Tina wooden vat biofilm: A safe and highly efficient lactic acid bacteria delivering system in PDO Ragusano cheese making

International Journal of Food Microbiology 132 (2009) 1–8

Contents lists available at ScienceDirect

International Journal of Food Microbiology

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

Tina wooden vat biofilm: A safe and highly efficient lactic acid bacteria deliveringsystem in PDO Ragusano cheese making

Sylvie Lortal a,c,⁎, Annalisa Di Blasi b, Marie-Noëlle Madec a,c, Concetta Pediliggieri b, Laura Tuminello b,Gaëlle Tanguy a,c, Jacques Fauquant a,c, Yohan Lecuona a,c, Patrizia Campo b,Stefania Carpino b, Giuseppe Licitra b

a INRA, UMR1253, Science et Technologie du Lait et de l'Oeuf, F-35 Rennes, Franceb CoRFiLaC, 25 km 5, S.P. Ragusa Mare, 97100 Ragusa, Sicily, Italyc AGROCAMPUS OUEST, UMR1253, Science et Technologie du Lait et de l'Oeuf, F-35 Rennes, France

⁎ Corresponding author. INRA-AgroCampus Ouest, 6Rennes Cedex, France. Tel.: +33 2 23 48 53 21.

E-mail address: [email protected] (S. Lortal

0168-1605/$ – see front matter © 2009 Elsevier B.V. Aldoi:10.1016/j.ijfoodmicro.2009.02.026

a b s t r a c t

Article history:

Keywords:Wooden vatCheeseBiofilmRagusano cheeseLactic acid bacteriaFood safety

In the Sicilian PDO Ragusanadded, lactic acid is produceto assess the safety and efficounts varied from 103 tocounts of yeasts and mouldO157:H7 were totally abseenrichment, highlighting thobserved by confocal and sentirely the surface of the wwere predominant in theobserved. Fifty litres of micr

o cheese making, raw milk is placed in a wooden vat called a Tina. As no starter isd by milk flora and flora released from the Tina biofilm. The aim of this work wasciency of this natural inoculation system. From 15 Tinas' biofilms, bacteria total106 CFU/cm2, with the predominance of thermophilic lactic acid bacteria. Lows were found in a few Tinas. Salmonella, Listeria monocytogenes, Escherichia colint, as assessed by conventional plating and the Bax detection system aftere safety of the system. From four Tinas out of the 15, micropieces of wood werecanning electron microscopy. The biofilm entrapped in a matrix covered almostood. Polysaccharides were detected in the four Tinas. In three of the latter, cocciecosystem whereas in the other one, cocci, bacilli, yeasts and moulds wereofiltrated milk (b10 CFU/mL) were poured in the four Tinas for 10 min of contact.

Enumeration of lactic acid bacteria, yeasts and enterococci were performed in the milk after contact.Depending on the Tina, from 5·104 to 106 CFU/mL of Streptococcus thermophilus were released into the milk,and from 104 to 105 CFU/mL of thermophilic lactobacilli. Spontaneous acidification after contact confirmedthe high efficiency of biofilm lactic acid bacteria delivery.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

Ragusano cheese is a Protected Denomination of Origin (PDO)cheese made in the Hyblean region of Sicily from raw milk usingtraditional wood tools, without commercial starters. In the manu-facture of this brine-salted pasta filata cheese, raw milk is directlyplaced in the traditional wooden vat (Tina) for cheese making andlactic acid is produced by natural milk flora and desirable flora fromthe biofilm of the surface of the Tina. The respective contributions ofthese two bacterial sources, raw milk and Tina biofilm, as well as theecosystem in the biofilm are still not completely elucidated. Recently,molecular exploration of Tinas coming from 5 different farms of theHyblean region demonstrated the predominance of lactic acid bacteriain the biofilm, and in particular S. thermophilus, the presence ofthermophilic lactobacilli, lactococci and of a few high GC% micro-organisms like coryneform bacteria (Licitra et al., 2007). Somevariability was shown among Tinas (2 to 10 co-dominant species in

5, rue de St-Brieuc, F-35042

).

l rights reserved.

the PCR-Temperature Temporal Gradient Electrophoresis (TTGE)profiles), which represents a valuable source of biodiversity; more-over, each molecular profile was shown to be stable in time at least forthe dominant species. Raw milk spontaneous acidification before andafter a fewminutes of contact with the Tinas was accelerated in 80% ofthe cases clearly supporting the idea of lactic acid bacteria inoculationvia the Tina biofilms (Licitra et al., 2007). Unfortunately, the initialvariable level of raw milk microflora did not allow quantitative andqualitative assessment of the efficiency of this natural inoculatingsystem.

