Lait (1999) 79, 201-209© InralElsevier, Paris

201

Original article

Propionibacteria flora in Swiss raw milk fromlowlands and Alps

Denise Fessler-", Michael G. Casey'?", Zdenko Puhan''

a Laboratory of Dairy Science, Swiss Federal Institute of Technology, Zurich, Switzerlandb Federal Dairy Research Institute, 3003 Bem, Switzerland

(Received 22 September 1997; accepted 8 September 1998)

Abstract - The propionic acid bacterial tlora in Swiss raw milk was investigated. Four hundred andfifty-three strains from lowland milk and 21 strains from Alpine raw milk were c1assified to thespecies level by protein profile analysis and restriction analysis of the 23S ribosomal ribonucleicacid (rRNA). Plasmid profiles and random amplified polymorphie deoxyribonucleic acid (RAPD) wasused to differentiate the strains. The tlora was found to be extremely diverse. Ali four dairy Propi-onibacterium species were found in lowland raw milk: 71 % were P.freudenreichii, 19 % P.jensenii,8 % P. acidipropionici and 2 % P. thoenii. P. acidipropionici was not found in alpine milk butP. freudenreichii made up 55 %, P. jensenii 15 % and P. thoenii 30 % of the total. Among 278P.freudenreichii strains, 219 (79 %) different strains were identified by RAPD to the strain level. Forthe other species strain diversity was even greater. Only 30 % of the strains analyzed carried plasmids.The Swiss chee se industry has a large reservoir of Propionibacterium strains in raw milk for futureapplications and developments. © InralElsevier, Paris

propionibacterium / milk / electrophoresis / PCR

Résumé - La flore propionique dans le lait cru suisse en provenance de la plaine et des alpages.La flore propionique dans le lait cru suisse a été analysée. L'analyse des profils protéiques et l'ana-lyse de restriction de l'ARNr 235 ont été faites avec 453 souches bactériennes du lait en provenancede la plaine et 21 souches du lait d'alpages, en vue de les classer en espèces. Leur contenu en plas-mides et la méthode de l'amplification génique au hasard de l'ADN ont été utilisés pour la diffé-renciation des souches. La tlore propionique dans le lait cru suisse s'est avérée être extrêmement variée.Les quatre espèces laitières de Propionibacterium ont été trouvées dans le lait cru de plaine; 71 %étaient des P. f reudenreichii, 19 % des P. jensenii, 8 % des P. acidipropionici et 2 % des P. thoenii.Dans le lait cru d'alpage P. acidipropionici n'a pas été trouvé, P. freudenreichii formait 55 %, P.jen-senii 15 % et P. thoenii 30 % des bactéries propioniques. Parmi les 278 souches de P.freudenreichiiisolées 219 (79 %) souches différentes ont été identifiées par RAPD. La diversité chez les autres

* Correspondence and reprints. michaeI.casey@mbox.fam.admin.ch

202 D. Fessier et al.

espèces était encore plus grande. Seulement 30 % de toutes les souches analysées possédaient des plas-mides. L'industrie suisse du fromage dispose grâce au lait cru d'un reservoir important de souches dePropionibacterium en vue d'applications et développements futurs. © InralElsevier, Paris

propionibacterium liait 1électrophorèse 1PCR

1. INTRODUCTION

Propionibacteria are used in the manu-facture of Emmental (or Swiss-type) chee seto produce the characteristic eyes and thenutty flavour [18, 19]. Orla-Jensen [23] firstisolated propionibacteria from milk anddescribed them in detail.

Propionibacteria are gram-positive, non-motile, asporogenous, anaerobie to aerotol-erant and generally catalase-positive pleo-morphic rods [Il]. The genus Propioni-bacterium is separated into eight species.Four have been isolated from human skinand are generally not found in milk. Theother four species, the so-called 'dairypropionibacteria', are P. freudenreichii,P. jensenii, P. thoenii and P. acidipropi-onici. Few investigations exist on the nat-ural habitat of dairy propionibacteria. Alispecies have been isolated from soil andsilage [5], sorne species from fodder anddung [21], and anaerobie fermenters [29].Plastourgos and Vaughn [27] isolated P. aci-dipropionici and P. jensenii strains fromspoiled olives. In the rumen, Propionibac-terium sp. are, together with other species,responsible for the breakdown of urea andthe release of ammonia [36].

