Effects of free fatty acids on propionic acid bacteria

HAL Id: hal-00929416https://hal.archives-ouvertes.fr/hal-00929416

Submitted on 1 Jan 1995

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

Effects of free fatty acids on propionic acid bacteriaP Boyaval, C Corre, C Dupuis, E Roussel

To cite this version:P Boyaval, C Corre, C Dupuis, E Roussel. Effects of free fatty acids on propionic acid bacteria. LeLait, INRA Editions, 1995, 75 (1), pp.17-29. �hal-00929416�

Lait (1995) 75, 17-29© Elsevier/INRA

17

Original article

Effects of free fatty.acids on propionic acid bacteria

P Boyaval 1, C Corre 1, C Dupuis 1, E Roussel 2

1 Laboratoire de Recherches de Technologie Laitière, INRA, 65, rue de St Brieuc,35042 Rennes Cedex;

2 Standa-Industrie, 184, rue Maréchal-Galliéni, 14050 Caen, France

(Received 10 May 1994; accepted 21 November 1994)

Summary - The seasonal variations in milk fat composition, especially du ring the grazing period, oftenlead to poor eye formation in Swiss-type cheese. The influence of free fatty acids on the grow1h andmetabolism of the dairy propionibacteria has been studied in this work. Linoleic (C1B:2), laurie (C12:0),

myristic (C14:0) and oleic acids (C1B:1) inhibited the growth and acid production of P treudenreichiisubsp shermanii in the reference medium. The antibacterial activity of Iinoleic acid can be overcomeby additions of cholesterol and soya lecithin to the medium. The four species and !WO subspecies of dairyPropionibacterium could be divided into !WO groups according to their high susceptibility to unsaturatedfatty acids (P freudenreichii subsp shermanii and subsp freudenreichit) or low susceptibility (P acidipro-pionici, P jensenii and P thoeniJ). Nevertheless, this inhibitory action of free fatty acids was not observedin milk, retentate media or lactic curd. A total extraction of the milk lipids led to the recovery of thisinhibitory effect on Propionibacterium. The possible modes of action of these molecules are discussedon the basis of the observed potassium effluxes and disturbances in the cell membranes.

free fatty acid 1 Propionibacterium 1 inhibition llinoleic acid Ilauric acid

Résumé - Effet des acides gras libres sur les bactéries propioniques. Les variations saisonnièresde composition du lait peuvent avoir des répercussions importantes sur la qualité des fromages. Leslaits de printemps conduisent souvent à de graves défauts d'ouverture en fabrication de fromage à pâtepressée cuite (PPC) et notamment d'emmental. Ce travail montre l'action fortement inhibitrice defaibles quantités d'acides gras libres (10 à 100 mg/t-I) sur le développement, dans leur milieu de cul-ture de référence, des bactéries propioniques, agents de l'ouverture de ces fromages. Les acideslinoléique (CIB:2), laurique (CI2:0), myristique (CI4:0) et oléique (CIB:I) inhibent la croissance et la pro-duction d'acides propionique et acétique de Propionibacterium freudenreichii subsp shermanii. Laprésence de cholestérol ou de lécithine de soja dans le milieu de culture restaure l'activité cellulaire.Deux grands groupes de susceptibilités différentes à ces actions inhibitrices peuvent être formésparmi les souches des 4 espèces de bactéries propioniques laitières: les 2 sous-espèces de P freu-denreichii (grande susceptibilité) d'une part et P acidipropionici, P jensenii et P thoenii (faible sus-ceptibilité) d'autre part. Néanmoins, l'action inhibitrice de l'acide linoléique ne se retrouve ni lors de cul-ture sur lait ou sur rétentat (facteur de concentration volumique = 5), ni dans des caillés préparésselon la technologie des fromages PPC. Une délipidation totale du lait permet à nouveau l'expression

1

1

1

1

1

1

1

18 P Boyaval et al

du caractère inhibiteur de cet acide gras libre sur les bactéries propioniques. Les modes d'action pos-sibles de ces acides gras sont discutés au regard des résultats de pertes de potassium et des per-turbations engendrées par ces acides gras libres au niveau des membranes cellulaires.

acide gras libre /Propionibacterium / inhibition / acide linoléique / acide laurique

INTRODUCTION

The inhibitory action of rancid milk on thegrowth of numerous bacteria has beenshown (Tarassuk and Smith, 1940) and con-firmed later (Costilow and Speck, 1951).The inhibitory effects and also the growthpromoting properties of fatty acids werereviewed by Nieman (1954). The most pro-nounced antibacterial effects of low con-centrations of fatty acids have been notedwith Gram-positive bacteria.