Wood is a traditional and natural material used in cheese produc-tion (vat and shelves). European rule discussions highlight regularlythe question of food safety of this material. The French food safetyauthority (Afssa, Saisine n°2007-SA-0206) reaffirmed recently afavourable judgment relative to the use of wooden tools in traditionalcheese making, but strongly suggested increasing data about theirspecific roles. Indeed, little scientific work has been carried out on themicroflora from this environment, its efficiency in inoculation and theproper traditional way of cleaning for guaranteeing safety of thisnatural system (Richard, 1987; Mariani et al., 2007). The objectives ofthis work were to further describe the Tina biofilm composition and

Table 1Enumeration of biofilms from 16 Tinas, 15 Tinas daily used in the farms of Hybleanregion and one used from time to time in the pilot plant of the CoRFiLaC.

Tinas Total count CFU/mL Yeasts CFU/mL Moulds cfu/mL

Sp 2.3·104 260 10Ia 7.8·105 270 2.8·103

Sam 4.5·106 180 40LisG 1.4·106 9.2·104 2·103

Lau 7.5·104 610 b10TumSal 2.3·105 5.8·104 5.9·102

FloSal 6.2·106 7.5·103 10Scro 1.2·104 20 b10BatGio 1.9·106 10 2.7·102

Cas 2.45·106 1.3·103 60BatAlt 2.2·106 b10 20Cris 2.5·105 b10 b10LisS 8.9·106 2·105 3·103

Occ 3.8·106 380 10Gua 9.6·105 7.2·103 40

CoRFiLaC 2.4–2.7·102 b10 b10

Results expressed as CFU/mL (5×100 cm2 of surface collected in 40 mL of peptonewater).

2 S. Lortal et al. / International Journal of Food Microbiology 132 (2009) 1–8

organisation, to assess its safety by systematic pathogen enumera-tions, to question its way of cleaning andmaintenance and to quantify,using microfiltrated milk, the release of microflora, in particular lacticacid bacteria, from the Tina.

2. Material and methods

2.1. Origin of the Tinas

The 15 Tinas analysed in the present work were used daily for milkprocessing and cheese production in the Hyblean region. Theyrepresent 37% of the Tina existing in Hyblean region. The wood usedis Douglas-fir, genus Pseudotsuga. The height of this slightly conical vatis between 75 and 105 cm, the diameter between 54 and 81 cm, andthe capacity between 130 and 320 L (Fig. 1A). Sampling of biofilm,enumerations and molecular profiles were performed on the 15 Tinas.The biofilms of four of them (farms Sp, Ia, Sa, LisG) were then furtherexplored by confocal microscopy, SEM, lactic acid bacteria enumera-tion and for their ability to release lactic acid bacteria in microfiltratedmilk.

2.2. Sampling of biofilms from the Tinas

The Tina biofilm samples were collected from the inner surface ofthe wooden vat. Cells from the biofilmwere removed from 5 differentareas of 100 cm2, using sterile swabs suspended in peptone water(OXOID, Basingstoke, Hampshire, UK) and then stored at −60 °C. Thesampling area of 100 cm2, was delimited by a paper square previouslysterilized under UV in a laminar air flow hood (Fig. 1B); one area wastaken from the bottom and four from the side.