Propionibacteria were enumerated insamples of raw milk from different areas ofFrance over a period of lyear but the specieswere not identified [34]. Carcano et al. [6]analyzed 306 samples of milk used forGrana chee se production for the presenceof propionibacteria. Sixt y strains were iso-lated and the four different species identi-fied. Until now only a limited study on theanalysis and classification of the propioni-

bacterial flora in Swiss raw milk has beenperformed [3]. The wild propionibacterialflora may influence cheese quality and beresponsible for splitting [15, 25] and brownspot defects [4]. It may also serve as a reser-voir for future applications in the cheeseindustry.

The differentiation between the four dairyspecies, especially between P. jensenii andP. thoenii, using classical biochemical meth-ods is problematic. New methods, whichallow a better classification, have been pro-posed [1-3,12,14,30,31].

The purpose of this study was to screen alarge number of strains, classify them byprotein analysis into species and to differ-entiate them by random amplified poly-morphic deoxyribonucleic acid (RA PD) atthe strain level in order to obtain informationconcerning the diversity of the propioni-bacterial flora in Swiss raw milk. Milks fromthe lowlands in different areas of Switzer-land and a small number of milks fromalpine regions were studied, since approxi-mately 20 % of cows sojoum to the Alps insummertime.

2. MATERIALS AND METHODS

2.1. PropionibacteriaI strainsand culture conditions

Four hundred and fifty-three strains from358 milks, provided by nine laboratories of theDairy Inspection and Consulting Services, wereisolated by plating on YELA (24 mL·L-1 sodiumlactate (50 % v/v), 30 g·L-1 casein peptone,30 g·L-1 yeast extract and 15 g-L-1 agar, pH 6.8)and incubated for 10 d at 30 oc. The nine labo-ratories were situated in different regions of

Propionibacterial flora in Swiss raw milk

Switzerland: in the west, Vaud-Geneva/Neuchâ-tel (VDINE), Fribourg (FR) and Bern (BE); inthe north, Nordwestschweiz (NW), Aargau (AG),Nordostschweiz (NO), Thurgau (TG) andSaint-Galien/Appenzell Ausserrhoden (SG/AR);in the centre, Zentralschweiz (ZS). Preliminaryidentification was determined by noting thecolour of each colony, which ranged from beigeto red to brown, the morphology under the micro-scope and the presence of catalase tested with3 % (v/v) HzÛ2' Another 21 strains were isolatedfrom alpine milk from three different areas ofCanton Fribourg, where Gruyere cheese is pro-duced. Type strains were obtained from theAmerican Type Culture Collection (ATCC;Rockville, MD, USA) and Deutsche Sammlungvon Mikroorganismen und Zellkulturen GmbH(DSM).

Ali strains were purified by re-streaking threetimes on YELA for 10 d at 30 "C under anaero-bic conditions. Strains were subsequently grownon YEL broth for 5 to 7 d at 30 oc.

2.2. Analysis of the solublecell free protein extracts

Ten mL of a broth culture were centrifuged at10 000 g for 10 min at 4 "C and the cells sus-pended in 150 ~L TES buffer(1O mrnol-L:' Tris-CI, pH 7.5, 1 mmol-L"' EDTA, 100 mmol-L:'NaCI). Approximately 300 mg glass beads of100 urn diameter (BioSpec Products) were addedand the suspension was vortexed five times atI-min intervals with a cooling period of 1 minon ice between each vortexing period. Glassbeads and intact bacterial cells were removed bycentrifugation at 450 g for 10 min. The proteinconcentration in the supernatant was deterrninedaccording to the method of Lowry et al. [20].

Polyacrylamide gel electrophoresis was per-forrned according to Laemmli [17] as modifiedby Baer [1]. The gels were stained according toMorrissey [22], photographed with a Sony CCDcamera and the patterns analyzed with GelCom-par (Applied Maths, Kortrijk, Belgium). Classi-fication of profiles and their reproducibility wereverified using the unweighted pair group methodusing arithmetic averages [33].

