM15 in all or almost all characterized strainscould indicate that they were essential for the survival of bacteriaduring MLF. These markers may be responsible for resistance/response to stress through high sugar and ethanol concentrations(M9), cellular functions viz. the cell wall organization (M15) and thetranscription (M7). Indeed, it was showed a tendency for a highernumber of markers in the genome of strains fromwines under MLFcarried out without the addition of antimicrobial phenolic extracts(profiles 7 and 13), being the number of genetic markers in thecontrol wine from spontaneous MLF (profile 7) higher than in thecontrol wine from inoculated MLF (profile 13) (Table 4). These re-sults were in line with Renouf et al. (2008), where these 6 markerswere present with a higher percentage in the strains selectedduring the industrial winemaking of three wines. Moreover theyput forward that the type ofMLF could also influence the number ofmarkers and the biodiversity observed. As a whole, it suggests theidentification of a great diversity of genetic markers that allows us abetter evaluation of the capacity of resistance/survival of O. oeni tothe conditions in which the MFL takes places in presence of anti-microbial phenolic extracts.

In summary, we concluded that O. oeni was the speciesresponsible for MLF in the wines elaborated in the absence andpresence of antimicrobial phenolic extracts (eucalyptus leaves andalmond skins). DGGE gels showed a more favorable adaptation of LO. oeni strains than H strains to MLF. The high number of profilesrevealed in the REA-PFGE analysis indicated a rich biodiversity ofindigenous O. oeni strains in the winery. And finally, the strainsfromMLF, spontaneous or inoculated, carried out in the presence ofantimicrobial phenolic extracts (eucalyptus leaves and almondskins) presented differences in their genetic markers in comparisonwith strains from wines not exposed to antimicrobial phenolicextracts. Furthermore, this study contributes to providing a basisfor further investigation of the molecular and evolutionary mech-anisms leading to the prevalence of some O. oeni strains in winestreated with polyphenols as particular inhibitors.

Acknowledgments

This work has been funded by the Spanish Ministry for Scienceand Innovation (AGL2009-13361-C02-00, AGL2012-40172-C02-01,PRI-PIBAR-2011-1358, INIA-RM2011-00003-00-00 and CSD2007-00063 Consolider Ingenio 2010 FUN-C-FOOD Projects) and theComunidad de Madrid (ALIBIRD P2009/AGR-1469 Project). AGR isthe recipient of a fellowship by the JAE-Program (CSIC) and theDanone Institute. The authors would like to thank Bodegas MiguelTorres S.A. winery for their collaboration and the companies thatproduced the phenolic extracts by the samples supplied.

A. García-Ruiz et al. / Food Microbiology 36 (2013) 267e274274

References

Bilhère, E., Lucas, P.M., Claisse, O., Lonvaud-Funel, A., 2009. Multilocus sequencetyping of Oenococcus oeni: detection of two subpopulations shaped by inter-genic recombinations. Applied and Environmental Microbiology 75, 1291e1300.

Davis, C., Silveira, N.F., Fleet, G.H., 1985. Occurrence and properties of bacterio-phages of Leuconostoc oenos in Australian wines. Applied and EnvironmentalMicrobiology 50, 872e876.

Dicks, L.M.T., Vanvuuren, H.J.J., 1988. Identification and physiological-characteristicsof heterofermentative strains of Lactobacillus from South-African red wines.Journal of Applied Bacteriology 64, 505e513.

García-Ruiz, A., Bartolomé, B., Martínez-Rodríguez, A.J., Pueyo, E., Martín-Álvarez, P.J., Moreno-Arribas, M.V., 2008. Potential of phenolic compounds forcontrolling lactic acid bacteria growth in wine. Food Control 19, 835e841.

García-Ruiz, A., Cueva, C., González-Rompinelli, E.M., Yuste, M., Torres, M., Martín-Álvarez, P.J., Bartolomé, B., Moreno-Arribas, M.V., 2012. Antimicrobial phenolicextracts able to inhibit lactic acid bacteria growth and wine malolacticfermentation. Food Control 28, 212e219.

García-Ruiz, A., Rodríguez-Bencomo, J.J., Garrido, I., Martín-Álvarez, P.J., Moreno-Arribas, M.V., Bartolomé, B., 2013. Assessment of the impact of the addition ofantimicrobial plant extracts to wine. Volatile and phenolic composition. Journalof the Science of Food and Agriculture 93, 2507e2516.