2.3. Enumerations: pathogens and other flora

Enumerations of pathogens were done on biofilm samples freshlycollected, by using the BAX® System (DuPont Qualicon, UK) afterenrichment and results were confirmed by classical enumerations.Enrichment for Salmonella was carried out using modified tryptonesoya broth (Oxoid) at 35 °C for at least 20 h; 10 μL were thentransferred in 500 μL BHI (Brain Heart Infusion, Oxoid) for a furtherincubation for 3 h at 37 °C. Lysis was performed following the BAX®System instructions. Enrichment for Listeria was done in Listeriaselective enrichment broth base (Oxoid), at 30 °C, for 24 h. Avolume of0.1 mL of enriched broth was transferred in 9.9 mL of MoPS-BLED(Mops Bled Base: 3-(N-Morpholino) propanesulfonic acid-bufferedListeria enrichment broth base, DuPont Qualicon, UK) and incubated

Fig. 1. Tina wood vat characteristics (A) and biofilm sampling (B) (surface of the bluesquare, 10 cm×10 cm, five times per Tina).

at 35 °C during at least 18 h. Enrichment for Escherichia coli O157:H7was performed by incubating swabs in 10 mL of EC and novobiocinbroth (E. coli broth (Reduced Bile Salt) containing 20 mg/Lnovobiocin, Oxoid), at 35 °C for 20 h. Total counts and enumerationsof lactic acid bacteria were done as follows: for total count on PlateCount Agar at 30 °C for 3 days, for S. thermophilus using M17 agarplates (Difco, Sparks, MD; Terzaghi and Sandine, 1975), at 43 °C, inaerobiosis for 24 h; for thermophilic lactobacilli on MRS agar (Difco;De Man et al., 1960) at pH 5.4, 43 °C, 48 h, in anaerobiosis; forlactococci, M17 agar (Oxoid) at 15 °C for 10 days in anaerobiosis; formesophilic lactobacilli on MRS agar (Oxoid) at 15 °C, for 10 days inanaerobiosis; for enterococci KAA (Kanamycin Aesculin Azide Agarbase, Oxoid at 37 °C for 24 h and for yeasts on OGYE (Oxytetracycline-Glucose Yeast Extract Agar, Oxoid) at 22 °C for 5 days. All cultures wereperformed in duplicate.

2.4. PCR amplification and TTGE migration parameters

DNA was extracted from biofilm samples freshly collected asdescribed by Parayre et al. (2007). Universal primers (V3 region of the16S rRNA gene) and PCR amplification program were used asdescribed in Licitra et al. (2007). For TTGE analysis, the Dcodeuniversal mutation detection system (BIO-RAD Laboratories, Hercules,CA, USA) was used to separate the V3 region PCR products. Migrationwas performed at 41 V for 16 hwith a temperature increase from 63 °Cto 70 °C (rate of 0.4 °C/h). Gels were stained with ethidium bromide(Bio-Rad, Hercules, CA, USA) (0.6 μg ethidium bromide per mL of1.25× TAE (Tris Acetic acid EDTA, Bio-Rad) buffer) for 15 min, thenrinsed for 15 min in distilled water and photographed on a UVtransillumination table (E-Box 1000/26M, Euroclone, Siziano, Italy).The bands were analysed using the database developed by Parayreet al. (2007) for presumptive species identification.

2.5. Scanning electron microscopy

Thewood splinters from four Tinas (farms Sp, Ia, Sa and LisG) weresampled prior to the other microbiological samples. A little mark wasmade on three opposite sides of the internal surface of the Tina, atabout 20 cm from the bottom. By using a sterile blade, three smallwood splinters were removed (15×5×1 mm about), each onecontaining one mark on its marginal part (Fig. 1). The samples weredehydrated (Mallia et al., 2005), in a graded ethanol series (75, 85, 95and 100%) at room temperature for 12 h in each bath, and dried by thecritical point method in CO2 using a Polaron CPD7501 instrument(Polaron, Watford, UK). The dried samples were cut with a sterile

3S. Lortal et al. / International Journal of Food Microbiology 132 (2009) 1–8

blade into smaller fragments and each one was mounted on a SEMaluminium stub, with the surface covered by the biofilm faced up. Thesamples were gold-palladium coated in a Polaron SC7620mini sputtercoater, for 60 s. Two samples, for each farm, were analyzed by thescanning electron microscope Jeol JSM 5900LV (Jeol, Tokyo, Japan),operating at an accelerating voltage of 15 kV, a working distance of 9–11mm and amagnification of 200×, 500×, 2000× and 5000×. In orderto obtain a significant and complete representation of the biofilmcovering the internal surface of the Tina, the entire surface of eachsample was scanned and many images (about 65 images for eachsample) were recorded and stored in the form of Windows TIF imagefiles.