2.3. Preparation of crude DNAextracts for species identification

The preparation of crude DNA was perforrnedas described by FessIer et al. [14].

203

2.4. Purification of genomic DNAfor strain identification

Cultures (10 mL) were harvested at 10 000 gfor 10 min and washed in Iysis buffer(100 rnmol-L:' sodium borate, 10 mmolL:!EDTA, 25 % saccharose, pH 8.0). One mL ofIysis buffer containing 20 mg lysozyme wasadded, thoroughly mixed and incubated for 1 h at37 oc. The protoplasts were recovered by cen-trifugation at 4 000 g for 10 min at 4 "C, sus-pended in 4.75 mL 20 rnmol-L:' EDTA, pH 7.0and Iysed by the addition of 0.25 mL of 20 %sodium dodecyl sulphate (SDS). Protein wasremoved by the addition of 1.25 mL 5.0 mol-L:'sodium perchlorate and 3.0 mL chloroforrn fol-lowed by incubation for 15 min at room temper-ature and centrifugation for 10 min at 10 000 g.DNA was precipitated by the addition of 3 mL ofpropanol to 5 mL of the supematant, recovered bycentrifugation for 10 min at 10 000 g, dissolvedin TE buffer (10 mmol-L:' Tris-CI, 1 mmol-Lr!EDT A, pH 8.0) and the concentration adjusted to100 ug-ml, -1.

2.5. Polymerase chain reaction(PCR) and restriction analyses

Analysis of the insertion region within the23S ribosomal ribonucleic acid (rRNA) gene[32] was performed according to FessIer et al.[14].

2.6. Differentiation of P. freudenreichiisubsp. on minimum growth medium

Minimum growth medium (MGM) was amodification proposed by Baer (personal com-munication) of that of Crow [10] and includedeither 14 mmol-L:' lactose or 100 mmol-Lr!sodium lactate.

2.7. RAPD with primer SK2

Taq polymerase, PCR buffer and dNTP mixwere purchased from Boehringer Mannheim andthe primer SK2 (S'-GCC GCC GCC GCC-3')[12] from Microsynth (Balgach, Switzerland).The PCR-reaction mix contained 83.5 ~L steriledistilled water, 10 ~L PCR buffer, 2 ~L dNTPmix, 2 ~L of primer SK2 (100 nmol-mlr '), 0.5 ut,Taq polymerase (5 Ll-mlr") and 2 ~L purifiedDNA.

204 D. FessIer et al.

PCR amplification was performed on a ther-mocycler (Inotech, Dottikon, Switzerland) usingthe following conditions: an initial DNA denat-uration step at 94 "C for 4 min followed by4S cycles of denaturation at 94 "C for 1 min,primer annealing at 43 "C for 1 min, and primerextension at 72 "C for 2 min; a final extensionwas carried out at 72 "C for 10 min.

2.8. RAPD with primer DF4

RA PD analysis beads were purchased fromPharmacia Biotech (Uppsala, Sweden) and theprimer DF4 (S' -CGC CGC CGT CGC-3') fromMWG Biotech (Ebersberg, Germany). ThePCR-reaction mix contained one analysis bead,24 ul, sterile distilled water, 0.6 ul, of primerDF4 (SOnmol-ml..") and 004 ul, purified DNA.PCR amplification was performed under the sameconditions as for primer SK2.

PCR amplification products of both primerswere separated on a 4 % agarose gel (NuSieve3:1 agarose, FMC BioProducts) and visualizedwith ethidium bromide 1 rng-Lr! in 10 mmol-Lr!sodium borate, pH 8.0 under UV light. The DNAstandard was a S0-2 000 bp ladder (Bio-Rad)with bands of2 000,1 SOO,1 000,700, SOO,400,300, 200, 100 and SObp. Photographs were takenwith a Sony CCD camera and the patterns ana-Iyzed with GelCompar. Classification of profilesand their reproducibility were verified using theunweighted pair group method using arithmeticaverages [33].