Gindreau, E., Torlois, S., Lonvaud-Funel, A., 1997. Identification and sequence anal-ysis of the region encoding the site-specific integration system from Leuco-nostoc oenos (Oenococcus oeni) temperate bacteriophage B10MC. FEMSMicrobiology Letters 147, 279e285.

Lonvaud-Funel, A., 1999. Lactic acid bacteria in the quality improvement anddepreciation of wine. Antonie van Leeuwenhoek, International Journal ofGeneral and Molecular Microbiology 76, 317e331.

López, I., Tenorio, C., Zarazaga, M., Dizy, M., Torres, C., Ruiz-Larrea, F., 2007. Evidenceof mixed wild populations of Oenococcus oeni strains during wine spontaneousmalolactic fermentations. European Food Research and Technology 226, 215e223.

López, I., Torres, C., Ruiz-Larrea, F., 2008. Genetic typification by pulsed-field gelelectrophoresis (PFGE) and randomly amplified polymorphic DNA (RAPD) ofwild Lactobacillus plantarum and Oenococcus oeni wine strains. European FoodResearch and Technology 227, 547e555.

Moreno-Arribas, M.V., Polo, M.C., 2005. Winemaking biochemistry and microbi-ology: current knowledge and future trends. Critical Reviews in Food Scienceand Nutrition 45, 265e286.

Pozo-Bayón, M.Á., Pardo, I., Ferrer, S., Moreno-Arribas, M.V., 2009. Molecular ap-proaches for the identification and characterisation of oenological lactic acidbacteria. African Journal of Biotechnology 8, 3995e4001.

Reguant, C., Bordons, A., 2003. Typification of Oenococcus oeni strains by multiplexRAPD-PCR and study of population dynamics during malolactic fermentation.Journal of Applied Microbiology 95, 344e353.

Renouf, V., Claisse, O., Lonvaud-Funel, A., 2007. Inventory and monitoring of winemicrobial consortia. Applied Microbiology and Biotechnology 75, 149e164.

Renouf, V., Claisse, O., Miot-Sertier, C., Lonvaud-Funel, A., 2006. Lactic acid bacteriaevolution during winemaking: use of rpoB gene as a target for PCR-DGGEanalysis. Food Microbiology 23, 136e145.

Renouf, V., Delaherche, A., Claisse, O., Lonvaud-Funel, A., 2008. Correlation betweenindigenous Oenococcus oeni strain resistance and the presence of geneticmarkers. Journal of Industrial Microbiology and Biotechnology 35, 27e33.

Renouf, V., Gindreau, E., Claisse, O., Lonvaud-Funel, A., 2005. Microbial changesduring malolactic fermentation in red wine elaboration. Journal Internationaldes Sciences de la Vigne et du Vin 39, 179e190.

Renouf, V., Vayssieres, L.C., Claisse, O., Lonvaud-Funel, A., 2009. Genetic andphenotypic evidence for two groups of Oenococcus oeni strains and theirprevalence during winemaking. Applied Microbiology and Biotechnology X,1e13.

Rodtong, S., Tannock, G.W., 1993. Differentiation of Lactobacillus strains by ribo-typing. Applied and Environmental Microbiology 59, 3480e3484.

Ruiz, P., Seseña, S., Izquierdo, P.M., Palop, M.L., 2010. Bacterial biodiversity anddynamics during malolactic fermentation of Tempranillo wines as determinedby a culture-independent method (PCR-DGGE). Applied Microbiology andBiotechnology 86, 1555e1562.

Sanguinetti, C.J., Neto, E.D., Simpson, A.J.G., 1994. Rapid silver staining and recoveryof PCR products separated on polyacrylamide gels. Biotechniques 17, 914e921.

Spano, G., Lonvaud-Funel, A., Claisse, O., Massa, S., 2007. In vivo PCR-DGGE analysisof Lactobacillus plantarum and Oenococcus oeni populations in red wine. CurrentMicrobiology 54, 9e13.

van Vuuren, H.J.J., Dicks, L.M.T., 1993. Leuconostoc oenos: a review. AmericanJournal of Enology and Viticulture 44, 99e112.