Fig. 2. Enumeration of lactic acid bacteria, enterococci and yeasts in the biofilm of the Tinasindependent experiments (week 2 andweek 9 ). Sampling of the biofilmwas carried outhe microfiltrated milk was below 10 CFU/mL. TC: Total count; St: Streptococcus thermopenterococci and Y: yeasts.

2.6. Confocal microscopy

On samples of wood collected as described in the previousparagraph, 5 μL of syto9 stain (Live-dead Invitrogen L13152, Cergy-Pontoise, France) was added directly to the wood for micro-organismdetection or 5 μL of concanavalin A linked with alexa fluor 633(Invitrogen 21402) for polysaccharide detection. The samples wereincubated for 15 min in the dark and observed using a confocalMicroscope NIKON C1-Si (Champigny-sur-Marne, France) at differenthighlightings (objectives 20×, 40× and 60×) with an excitationwavelength of 488 nm or 638 nm, and an emission wavelength rangefrom 500 to 530 nm or over 650 nm, respectively.

Sp, Sa, Ia and LisG and in the microfiltrated milk after contact with these Tinas, on twot on 5 surfaces (100 cm2 each) in 40mL peptonewater. Before contact, the total count ofhilus; ThL: thermophilic lactobacilli; MeL: mesophilic lactobacilli; Lc: lactococci; Ent:

4 S. Lortal et al. / International Journal of Food Microbiology 132 (2009) 1–8

2.7. Microfiltration of raw milk and experimental design

Six hundred litres of raw milk were skimmed at 40 °C using acream separator (Magnobosco, Italy, 70 L/h). The skimmed milk waskept at 40 °C and microfiltered at this temperature using a TIAmicrofiltration unit (TIA, Bollène, France) equipped with a sterilox GPmembrane (Pall Exekia, Saint Germain en Laye, France) of whichsurface and pore size were 0.24 m2 and 0.8 μm, respectively. The flowrate of microfiltered milk (MF milk) was 70 L/h whereas the one ofbacterial retentate was 4.5 L/h. Fifty litres of microfiltered milk werepoured in each Tina. For Sp Tina, 3 batches of 50 L were successivelyfilled. The entire experiment was repeated with a new batch of 600 Lof raw milk one month later.

2.8. Acidification

Microfiltrated milk (blank) and microfiltrated milk after 10 mincontact with the Tinas (farms Sp, Ia, Sa, and LisG) were frozen in liquidnitrogen, and stored at −60 °C. Then, they were thawed at roomtemperature, and submitted to spontaneous acidification in a waterbath at 37 °C. The acidification was monitored by pH decrease in

Fig. 3. Confocal microscopy observations of the Tina biofilm directly on a cut sample of woodecosystem composed of yeast and bacteria in the LisG Tina biofilm; C) cocci consortium in Sstaining) and highlighting of EPS producing cocci surrounded by a more intense green halo

duplicate. Two independent repetitions were done with one monthinterval; the curves presented are the average of two acidifications.

2.9. Temperature and pH measurement at the surface of the Tina

Thermotracks (Laboratoire Nephrotek, Rungis, France) were fixedin three different places at the surface of the Tina and the temperatureevolution during the Ragusano cheesemaking process wasmonitored.The surface pH was measured on the bottom of the Tina and in tenplaces on the side using a surface electrode (HI 1413 B, Hanna,Tanneries, France).