2.9. Plasmid analysis

Cul tures (10 mL) were harvested at 10 000 gfor 10 min and washed in Iysis buffer. One mL oflysis buffer containing 20 mg lysozyme and2 000 U mutanolysin was added, thoroughlymixed and incubated 2 h at 37 oc. The proto-plasts were recovered by centrifugation at 4 000 gfor 10 min at 4 "C and Iysed by the addition of1 mL buffer (SO rnmol-L -1 sodium borate,10 mmol-Lr! EDTA, 1 % SDS, pH 1204). O.SIlLS mol-L:' sodium acetate, pH 4.7 was added,incubated at 4 "C for 1 h, and centrifuged at 4 "Cfor 20 min at 10 000 g. The plasmid in the super-natant was further purified using the Plasmid-Safe kit (Epicentre Technologies, WI, USA)according to the manufacturers instructions.4S ul, of the solution were run overnight on an0.8 % agarose gel (Bio-Rad, 162-0126) at SOVin TBE buffer (90 rnmol-L:' Tris, 90 rnrnol-L'"borate, 2 mrnol-L'" EDTA, pH 8A-8.S). The

plasmids were visualized under UV light inthe presence of ethidium bromide buffer(10 mmol-L -1 sodium borate, pH 8.0 containing1 mg-L':' ethidium bromide). The standard wasthe supercoiled DNA ladder from Sigma (DS292)with bands of 16, 14, 12, 10,8,7,6, S, 4, 3 and2 kb. Photographs were taken with a PolaroidMP4 Land Camera.

3. RESULTS AND DISCUSSION

3.1. Propionibacterium speciesin lowland milk

Of the 453 strains isolated, 379 wereidentified as being propionibacteria: 267were identified as P. freudenreichii, 32 asP. acidipropionici, 72 as P. jensenii and 8 asP. thoenii. The identification was based onthe comparison of the protein profiles withthe type strains using the GelCompar soft-ware. With SDS-polyacrylamide gel elec-trophoresis (PAGE) reproducibilities of92-98 % [7, 8], 95.9 ± 1.7 % [9] and 93-97 %[35] were found for the identification of thespecies. It was considered that a strainbelonged to a particular species when therewas more than 70 % identity of the proteinprofile with the type strain. Species identi-fication was confirmed by PCR and restric-tion analysis of crude DNA extracts.

Figure J shows the protein profiles ofthe four Propionibacterium-type strains,which were used for comparison. The resultsof restriction analysis of the 23S rRNA geneare to be found in FessIer et al. [14].

With the restriction analysis method areproducibility of 100 % was achieved. Thismethod has the advantage that with propi-onibacteria only three or four bands wereobserved and the species were, therefore,easily recognized. Only 3 % of the strainstested were classified differently by 23SrRNA restriction analysis and protein profileanalysis. Restriction enzyme analysis is eas-ier to evaluate and gave better repro-ducibilities than protein profiling and is,therefore, the preferred method for speciesidentification.

Propionibacterial tlora in Swiss raw milk

Figure 1. Protein profiles of the type strains ofPropionibacterium.1 = P. acidipropionici ATCC 25562; 2 =P. thoenii ATCC 4874; 3 = P. freudenreichiiATCC 6207; 4 = P. jensenii ATCC 4868.Figure 1. Profiles protéiques de souches typesde Propionibacterium.1 = P. acidipropionici ATCC 25562 ; 2 =P. thoenii ATCC 4874 ; 3 = P. freudenreichiiATCC 6207 ; 4 = P. jensenii ATCC 4868.

Figure 2 shows the distribution of thestrains from the are as covered by the DairyInspection and Consulting Service labora-tories. The distribution of the four Propi-onibacterium species was not the same ineach region. However, definite conclusionscould not be drawn because the number of

205

strains isolated from sorne regions was rel-atively smal!. Nevertheless, P. freudenre-ichii and P. jensenii were encountered in ailregions. P. acidipropionici was not foundin milks from VDINE, NW, NO or SO/AR,and P. thoenii was absent in milks fromVDINE, TO and SOI AR. Only milk fromthree regions, FR, AO and ZS, harbouredail four species of propionibacteria.

The only previous study performed onthe population of propionibacteria in Swissraw milk [3] showed that P. freudenreichiiwas almost exciusively present, whereasP. acidipropionici and P. jensenii wererarely detected and P. thoenii not at al!.