The antibacterial activity of free fatty acidswas shown at the same time, especially onlactic acid bacteria (Humfeld, 1947; Poz-nanski et al, 1968). This activity is some-times bactericidal but, with most microor-ganisms, the antibacterial action of fattyacids is bacteriostatic (Nieman, 1954).Kodicek and Worden (1945) have shownthat oleic and linolenic acids inhibit thegrowth of Lactobacillus helveticus. Lauricand myristic acids also proved to beinhibitory for this microorganism (Williamsand Fieger, 1946). Later, Anders and Jago(1964a, b) showed that oleic acid, accumu-lated early during the cheese ripening pro-cess, inhibited the viability of lactic acidstreptococci. This oleic acid alters the pyru-vate metabolism of group N streptococci(Anders and Jago, 1970). The cis-form ofunsaturated fatty acids exhibited a greaterantibacterial activity than the correspond-ing trans isomers on L helveticus, M tuber-culosis and B subtilis (Nieman, 1954; Gal-braith and Miller, 1973; Kabara, 1979).

Milk fat mainly consists of triglyceridesof fatty acids (97 to 98%) (Kurtz, 1965).Many factors determine the proportion of

fatty acids in milk fat, especially the feed-ing, the lactation stage, the animal speciesor the breed of cow. The seasonal varia-tions in the composition of milk fat influencedairy processing. This assertion is particu-larly true in cooked hard cheese process-ing, where the milk fat represents 45 to 55%of the dry matter. This fat influences therheological properties (Creamer and Oison,1982), the flavour (Adda et al, 1982) andthe microbial transformation of the cheese(Chen et al, 1979 ).

As the natural lipase of milk is almostcompletely destroyed by the cooked hardcheese processing (Driessen, 1983), themain Iipolytic activities in this type of cheeseare of microbial origin: psychrotrophic bac-teria (Law et al, 1976), lactic and propionicstarters and surface flora for Comté andBeaufort.

Lactic acid bacteria exhibit a low lipolyticactivity (Stadhouders and Veringa, 1973),but propionic acid bacteria, which are themain ripening agents in Swiss-type cheeses,are lipolytic (Knaut and Mazurek, 1974;Dupuis and Boyaval, 1993; Dupuis et al,1993). Free fatty acid concentrations as highas 14 9 kg-1 of hard cheese were reportedafter ripening (Woo et al, 1984). After pro-pionic and acetic acids, which are the directresult of propionic acid bacteria activity, themajor fatty acids are oleic (C18:1), palmitic(C16:0), stearic (C18:0) and myristic acids(C14:0), as in milk fat (Langler and Day,1966; Masson et al, 1978; Deeth et al, 1983;de Jong and Badings, 1990).

The seasonal variations which changethe proportions of these acids, especially inspring (Decaen and Journet, 1966; Gray,

Fatty acids and propionibacteria

1973; Masson et al, 1978), affect the Swiss-type cheese ripening. Poor eye formationand lower levels of propionic and aceticacids in the curd suggest a poor propionicacid bacteria activity. This problem has aconsiderable economic impact for the manu-facturers who need to know the exact factorsleading to these depreciated cheeses.

ln this study, we have examined theinfluence of free fatty acids on growth, acidproduction (propionic and acetic acids) andlactate consumption of Propionibacteriumfreudenreichii subsp shermsnii, the mostfrequently found propionic acid bacteria inSwiss-type cheese and of a type-strain ofeach dairy species of this genus. Moreover,potassium efflux rates, oxygen consump-tion and ~\jI (transmembrane electricalpotential) alterations generated by Iinoleicmolecules on Propionibacterium have beenmeasured in order to understand the natureof this influence.

MATERIALS AND METHODS

Strains and growth media

Five type-strains (Cummins and Johnson, 1986),of each species or subspecies of dairy Propioni-bacterium were used during this study: P freuden-reichii subsp freudenreichii CIP 103026 (Collec-tion de l'Institut Pasteur, Paris, France),P freudenreichii subsp shermanii CIP 103027,P jenseniiCIP 103028, P thoeniiCIP 103029 andP acidipropionici DSM 4900 (Deutsche Samm-lung von Mikroorganismen und Zellkulturen,Braunschweig, Germany). Moreover, one strainfrequently used in cheese factories, Propioni-bacterium freudenreichii subsp shermanii LRTL30 (INRA, Rennes, France), was also usedthroughout this study. Strains were propagatedin a lactate-yeast extract based medium (YEL)(Malik et al, 1968) at 30°C without shaking. Via-bility was evaluated by plating 1 ml of cell sus-pension on YEL agar and incubated at 30°C for 4days.