2.10. Statistical analysis

The datawere analysed with a three-way ANOVA (4×2×2) for eachflora using the following design: variable Tina (4 levels: Sp, Sa, Ia andLisG), variable Week (2 levels: Wk2 or Wk9) and variable Sample (2levels: Biofilm andMilk). Amissing value (enterococci in LisG biofilm atweek 9) was predicted using the results from ANOVA performedwithout enterococci counts. A Principal Component Analysis (PCA) wasperformed without using the total count as a variable, and the

: A) distribution of the biofilm on the wood of the LisG Tina (syto9 staining); B) complexp Tina biofilm; D) presence of polysaccharides covering the wood in blue (concanavalinin the LisG Tina biofilm.

5S. Lortal et al. / International Journal of Food Microbiology 132 (2009) 1–8

individuals were plotted on the two first axes, which explain more than70% of the variability. All statistical analyses and graphic representationwere performed using R (R Development Core Team, 2008).

3. Results

3.1. Absence of pathogens in 15 Tina biofilm samples

The enumerations (total counts, yeasts and moulds) performed onbiofilm samples from 15 Tinas daily used in Hyblean farms for Ragusanocheesemaking are summarized inTable 1. The total count varied from104

to almost 107 CFU/mL, yeasts from 0 to 105 CFU/mL and four Tinascontained moulds (with a maximum of 103 CFU/mL), without anyobvious correlation between the presence of yeasts and the presence ofmoulds. The variability of composition of the Tina biofilms, that canexhibit, or not, significant amount of yeasts (5 log magnitude), andeventuallymore rarely somemoulds,was clearlyhighlighted fromTable1.LisG Tina exhibited the richest biofilms in terms of total count and yeast/mould levels. By contrast, the Tina from CoRFiLaC pilot plant showed avery low total count, which can be explained by the irregular use of thisTina. After enrichment from the 15 Tina biofilm samples, Salmonella,L. monocytogenes, E. coli O157:H7were not found using the BAX® System(data not shown) highlighting the complete absence of these pathogens.Staphylococcus aureus was assayed by conventional plating and wasabsent in 13 of the 15 Tinas. In Ia Tina,10 CFU/mL of S. aureuswere foundand 6·102 CFU/mL in Sp Tina. Coliforms were totally absent in 9 Tinas aswell asE. coli, very low in three Tinas (b70CFU/mL) and between103 and104 CFU/mL in three other ones (LisG, LisS and Occ).

3.2. Lactic acid bacteria and yeasts in biofilm samples

S. thermophilus, thermophilic lactobacilli, mesophilic lactobacilli,lactococci, enterococci and yeasts counts in the biofilm were

Fig. 4. SEM observations of the Tina biofilms: A) lactobacilli in LisG; B) complex ecosystemextracellular matrix in Ia (magnification: A) and C), 2000×; B) and D), 5000×).

determined for four Tinas (Sp, Ia, Sa and LisG) at two periods (week2 and week 9 of 2008). The results are presented in Fig. 2. Thepresence of high viable counts of S. thermophilus in all Tinas (PN0.1)and the presence of high counts of thermophilic lactobacilli too, inparticular in LisG Tina, were confirmed as well as the variability in thesubdominant species (mesophilic lactic acid bacteria, yeasts andenterococci). The presence of S. thermophilus as a predominantspecies was confirmed by PCR-TTGE profiles for the 15 Tinas (datanot shown).

3.3. Microstructure of the biofilms

Micropieces of wood coming from four selected Tinas wereobserved by confocal microscopy (Fig. 3) and by SEM (Fig. 4). Usingconfocal microscopy and bacterial specific staining, the biofilm wasshown to cover the wood almost entirely as illustrated on Fig. 3A. Acomplex ecosystem consortium (cocci, bacilli, yeasts andmoulds) wasobserved in the case of LisG Tina (Fig. 3B) and simpler ecosystemswith predominant cocci in Sp, Ia, and Sa Tina (Fig. 3C). The conca-navalin staining indicated the occurrence of a polysaccharide layer inthe biofilm (Fig. 3D).