3.2. Propionibacteriafrom alpine regions

Twenty-one strains were isolated from38 milks from three different alpine regions.The techniques used for identification werethe same as those used for strains isolatedfrom lowland milk.

One strain was not a PAB and Il wereP. freudenreichii, three P. jensenii and sixP. thoenii. P. acidipropionici was not found.Figure 3 shows the distribution of propi-onibacteria in alpine milk and in lowlandraw milk from the FR region. The Alpswhere the milk originated from are also sit-uated in this region. In both groups,P. freudenreichii was the predominantspecies, with 62 % in lowland raw milk and55 % in alpine milk. In lowland raw milk,21 % of the strains were P. jensenii and inalpine milk 15 %. Of the alpine milk strains30 % were classified as P. thoenii comparedto only 2 % in lowland milk. The large pro-portion of P. thoenii in alpine milk shouldbe confirmed before any definite conclu-sions may be drawn.

3.3. P. freudenreichii subspecies

The two subspecies P. freudenreichiisubsp. freudenreichii and P. freudenreichii

206 D. FessIer et al.

53

not PropionibacteriumP. freudenreichiiP. acidtproplonici

III P. ienseniiIII P. thoenii

Figure 2. Distribution of propionibacteria in lowland raw milk.AG = Aargau; BE = Bern; FR = Fribourg; NO = Nordostschweiz; NW = Nordwestschweiz;SG/AR = Saint-GalIen/Appenzell Ausserrhoden; TG = Thurgau; VD/NE = Vaud-GenevalNeuchâtel;ZS = Zentralschweiz.Figure 2. Distribution des bactéries propioniques dans le lait cru de la plaine.AG = Aargau; BE = Bern ; FR = Fribourg; NO = Nordost-Schweiz ; NW = Nordwest-Schweiz ;SG/AR = Saint-Galien/Appenzell Ausserrhoden ; TG = Thurgau ; VD/NE = Vaud/Neuchâtel;ZS = Zentralschweiz.

subsp. shennanii could not be distinguishedby the molecular methods used in this work.Protein profiles and restriction profiles withMsp 1 of the 23S rRNA gene did not show

70,----------------,60

~ 50

5 403f! 30'Jii5 20

10

P. freudenreichii P. jensenii P. acidipropionici P. thoenii

Figure 3. Distribution of propionibacteria inalpine and lowland raw milk from the Fribourg(FR) region.Figure 3. Distribution des bactéries propioniquesdans le lait cru de la plaine et d'alpages de laregion FR.

any difference. However, for the dairy indus-try the difference is important: P. freuden-reichii subsp.freudenreichii does not growon lactose, whereas P. freudenreichii subsp.shermanii does. In order to distinguishbetween the two subspecies, the utilizationof lactose by P. freudenreichii strains wasdetermined. Of the total 278 P. freudenre-ichii strains, 136 did not grow on lactoseand were classified as P. freudenreichiisubsp. freudenreichii. The remaining 142strains grew on lactose and were classifiedas P. freudenreichii subsp. shermanii.

The existence of the two P. freudenre-ichii subspecies is disputed among workers[13,31]. The differentiation of P.freuden-reichii subsp. freudenreichii and P. freuden-reichii subsp. shermanii on the basis of

nitrate reduction and lactose fermentationmay be significant for the dairy industry butquestionable at the genetic level, since John-son and Cummins [16] reported a high DNAhomology between the two subspecies. The23S rRNA results in the present study con-firm that there is no difference at the geneticlevel.

3.4. Plasmid profiles

Plasmid profile analysis was carried outon the majority of propionibacteria fromlowland milk. Of 373 strains tested, only 30 %contained plasmids. Of the plasmid-con-taining strains, 83 % carried one, 13 % twoand 4 % more than two plasmids. Threestrains possessed five plasmids.

Of the P. freudenreichii strains, 29 %carried plasmids, of P. jensenii 37 % andof P. acidipropionici 30 %. Al! eight strainsof P. thoenii lacked plasmids. The size ofthe plasmids ranged from 1.2 kb to morethan 16 kb. Eighty-nine percent of al! theplasmid-carrying strains contained plas-mids > 16 kb. Plasmids with a size between6 and 8 kb were relatively frequent (62 %) inP. jensenii strains.