Stock cultures were maintained at -70°C inYEL medium containing 15% (vlv) glycerol (Pro-

19

labo, France). Ali media were sterilized by heattreatment (120°C, 15 min).

Cultures on milk were carried out with a lowheat milk powder (milk 'G') recommended byChamba and Prost (1989) for the evaluation ofacidifying activity of thermophilic lactic acid bac-te ria.

Cheese curds were prepared according toBuisson et al (1987) with microfiltrated milk(Trouvé et al, 1991). The phase of pressing wasnot carried out.

Analysis

Bacterial growth was followed by optical densitymeasurements (650 nm; Beckman spectropho-tometer DU 7400, Gagny, France) correlated withdry weight measurements. Lactic, propionic andacetic acid concentrations were measured usingan HPLC system (Beckman, Gagny, France)equipped with an UV detector (214 nm; Scan-ning detector module 167). Separation of sam-pie of 20 JlI took place in a 7.5 x 300 mm (AminexA6, ion exchange, Biorad) stainless steel column,operated at ambient temperature with H2S04

5 mmoll-1 (1 ml rnirr'") as eluent.

Chemicals reagents

Acetic acid (C2, Merck, Darmstadt, Germany);propionic acid (Cs, Sigma, St Quentin Fallavier,France); n-butyric acid (C4, Sigma); caproic acid(Ca, Sigma); caprylic acid (Cs' Merck); capric acid(C1Q, Merck); laurie acid (C12, Prolabo, Paris,France); oleic acid (C1S:1' Prolabo) and Iinoleicacid (C1S:2' Prolabo) were used in this study, dis-solved in ethyl alcohol of a high chemically pureform (Prolabo). Cholesterol was purchased fromProlabo, soya lecithin from Lucas-Meyer (France,'Emulfluid E') and bovine albumin (BSA, essen-tially globulin free) from Sigma.

Growth and activity in the presenceof free fatty acids

Cultures were carried out at 30°C in 100 ml glassbottles with 10 and 100 mg 1-1 of each fatty acidin the YEL medium. Controls were made for each

20 P Boyaval et al

solvent (ethanol) concentration employed. ThepH was not monitored.

Oxygen consumption measurements

Oxygen consumption was measured polaro-graphically using a Clarke-type electrode con-nected to a Gilson oxygraph (Model K-ITC-C,Middleton, Wl, USA). Measurements were car-ried out at 30°C on cells incubated in phosphatebutter (50 mmol 1-1, pH 6.50) at a final concen-tration of 0.3 mg ml-1 (dry weight: DW).

Potassium efflux measurements

The variations in the potassium content of thecells (K+in) were determined by measuring thechanges of the potassium concentration in theexternal mediurn (K+out) with a potassium-vali-nomycin-selective electrode (Radiometer) asso-ciated with a calomel reference electrode con-taining a secondary salt bridge filled with asolution of 100 mmol 1-1 NaCI (Boulanger andLetellier, 1988). The cation electrode potentialwas calibrated with KCI solutions of known con-centrations, which were prepared in the experi-mental butter. To estimate the total potassiumcontent of the bacteria, the cation was released bytreatment of the cells with 0.4 mmoll-1 Iinoleateor 2 mmol 1-1 laurate. The amount of releasedcation was th en estimated with the potassiumelectrode. K+in was calculated from the value ofK+out and expressed in nmol mg-1 ml-1 (cell DW)assuming that 4 x 109 cells correspond to 1 mgml-1 (cell DW). Whatever the product natureadded to the cells, the ethanol concentration wasalways less than 1% (v/v).

Measurement of the transmembraneelectrical potential (.6.0/)

tl.\jI was determined from the accumulation of[14C] TPP+ (tetraphenylphosphonium ion) (10umol lr", final concentration, 2.17 GBq mmol-1)(Ghazi et al, 1986). The cells were filtered onglass fibre filters (GF/F Whatman), washed with4 ml of butter and counted for radioactivity. TPP+uptake was corrected for unspecific binding bysubtracting a blank obtained under identical con-

ditions, except that the ce Ils were pretreated withthe protonophore TCS (3, 3', 4', 5 - tetrachloro-salicylanilide) (10 j.lmoll-1, final concentration).

RESULTS

Inhibitoryaction of free fattyacids (FFA)

Even at very low concentrations « 1a mg1-1), the palmitic (C16:0) and stearic acids(C18:0) were not soluble in ethanol. Theywere soluble in propanol-1 and chloroformbut formed a precipitate in the YEL medium.Consequently, they were not tested furtherin this study.