Using SEM, the biofilmwas shown to cover the wood entirely withvery rare areas where the wood itself can be seen, in particular in LisG(data not shown). The presence of a thick extracellular matrix whichlikely corresponds to the exopolysaccharides detected by confocalmicroscopy was observed in the four Tinas (Fig. 4C and D). In Sp andSa, cocci were clearly predominant as observed by confocal micro-scopy; a few isolated rods can also be seen, and no or very few yeasts.Some lactate crystals were also noticed in Sp (data not shown). In Ia,cocci were clearly predominant but abundant yeasts were alsoobserved in the first replica (biofilm week 2) but not in the secondone (Ia biofilm week 9), which is consistent with the respective yeastenumeration (2 log less yeasts). In LisG, lactobacilli were predominant

composed of lactobacilli, cocci and yeasts in LisG; C) cocci in Sp; D) cocci in a thick

Fig. 5. Principal component analysis (PCA). Variables factor map (A) and individualsfactor map (B) on the first two dimensions. For each individual are mentioned: thename of the Tina, week of sampling and type of sample; e.g. “LisG2m” represents themilk from Tina LisG sampled on week 2.

Fig. 6. Spontaneous acidification at 37 °C of themicrofiltratedmilk before (blank,♦) andafter 10 min contact with Tina Ia (■), LisG (▲) and Sa (●); each curve represents theaverage of 2 repetitions from 2 independent acidifications.

6 S. Lortal et al. / International Journal of Food Microbiology 132 (2009) 1–8

over cocci, and many yeasts can be seen (Fig. 4A, B) which isconsistent with the yeast enumeration (N104 CFU/mL, Fig. 2D),confirming the more complex ecosystem in LisG already detected byconfocal microscopy (Fig. 3B).

3.4. Cleaning of the Tinas: summary of traditional practices

In order to better assess the way of cleaning and maintenance ofactive Tinas, a questionnaire was prepared for the farmers of theHyblean region. All the farmers clean the Tina daily after the cheesemaking. A majority of farmers (11/15) brush the Tina and clean withhot water (# 55 °C), a minority (4/11) with hot water only withoutbrushing. Then, all let the Tina dry upside down until the next cheesemaking the day after.

3.5. Temperature and pH in the biofilm

The registration by thermotracks fixed in the bottom of the Sp Tinawooden vat during all the cheese making process indicated thetemperature never exceeded 45 °C.

The surface pH just after washing the Tina was acid, between 4.5and 5.0; the lowest values were in the bottom. This low surface pH islikely related to the local lactic acid production of lactic acid bacteria.

3.6. Inoculation of milk

To assess the release of microflora from the biofilm of these fourTinas, raw milk was skimmed and microfiltrated (GP membrane0.8 μm). Fifty litres of microfiltratedmilk (b10 CFU/mL)was poured inthe Tinas and left in contact for 10 min. The whole experiment wascompletely repeated at week 2 and week 9. Enumeration of lactic acidbacteria, yeasts and enterococci was then performed in the biofilmand in the milk after contact and the results are reported in Fig. 2.More than 70% of the variability is explained by the first twocomponents of the PCA (Fig. 5A). A highly significant Tina effect wasobserved for enterococci and yeasts (Pb0.001). The Tina LisG, locatedat the right of Fig. 5B, differs from the other Tinas mainly because ofhigh counts for yeasts and mesophilic lactobacilli. Depending on theTina, from 5·104 to 106 CFU/mL of milk of S. thermophilus, and from104 to 105 CFU/mL of thermophilic lactobacilli were inoculated bycontact. As the microfiltrated milk was almost sterile (b10 CFU/mL)this high lactic acid bacteria level can only be attributed tospontaneous release from the Tina biofilms. The high level oflactobacilli for LisG was consistent with the high level found in thebiofilm. Yeasts were also released in good agreement with thepresence of yeasts in the biofilm itself. The similarity of the patternbetween biofilm and milk is shown in Fig. 5B, where the biofilm andmilk samples from the same Tina are clustered.