Perez Chaia et al. [26] found plasmids in25 % of the propionibacteria strains ana-lyzed by them. However, their studyincluded only 27 P. freudenreichii and 3P. acidipropionici and no P. jensenii strainscompared to the 373 strains in this study.In the present study, P. jensenii had thehighest percentage of plasmid-carryingstrains. The size of plasmids varied between3.2 and 47 kb; smaller plasmids were notobserved. Panon [24] detected plasmids in38 % of a total of 53 strains, which includedal! the propionibacterial species.

The functional properties of Propioni-bacterium plasmids are not known. It wassuggested [28] that lactose utilization inP. freudenreichii might be plasmid linked.This was not supported by the results in thepresent study since only 38 of the 78 plas-mid-carrying P. freudenreichii strains were

207

able to ferment lactose and were, conse-quently, classified as P. freudenreichiisubsp. shermanii, and 40 were identified asP. freudenreichii subsp. freudenreichii. Inaddition, 137 P.freudenreichii subs P. sher-manii strains and 141 P. freudenreichiisubsp. freudenreichii carried no plasmids.Since not al! the strains carried plasmids, itwas not possible to differentiate aIl thestrains by plasmid analysis.

3.5.RAPD

The method of choice for investigatingstrain variety in Swiss raw milk proved to beRAPD. Several primers were studied andthe two (SK2 and DF4) that gave the mostconsistent number and clearest bands werechosen. Both were 12-mers with 100 % G +C. Primer SK2 was used for al! propioni-bacteria strains, whereas primer DF4 wasused for P.freudenreichii strains only. Theaverage intra-gel reproducibility with 15strains was 94 ± 2 % and inter-gel repro-ducibility with 12 strains was 90 ± 2 %. Thetotal of 278 P.freudenreichii strains yielded,at the 88 % similarity level, 154 differentprofiles with primer DF4 and 112 profileswith primer SK2.

The identification of the strains wasprimer dependent, because with RAPD onlya smal! part of the total DNA is amplified.The RAPD profiles for two strains may,therefore, be identical with one primer, butdifferent with another primer.

Only 59 strains gave the same profilewith both primers, so that finally 219 dif-ferent P.freudenreichii strains were obtained.Identical strains usually came from the sameregion. The P. freudenreichii strain diver-sity in Swiss raw milk is important to note,since it offers a large reservoir of strains foruse in the dairy industry. Figure 4 showsan example of the RAPD profiles fromP. freudenreichii strains from lowland milk.

Primer SK2 permitted classification ofthe 32 P. acidipropionici, 14 P. thoenii and

208 D. Fessier et al.

Figure 4. RAPD profiles with primer SK2 ofP.freudenreichii strains from lowland raw milk.2,3,4,5,6,7,8,9, 10 = P.freudenreichii strains;1, Il = 50-2 000 bp ladder.Figure 4. Profiles de RAPD avec amorce SK2 desouches de P. freudenreichii du lait cru de laplaine (2, 3, 4, 5, 6, 7, 8, 9, 10 = P.freudenreichiistrains; l, II = 50-2000 bp ladder).

75 P. jensenii strains into 30, 12 and 50 dif-ferent groups, respectively. Because of thelesser significance of these species com-pared to P. freudenreichii in Emmentalcheese, only one primer was used to differ-entiate the strains.

It is also interesting, that of the four pairsof strains having identical RA PD patternsand originating from the same milk, two didnot have the same plasrnid profile. This mayindicate horizontal transfer of plasmidsbetween bacteria and/or spontaneous lossof plasmids.

4. CONCLUSION

The propionibacterial flora in Swiss rawrnilk is extremely diverse. In raw milk fromthe lowlands, ail four dairy species werefound. In raw milk from the Alps, P. aci-dipropionici was absent and P. thoenii madeup 30 % of the strains compared to only 2 %in lowland raw milk. The reason for thediversity of Propionibacterium species in

Swiss milks can only be guessed at, sincedifferent factors such as grass feeding,sojoum of cows on the Alps during the sum-mer, propionibacterial flora on the farm andin the cheese factory cou Id play a role. Mostof the strains were only detected once byRAPD. As only 30 % of ail analyzed pro-pionibacteria carried plasmids, plasmid con-tent was not useful for further strain identi-fication. The plasmids might, however, beinteresting for future studies, as the charac-teristics coded on them remain unknownand there is a possibility of horizontal trans-fer and/or spontaneous loss. The cheeseindustry has in Swiss raw rnilk a large reser-voir of Propionibacterium strains for futureapplications and developments.