At 10 mg 1-1, growth of P freudenreichiisubsp shermanii LRTL 30 on YEL mediumwas clearly inhibited by the linoleic acid andslightly by the lauric (C12:0), myristic (C14:0)and oleic (C18:1) acids. These bacteriostaticeffects were clearly evidenced at 100 mg1-1, particularly for the lauric (C12:0) andIinoleic (C18:2) acids (table 1). A restorationof growth after 25 h was observed for theoleic acid (C18:1). The lactate consumptionwas affected in the same way with nodecrease for laurie (C12:0) and linoleic acids(C18:2) and a slight decrease for myristic(C14:0) and oleic acids (C18:1). The propi-onate and acetate productions reflectedexactly the lactate evolution. Moreover, alow positive (or no) effect of the fatty acidsfrom acetic (C2:0) to capric acids (ClO:0) onthe initial (24 h) lactate consumption andpropionic and acetic acids production wasnoted.

The effects of caproic acid (C6:0) andlinoleic acid (C18:2) on the type-strains ofthe four species and two subspecies of dairypropionibacteria are shown in table II. Aslight positive effect of the caproic acid onthe growth of the two P freudenreichii strainswas observed. This effect was not signifi-cant for the three other species. Linoleicacid drastically affected the growth of

Fatty acids and propionibacteria 21

Table 1. Effect 01 caproic acid (Cs:o) and linoleic acid (C1S:2) at 100 mg 1-1 on the growth 01 the livetype-strains 01 dairy Propionibacterium.Effets des acides caproïque (C6:0) et linoléique (CI8:2) à 100 mg 1-1 sur la croissance des 5souches-type de bactéries propioniques laitières.

FFA Strain

103026 103027 4900103028 103029

C6:0

C18:2

o o o

Inhibitory action: --- high: .- medium: - low; 0, no activity.Action inhibitrice: --- forte; -- moyenne; • faible; 0 : pas d'activité.

Table Il. Inhibiting action 01 Iree latty acids onthe growth, propionate production and lactateconsumption 01 P freudenreichii subsp sher-manii LRTL 30 compared ta a control withoutlattyacid.Action des acides gras libres sur la croissance, laproduction de propionate et la consommation delactate de Propionibacterium Ireudenreichii subspshermanii LRTL 30 (comparée à un témoin sansaddition d'acide gras libre).

Fattyacid Growth Propionate Lactateproduction consumption

C2 0 + +C3 0 + 0C4 0 + 0Cs 0 + 0Cs 0 +C1Q +C12

C14

C1S:1

C1S:2

Inhibition: -", high; -', medium; " low. Activation: +++,high; ++, medium; +, low; 0, no effect.tnhibition: --', forte; --, moyenne; " faible. Activation:+++, forte; ++ moyenne; + faible; 0 sans effet.

P freudenreichii subsp freudenreichii and,to a lesser extent, the growth of P freuden-reichii subsp shermanii, P theonii. P jenseniiand P acidipropionici growth was affectedduring the first 30 h, but the number of cellssharply increased afterwards to reach a levelhigher than the control cultures (28% morefor P jensenii and 33% for P acidipropionicl).Globally, these results were confirmed bythe lactate consumption and the propionicand acetic acid production.

As the natural medium of Propionibac-terium multiplication in Swiss-type cheesetechnology is milk, we have tried to repeatour observations on milk. The main resultsare summarized in figure 1. Even in a milkcontaining a low lipid concentration (500 mg1-1), linoleic acid, at concentrations up to 5 91-1, had no effect on cell multiplication or onacid production. The inhibitory effect ofIinoleic acid was recovered after a previousextraction of the lipids from the milk pow-der with hexane before milk rehydration.Moreover, the inhibition of Propionibacteriumfreudenreichii subsp shermanii LRTL 30was also not evidenced in milk retentate(volume concentration factor of 5, obtainedby ultrafiltration; not shown). In order to bet-ter master the possible influence of free fatty

caproic acid (C6:0), nor by the presence ofbiotin at 10-5 or 10-4 9 1-1. However, soyalecithin (100 mg 1-1) and cholesterol (100mg 1-1) greatly reduced this negative effect(fig 2).

22 P Boyaval et al

acids in Swiss-type cheese technology, wehave examined the development of thisstrain in a cheese curd made from microfil-trated milk (which contained only lacticstarters before addition of propionic acidbacteria). The cell multiplication, during the340 h of the trial, was not affected by thepresence of 5 9 1-1 of linoleic acid (notshown).