Spontaneous acidification of milks after contact confirmed thehigh efficiency of lactic acid bacteria release from the Tina biofilm. ThepH of the blank (microfiltrated milk, total count below 10 CFU/mL)remained unchanged during 18 h (Fig. 6). After contact with the Tinasa spontaneous strong acidification occurred, leading to a final pH ofabout 4. Significant differences in lag phases and rates of acidificationwere observed. The lag phase was 2 h, 3 h and 4 h for Sa, LisG, and Ia,respectively; the Vmax was−0.48 upH/h for LisG,−0.465 upH/h forSa and −0.42 upH/h for Ia. The final pH reached (above or below 4)was also different, all these differences reflecting the specificity andbiodiversity in lactic acid bacteria released by each Tina. For one Tina(Sp) the experiment was repeated three times successively in the

7S. Lortal et al. / International Journal of Food Microbiology 132 (2009) 1–8

same day with three different sets of microfiltrated milk. Interestingly,as shown by the counts of S. thermophilus, lactobacilli and yeasts aftercontact, the inoculation efficiency decreased as well as the acidifica-tion rate (Fig. 7). The lag phase increased from 2 h 40 min to 7 h. Theefficiency decreased mainly between the first and second times, butthen the release was stable. This might indicate a more labile layerthat can be detached easily during the first contact with milk.

4. Discussion

Wood is used in traditional cheese making as vat or as shelves, allaround the world, in particular in PDO cheese. As any surface incontact with nutrients such as milk or cheeses components, it can bemore or less rapidly covered by amicrobial biofilm. However, very fewreports exist in the literature regarding the organization and safety ofthis biofilm and its ability to release microflora either in milk or on thecheese surface directly in contact with the wood (Richard, 1987;Mariani et al., 2007). The presence of a biofilm containing lactic acidbacteria in the Tina wood vat used in PDO Ragusano cheese makingwas already assessed (Licitra et al., 1998, 2007). In the present work,the safety, composition, microstructure and role of this biofilm in rawmilk inoculationwere further explored for several Tinas. Tina woodenvats are made by two Sicilian producers from Douglas fir wood. Eachfarmer has his own Tina, which is used and cleaned daily afterRagusano cheese making. The high bacterial total counts found on thesurface confirmed the occurrence of a rich biofilm.

The complete absence of Salmonella, Listeria and E. coli O157:H7was demonstrated in these 15 Tinas, which represent 37% of the activeTinas in the Hyblean region, strongly reinforcing the idea of the safetyof this wood vat system. To explain the inability of pathogens toadhere or to survive in these biofilms, the main factors are the acidicconditions measured at the surface of the biofilm (pH below 5), thecompetition for nutrients, as well as the cooking temperature above

Fig. 7. Decrease in LAB inoculation (A) and spontaneous acidification rate (B) when 3

40 °C. However, the additional hypothesis of bacteriocin producersinside the biofilm cannot be excluded and should be explored.

The further characterization of four Tinas confirmed that thermo-philic lactic acid bacteria are the most abundant bacteria in thebiofilm, reaching values as high as 106 per cm2, S. thermophilus aloneor with thermophilic lactobacilli being predominant in enumerationsas previously indicated by molecular tools (Licitra et al., 2007). Thelevel of yeasts, moulds and enterococci was extremely variable fromone Tina biofilm to another, confirming the farm specificity of themicrobial profile. Confocal as well as scanning electron microscopyfurther highlighted this biodiversity by showing biofilms composedeither almost exclusively of cocci or of much more complexecosystems (cocci, bacilli, yeasts and moulds tightly bound together)like in LisG. This variability can be likely related to slight variations ofprocess parameters depending on the farms, like for exampletemperature of cooking or to different ways of washing.

By confocal as well as scanning electron microscopy, the biofilmwas shown to cover entirely the surface of the wood. An extracellularmatrix was clearly seen by both methods, and likely corresponded tobacterial exopolysaccharides by specific concanavalin staining andconfocal microscopy observations. The systematic presence of poly-saccharides in each Tina analysed led to questions about theirpotential role. Are they crucial in the initial steps of adhesion, helpingthe adherence of producing strains? Are polysaccharides protecting orhelping the survival of the bacteria inside the biofilm, by bindingwater or catching somemilk/whey components by ionic interactions?The identification of the producing strains as well as the biochemicalcharacterization of these polysaccharides would help in answeringthese questions.