REFERENCES

[Il Baer A., Identification and differentiation ofpropionibacteria by e1ectrophoresis of their pro-teins, Milchwissenschaft 42 (1987) 431--433.

[2] Baer A., Ryba 1., Identification of propionibac-teria and streptococci by immunoblotting, Milch-wissenschaft 46 (1991) 292-294.

[3] Baer A., Ryba 1., Serological identification ofpropionibacteria in milk and cheese sampI es,Int. Dairy J. 2 (1992) 299-310.

[4] Baer A., Ryba 1., Grand M., Ursachen derEntstehung von braunen Tupfen im Kâse,Schweiz. Milchw. Forschung. 22 (1993) 3-7.

[5] Beerens H., Neut C., Romond C., Importantproperties in the differentiation of gram-posi-tive non-sporing rods in the genera Propioni-bacterium, Actinomyces and Bifidobacterium,in: Barnes E.M., Mead G.c. (Eds.), AnaerobieBacteria in Habitats other th an Man, Vol. 2,Blackwell Scientific Publications, UK, 1986,pp. 1346-1353.

[6] Carcano M., Todesco R., Lodi R., Brasca M.,Propionibacteria in Italian hard cheeses, Lait 75(1995) 415--426.

[7] Costas M., Numerical analysis of sodium dode-cyl sulphate-polyacrylarnide gel electrophoreticprotein patterns for the classification, identifi-cation and typing of medically important bac-teria, Electrophoresis II (1990) 382-391.

[8] Costas M., Pot B., Van dam me P., Kersters K.,Owen RJ., Hill L,R., Interlaboratory compara-tive study of the numerical analysis of one-dimensional sodium dodecyl sulphate-poly-acrylamide gel electrophoretic protein patternsof Campylobacter strains, Electrophoresis II(1990) 467--474.

Propionibacterial !lora in Swiss raw milk 209

[9] Costas M., Holmes B., Ganner M., On S.L.W., [23] Orla-Jensen S., Die Hauptlinien des nalürlichenHoffman P.N., Worsley M.A., Panigrahi H., Bakteriensystems, Zentralblall für Bakteriolo-Identification of outbreak-associated and other gie, Parasitenkunde und Infektionskrankheiten,strains of Clostridium difficile by numerical Abteilung 2 (1909) 305-346.analysis of SDS-PAGE protein patterns, Epi-

[24] Panon G., Presence of plasmids in propionicdemiol. Infect. 113 (1994) 1-12.[10] Crow V.L., Utilization of lactate isomers by

acid bacteria, Lait 68 (1988) 103-107.

Propionibacterium freudenreichii subsp. sher- [25] Park H.S., Reinbold G.W., Hammond E.G., Rolemanii: regulatory role for intracellular pyruvate, of propionibacteria in split defect of SwissAppl. Environ. Microbiol. 52 (1986) 352-358. cheese, J. Dairy Sci. 50 (1967) 820-823.

[II] Cummins CS., Johnson JL, Propionibacterium, [26] Perez Chai a A., Sesma F., Pesee de Ruiz Hol-in: Sneath P.H.A., Mair N.S., Sharpe M.E., gado A.P., Oliver G., Screening of plasmids inHolt J.G. (Eds.), Bergey's Manual of Systematic strains of Propionibacterium and mesophilicBacteriology, The Williams & Wilkins Co., Bal- lactobacilli isolated from Swiss-type chee ses ,timore, 1986, pp. 1345-1353. Microbiol. Aliments Nutr. 6 (1988) 171-174.

[12] De Carvalho A.F., Systématique des bactéries [27] Plastourgos S., Vaughn R.H., Species of Pro-laitières: classification, nomenclature et identi-fication, dissertation 9429, 69, ENSA de Rennes, pionibacterium associated with zapatera spoilage

1994. of olive, Appl. Microbiol. 5 (1957) 267-271.