Alterations of the FFA inhibition activitybyaddition of compounds

No growth stimulating effect of Tween 40,Tween 60, Tween 80 (at 100 mg 1-1) andbiotin (at 10-5 or 1Q-4 9 1-1)was observed.On the other hand, Tween 80 at 30 9 1-1counteracted the inhibitory effect of linoleicacid for the P jensenii CIP 103028 culture(not shown). The inhibitory effect of linoleic .acid (C18:2) at 100 mg 1-1 was reversed nei-ther by the presence of 10 or 100 mg 1-1 of

_ 1.8

o 0

60 80 100

Time(h)

Fig 1. Effect of linoleic acid (5 g 1-1) on propio-nic acid production by P freudenreichii subspshermanii LRTL 30 in milk G (.) and in milk Gpreviously treated with hexane (.À.). Controlwithout addition of linoleic acid (0). Contrais arecarried out with the same solve nt concentrationsas in trials.Effet de J'acide linoléique (5 g rI) sur la pro-duction d'acide propionique par des cellules deP freudenreichii subsp shermanii LRTL 30 cul-tivées sur lait G (.) et sur lait G après traite-ment à J'hexane (~). (0) représente la produc-tion d'acide propionique sans addition d'acidelinoléique. Les témoins sont réalisés avec lesmêmes concentrations de solvant que dans lesessais.

.~ 3.5<JlCCI>

"ai 2.5.2a0

1.5

0.5

o ~=~:!!:=:==~!:::::;:::~----<o 20 60 80

Timelh)

40

4

3.5

1.5

0.5

ot:=~~~~~~~~-----io 20 40 60 80

Timelh)

120Fig 2. Effect of linoleic acid (C18:2) (100 mg 1-1)

(.); linoleic (100 mg 1-1) + cholesterol (100 mg 1-1)

(0); linoleic (100 mg 1-') + biotin (10-4 g 1-1) (+);linoleic (100 mg 1-1) + soya lecithin (100 mg 1-1)

(e); linoleic (100 mg 1-1) + caproic acid (C6:0) 10mg 1-1 (.À.) and 100 mg 1-1 (1:\) on the growth of Pfreudenreichii subsp shermanii LRTL 30. Contraisare carried out with the same solve nt concentra-tions as in trials.Effets des acides linoléique (CI8:2)(100 mg rI)(.); linoléique (100 mg rI) + cholesterol (100 mgrI) (0); linoléique (100 mg rI) + biotine (1D-4 grI) (+); linoléique (100 mg rI) + lécithine de soja(100 mg t:'} (e); linoléique (100 mg rI) + acidecaproïque (C6:0) à 10 mg.rl (~) et à 100 mg r!(1:\) sur la croissance de P freudenreichii subspshermanii LRTL 30. Les témoins sont réalisésavec les mêmes concentrations de solvant quedans les essais.

Fatty acids and propionibacteria

Alterations of the cell membraneintegrity: potassium efflux

Total cellular K+ initially present in the Pro-pionibacterium cells was evaluated at 1650nmol mg-1 dry weight, a value close to the1800 nmol mg-1 determined by Duperrayet al (1992) for the closely related species ofCorynebacterium glutamicum. Figure 3 re-presents the effect of increasing concen-trations of linoleic acid on the initial rate of K+efflux. The minimum concentration of linoleicacid necessary to detect an efflux of K+ incells incubated at pH 6.50 is 6 Ilmoll-1 (20Ilmoll-1 for laurie acid). Addition of 1251lmol1-1of linoleic acid induced a complete andrapid efflux of K+ ions (1600 urnol min-1mg-1 (DW)). Lauric acid also induced anefflux of cytoplasmic K+. This efflux wasmaximal only for a laurie acid concentrationof 900 Ilmoll-1, more than 7 times the con-centration of linoleic acid for the same rate(fig 4). Addition of 50 umol t' of BSA to thecell suspension after addition of the freefatty acid totally counteracted the action of251lmoll-1 of linoleic acid on K+ efflux (thesame effect was observed for 200 Ilmoll-1BSA after addition of 51 umol r' of linoleicacid). Moreover, a 1 min incubation with 50umol r' BSA totally protected the internai K+cellular pool against the action of 51 ~011-1of linoleic acid. The addition of MgS04 at 1mmol 1-1 completely counteracted the K+efflux induced by the previous adjunction of52 urnol 1-1of linoleic acid to the bacterialcells. Morever, in the presence of 20 mmol1-1 MgS04 in the medium, 50 urnol 1-1 oflinoleic acid have no effect on the K+ effluxfrom the P freudenreichii subsp shermaniicells. A previous incubation of the Propio-nibacterium cells with caproic acid (780 umol1-1) had no protective effect on K+ effluxinduced later by linoleic acid (25 urnol 1-1)(not shown).