Using microfiltrated milk, almost devoid of any microbial cells, theability of Tina biofilm to release microflora into the milk has beenassessed and quantified here for the first time. After a few minutes ofcontact the Tina biofilms can release into the microfiltrated milk 6 log

successive sets (A, B, C) of MF milk in the same day were placed in the Sp Tina.

8 S. Lortal et al. / International Journal of Food Microbiology 132 (2009) 1–8

of lactic acid bacteria per mL, which revealed a spectacular efficiencyof milk inoculation. The rate of spontaneous acidification generated bythis inoculation was again “Tina dependant”, reflecting the biodiver-sity of these biofilms. The lactic acid bacteria released by the Tinabiofilm can regulate the variability of the lactic flora coming from theraw milk, making the day to day repeatability of the milk acidificationeasier to achieve. As the Tina biofilm contains also a variety of nonlactic species, the subsequent role, if they have any, of all thesebacteria in generating variability of flavour compounds during thecheese ripening should now be investigated.

In conclusion, the Tina wooden vat is a safe and highly efficientsystem to deliver thermophilic lactic acid bacteria into the raw milk,contributing to the acidification step. As the biofilm contains, inaddition to the predominant S. thermophilus, many other species, likeother lactic or other bacteria including high GC% (Licitra et al., 2007),yeasts and enterococci (in this work), Tina wooden vat biofilm likelycontributes to the biodiversity of the Ragusano cheese ecosystem.

Acknowledgements

The authors gratefully acknowledge the Assessorato IndustrieSicilian Region and the Ministero dell'Istruzione dell'Universita e dellaRicerca (MIUR) for funding this project in the “Accordo ProgrammaQuadro” action. The authors are grateful to Anne Thierry and PascalPachot for the statistical analysis of the data.

References

De Man, J.C., Rogosa, M., Sharpe, M.E., 1960. A medium for the cultivation of lactobacilli.Journal of Applied Bacteriology 23, 130–135.

Licitra, G., Portelli, G., Campo, P., Longombardo, G., Farina, G., Carpino, S., Barbano, D.,1998. Technology to produce Ragusano cheese: a survey. Journal of Dairy Science81, 3343–3349.

Licitra, G., Ogier, J.C., Parayre, S., Pediliggieri, C., Carnemolla, T.M., Falentin, H., Madec, M.N.,Carpino, S., Lortal, S., 2007. Variability of the bacterial biofilms of the “tina”woodvat usedin the Ragusano cheese making process. Applied and Environmental Microbiology 73,6980–6987.

Mallia, S., Carpino, S., Corralo, L., Tuminello, L., Gelsonimo, R., Licitra, G., 2005. Effects onaroma profile of Piacentinu and Ricotta cheese using different tools materialsduring cheese making. In: Spanier, A.M., Shahidi, F., Parliment, T.H., Mussinan, C.,Ho, C.-T., Tratras Contis, E. (Eds.), Food Flavor and Chemistry: Explorations into the21st Century. Royal Society of Chemistry, Cambridge, UK, pp. 23–34.

Mariani, C., Briandet, R., Chamba, J.F., Notz, E., Carnet-Pantiez, A., Eyoug, R.N., Oulahal, N., 2007.Biofilm ecology of wooden shelves used in ripening the French raw milk smear cheesereblochon de Savoie. Journal of Dairy Science 90, 1653–1661.

Parayre, S., Falentin, H., Madec, M.N., Sivieri, K., Le Dizes, A.S., Sohier, D., Lortal, S., 2007.Easy DNA extraction method and optimisation of PCR-Temporal Temperature GelElectrophoresis to identify the predominant high and lowGC-content bacteria fromdairy products. Journal of Microbiological Methods 69, 431–441.

R Development Core Team, 2008. R: a language and environment for statistical computing. RFoundation for Statistical Computing, Vienna, Austria. http://www.R-project.org.

Richard, J., 1987. Utilisation du bois comme matériau au contact des produits laitiers.Comptes Rendus de l'Académie d'Agriculture de France 83 (5), 27–34.

Terzaghi, B.E., Sandine, W.E., 1975. Improved medium for lactic streptococci and theirbacteriophages. Applied and Environmental Microbiology 29, 807–813.