[13] De Carvalho A.F., Gautier M., Grimont P.A.D., [28J Rehberger T.G., Glatz B.A., CharacterizationIdentification of dairy Propionibacterium spccies of Propionibacterium plasmids, Appl. Environ.by rRNA gene restriction patterns, Res. Micro- Microbiol. 56 (1990) 864-871.biol. 145 (1994) 667-676. [29] Riedel K.HJ., Britz TJ., Propionibacterium

[14] Fessier D., Casey M.G., Puhan Z., Rapid iden- species diversity in anaerobie digesters, Bio-tification of dairy Propionibacterium species divers. Conserv. 2 (1993) 400-411.by restriction analysis of the insertion within

[30] Riedel K.HJ., Britz TJ., Justification of thethe 23S rRNA gene, Lait 78 (1998) 203-216.[15] Hettinga D.H., Reinhold G'w., Vedamuthu E.R.,

'classical' Propionibacterium species conceptby ribotyping, Syst. Appl. Microbiol. 19 (1996)

The split defect of Swiss cheese: I. Effect of 370-380.strain of Propionibacterium and wrapping mate-rial, J. Milk Food Technol. 37 (1974) 322-328. [31] Riedel K.HJ., Wingfield B.D., Britz TJ., Jus-

[16] Johnson J.L., Cummins CS., Cell wall cornpo- tification of the 'classical' Propionibacteriumsition and deoxyribonucleic acid similarities species concept by restriction analysis of theamong the anaerobie corynefonns, classical pro- 16S ribosomal RNA genes, Syst. Appl. Micro-pionibacteria, and strains of Arachnia propi- biol. 17 (1994) 536-542.onica, J. Bacteriol. 109 (1972) 1047-1066. [32] Roller C, Ludwig W., Schleifer K.-H., Gram-

[17] Laemmli U.R., Most commonly used discon- positive bacteria with a high DNA G+C con-tinuous buffer system for SDS electrophoresis, tent are characterized by a corn mon insertionNature 227 (1970) 680. within their 23S rRNA genes, J. Gen. Micro-

[18] Langsrud T., Reinbold G.W., Flavor develop- biol. 138 (1992) 1167-1175.ment and microbiology of Swiss cheese. 1. Milk [33] Sokal R.R., Michener CD., A statistical methodquality and treatments, J. Milk Food Technol. for evaluating systematic relationships, Univ.36 (1973) 487--490. Kansas Sci. Bull. 38 (1958) 1409-1438.

[19] Langsrud T., Reinbold G.W., Flavor develop-[34] Thierry A., Madec M.N., Enumeration of pro-ment and microbiology of Swiss cheese.

II. Starters, manufacturing processes and pro- pionibacteria in raw milk using a new selectivecedures, J. Milk Food Technol. 36 (1973) medium, Lait 75 (1995) 315-323.531-542. [35] Tsakalidou E., Manolopoulou E., Kabaraki E.,

[20] Lowry O.H., Rosebrough NJ., Farr A.L., Ran- Zoidou E., Pot B., Kersters K., Kalantzo-dall RJ., Protein measurement with folin phenol poulos G., The combined use of whole-cell pro-reagent, J. Biol. Chem. 193 (1951) 265-275. tein extracts for the identification (SDS-PAGE)

[21] Mantere-Alhonen S., Über neuere Unter- and enzyme activity screening of lactic acid bac-suchungen der Propionsâurebakterien. Eine teria isolated from traditional Greek dairy prod-Ubersicht aus der Literatur, Meijeritiet Aikak ucts, Syst. Appl. Microbiol. 17 (1994) 444--448.35 (1977) 88-99. [36] Wallace RJ., Colla M.A., Metabolism of nitro-

[22] Morrissey J.H., Silver stain for proteins in poly- gen-containing compounds, in: Hobson P.N.acrylamide gels: a modified procedure with (Ed.), The Rumen Microbial Ecosystem,enhanced unifonn sensitivity, Anal. Biochem. Elsevier Applied Science, London, 1988,117 (1981) 307-310. pp. 185-216.