The presence of 15 mg of Tween 80 inthe cell suspension had a total protectiveeffect against linoleic acid action (on K+

23

200

50 100 150 200Linoleic acid (/Lmol-1)

Fig 3. Effect of Iinoleic acid on inhibition of O2consumption (e), K+ efflux rate (0) and L'l'V (.) inPropionibacterium freudenreichii subsp sherma-niiLRTL30.Effet de l'acide linoléique sur la consommationd'02 des cellules de Propionibacterium freuden-reichii subsp shermanii LRTL 30 (e), la vitesse desortie du potassium intracellulaire (0) et le L'l'V(.).

200

800 1000Laurie acid (JUllol-1)

Fig 4. Effect of laurie acid on inhibition of O2consumption (e), K+ efflux rate (0) and L'l'V (.) inPropionibacterium freudenreichii subsp sherma-niiLRTL30.Effet de l'acide laurique sur la consommationd'02 des cellules de Propionibacterium freu-denreichii subsp shermanii LRTL 30 (e), lavitesse de sortie du potassium intracellulaire (0)et le L'l'V (.).

24 P Boyaval et al

efflux and on oxygen consumption), at leastup to 200 urnol 1-1 ! Moreover, addition ofTween 80 (15 mg) after addition of linoleicacid (50 Ilmoll-1) immediately stopped theK+ efflux. A 2 min incubation of the cellswith 200 Ilmoll-1 of cholesterol reduced by74% the K+ efflux induced by 1281lmoll-1 oflinoleic acid (not shown). But the additionof 100 urnol 1-1 or 200 urnol 1-1 of choles-terol after addition of 50 urnol t' of linoleicacid had no effect on that K+ efflux.

Action of FFA on the transmembraneelectrical potential

At pH 6.50, the Ô\j1 was 158 mV, a valueclose to the Ô\j1 determined for other Gram-positive bacteria (162 mV for S aureus and113 mV for S Jaetis (Kashket, 1981) and190 mV for Corynebacterium glutamicum(Duperray et al, 1992)). The cell membranealso became more permeable to H+ afteraddition of linoleic acid since a decrease ofthe Ô\j1 was observed (fig 3). This decreas-ing Ô\j1 fram 158 mV to a mV was evidencedin the same range of linoleate concentra-tion than K+ efflux. At a concentration of 900urnol l"" of laurie acid, the ô\j1 was a (fig 4).

Inhibition of oxygen consumption

The inhibition of oxygen consumption, ini-tially evaluated at 15 ± 2 nmol O2 min-1mg-1 (DW) was shown within the samerange as Ô\j1 decrease (fig 3). Lauric acidinhibitory effect towards oxygen consump-tion was roughly the same as observed forlinoleic acid except that total inhibitionoccurred at 250 Ilmoll-1 for laurie acid andat 100 Ilmoll-1 for linoleic acid (fig 4).

Caproic acid at concentrations up to 800Ilmoll-1 showed no action on oxygen con-sumption and K+ intracellular level (notshown), confirming the above results on cell

growth. Addition of 60 Ilmoll-1 of B8A hadno action on the cell oxygen consumption,but protected the cells against the action of100 Ilmoll-1 linoleic acid. But even with thisprotective action, 180 urnol 1-1 of Iinoleicacid induced a 90% decrease in oxygenconsumption. Lauric acid at 440 urnol 1-1achieved the same inhibition. This is thesame ratio observed as without B8A pro-tection (2.4 more laurie acid than Iinoleicacid).

DISCUSSION

Inhibitoryaction ofFFA

The main effects of the fatty acids on Pro-pionibacterium freudenreiehii subsp sher-manii grawth on YEL medium enable a sep-aration into three classes: c1ass l, from C2:0to ClO:0 no effect on growth and a low pos-itive effect on the fermentative capacities;class Il, C14:0 and C18:1 negative effect ongrowth and acid production, intermediateeffect between classes 1 and III; class III,C18:2 and C12:0 complete inhibition of thegrowth and metabolism of these bacteria.

To our knowledge, no information hasbeen published on the action of free fattyacids on propionic acid bacteria. Lactobacilli,although generally not sensitive to saturatedfatty acids, can be sometimes inhibited bythem if the compounds have a chain lengthof around C12, these being the most active(Hassinen et al, 1951). In this study, wehave observed the same effects on propio-nic acid bacteria.

A few explanations on the exact influ-ence of the inhibitory fatty acids on microbialcells have been proposed but not confirmed:i) they decrease the bacterial respiration(Galbraith and Miller, 1973); ii) they buildan adsorption layer around the cell whichchanges the cell permeability, leading to anoutward diffusion of vital cellular cornpo-nents or bloc king the adsorption of essential

Fatty acids and propionibacteria 25

nutrients (Nieman, 1954); iii) they are able,in the form of soaps, to lower the surfacetension of the media (Nieman, 1954) andthey inhibit the active transport of aminoacids (Galbraith and Miller, 1973).

As numerous works have shown that sev-eral surface active compounds do not inhibitthe growth of microorganisms, the first andsecond hypotheses must be controlledexperimentally. The fact that Gram-positivebacteria are more sensitive than Gram-neg-ative bacteria supports the intervention ofthe cell membrane structures in this phe-nomenon. Moreover, the work of Moss et al(1969) on the fatty acid composition of pro-pionibacteria leads to the separation of thespecies into two groups: P freudenreichiisubsp freudenreichii and subsp shermaniion the one hand and the three other specieson the other hand. This separation is in closeagreement with the two groups of Propioni-bacterium species reactivity towards fattyacids found in this study.

Alterations of the FFA inhibition activitybyaddition of compounds

The presence of protein can counteract theinhibition of fatty acids as observed by Dubos(1947) for the serum albumin, which has highaffinity binding sites for FFA (Frapin et al,1993). Many other components act similarly:saponin, lecithin, charcoal, starch, choles-terol (Kodicek and Worden, 1945). The anta-gonistic effect of cholesterol and lecithin on anantibacterial fatty acid, as observed in thisstudy, was already noticed by Wynne andFoster (1950), working with Micrococcuspyogenes var aureus and oleic acid. No clearinterpretation of this type of results has beenproposed to our knowledge.

ln 1990, Somkuti and Johnson showedthat P freudenreichii cells were able to bindthe cholesterol present in the medium bypolysaccharide or membrane bound pro-

teins or by other means. These indicationslead to an investigation into the role of thesecompounds, at the cell membrane level,with and without fatty acids in the medium.

ln this study we were unable to detectany positive effect of the polyoxyethylenederivative of fatty acid monoesters of sor-bitan (ester of oleic acid = Tween 80; esterof stearic acid = Tween 60 and ester ofpalmitic acid = Tween 40) at 100 mg 1-1 onthe growth of P freudenreichii subsp sher-manii LRTL 30. This growth stimulation wasfrequently evidenced in the past (Dubos,1947; Ledeoma et al, 1977; Cummins andJohnson, 1986). The neutralization of theantimicrobial properties of Iinoleic acid byTween 80 at 30 9 1-1 was also evidencedby Baker et al (1983) for laurie acid. Theexact mechanism of the Tween in neutral-izing the activity of this acid has not beenexplained in his paper.

The use of saturated fatty acids to inhibitthe negative action of unsaturated fatty acidsobserved by Hassinen et al (1950) onL bifidus was not confirmed using our strain.If the mode of action of these acids is aninteraction with the membrane layers of thecells, the affinity of the membranes for thefatty acid considered and a concentrationdependence of the phenomena are proba-bly involved. The ratios under investigationhere may be out of the correct range toobserve a detoxification.

Biotin is required for the growth of moststrains of dairy propionibacteria (Delwich,1949). No positive effect on growth of dairypropionibacteria by additions of biotin wasobserved in this study. The high level ofbiotin of our medium had probably hiddenthis potential effect. Williams et al (1947)supposed that biotin promotes, in some way,the formation of unsaturated fatty acidsessential for the bacteria. The relationshipsbetween fatty acids and biotin in the nutritionof microorganisms is still far from beingclearly defined.

26 P Boyaval et al

Alteration of the cell membrane activity

If ~'" = 0, at 100 urnol 1-1 of linoleic acid,~'" remained at 60% of its initial value afteraddition of 100 Ilmoll-1 of laurie acid. More-over, 250 umol 1-1 of lauric acid inducedonly a very slow K+ efflux in the cells of Ptreudenreichii subsp shermanii. It was thenevidenced that linoleic acid is a much moreactive compound than laurie acid as a Pro-pionibacterium cell membrane disturber.

The addition of linoleic or laurie acids inPropionibacterium cell suspensions resultsin an immediate perturbation of the permea-bility properties of the cytoplasmic mem-brane and of the bacterial energetic state:cells lose cytoplasmic potassium, theybecome partially or totally depolarized andtheir respiration is inhibited. Ali thesechanges indicate that the target of the freefatty acids is the cytoplasmic membrane.

To our knowledge, no information is avail-able about the K+ transport system(s) inPropionibacterium cells. However, we havecalculated in our experiments that one cellreached by fatty acid lost potassium at arate of 4.8 106 K+ S-1. This rate is manyorders of magnitude higher than one expectsfor passive efflux through the lipid bilayerestimated at 20 ions s-1/cell by Gomperts(1976) (if a permeability coefficient of 10-14m S-1 is assumed) . This suggests that theinhibitory action of free fatty acids occurredby some globallipid bilayer disturbances.

The K+ efflux is not the consequence ofmembrane depolarization since the cellsincubated with the protonophore TCS at 10Ilmoll-1, do not show K+ efflux (not shown).Even if the time necessary to reach the com-plete loss of K+ by the cells increased whenthe free fatty acid concentrations decreased,the level "0" is always attained.

Free fatty acids are known to uncoupleoxidative phosphorylation in mitochondria,chloroplasts and bacterial cells (Rottenbergand Hashimoto, 1986). Unsaturated long-

chain free fatty acids were evidenced as farmore pote nt uncouplers than the saturatedacids (Borst et al, 1962). But we cannotexclude the possibility of some cofactor leak-age from the cells. The protective effect ofBSA against the inhibitory effect of anuncoupler, as we observed in this study,has already been observed on mitochon-dria since 1956 (Pullman and Racker, 1956).It has been related to the presence of anhydrophobic site in that protein, which has agreat affinity for fatty acids.

As underlined by Borst et al (1962) themembrane disturbing activity of free fattyacids may probably be distinguished fromthat brought about by agents breaking upmitochondrial or bacterial envelope struc-ture in that it is readily reversible by the sub-sequent addition of serum albumin. How-ever, we have no c1ear interpretation of theprotective action of MgS04 on Propioni-bacterium cells.

The fatty acid inhibitory effects observedon the growth and metabolism of dairy Pro-pionibacterium are of great industrial impor-tance. Indeed, the interest in propionic acidproduction by fermentation is increasing(Boyaval and Corre, 1987; Boyaval, 1992).This revival for biological propionic acid pro-duction is based on the highly increasedproductivities allowed by the technology ofmembrane bioreactors. As yeast extracts,which are very frequently used in that typeof fermentation, contain unsaturated fattyacids (Eddy, 1958) the producer must takeinto account the susceptibility to theseinhibitory effect in the selection of the pro-pionic acid strains.

But the most important point is the pos-sible involvement of these fatty acids in theinhibition of dairy Propionibacterium in hardcheese manufacturing. These cheeses re-present more than 212 000 tons per yearin France, only for Emmental (CNIEL, 1994).The highly predominant species employedin hard cheese technology is P treudenreichiisubsp shermanii. If the response of most of

Fatty acids and propionibacteria

the strains of this species is similar to theresponse of the strain LRTL 30, it will prob-ably beinteresfinq to mix them with strainsof other species in spring, when the problemof eye formation occurs. Unfortunately, wehave been unable to find any informationon the level of free fatty acids in the cheesecurd at the end of the cold ripening, justbefore the development of the propionibac-teria. These analyses are currently underinvestigation. This finding underlines onceagain that the lipid content of the mediumhas a drastic protective effect on the cells,probably by integration of the free fatty acidin the lipid globules. But scanning electronmicroscopie examinations of Saint Paulincheeses suggest that the globule membranecould be disorganised by change in pH,enzymatic action and by the mechanicalaction of pressing (Rousseau. 1988). Mem-brane debris and fractured fat globule wereobserved. Moreover, some water was pref-erentially localized around the globules,leading to a kind of 'barrier' between theglobules and the hydrophobie fatty acids.These observations were completed by astudy on Emmental cheese where fat glob-ules had lost their initial structure to givelarge masses with diverse forms (Rousseauand Le Gallo, 1990). These points under-line that the fat globules, with such modifi-cations and reduced surface have adecreased ability to trap free fatty acids inthe cheese body. Our test with a curd whichhas not followed the complete cheese tech-nology process (not pressed), was proba-bly not sufficient to answer the questionasked. Moreover, the alterations of fat glob-ules, during processing and ripening of thesecheeses, could overcome their trappingeffect on free fatty acids as observed in milkand fresh cheese curd. Even if we have noclear biochemical data on the evolutionofthe lipid phase in a cheese curd from thebeginning of the technology to the end ofthe ripening period, these results cast doubton the major role of free fatty acids in thealteration of the eye formation during the

27

spring. The cycle of cheese manufacturingmust be completed in the initial presenceof Iinoleic acid in arder to discard this hypoth-esis or not.

ACKNOWLEDGMENTS

The authors gratefully acknowledge the regionof Brittany for the financial support. E Faivre forthe preparation of the cheese curd, A Ghazy andL Letellier for their expert help during K+ effluxmeasurements and paper elaboration and CGuinard and E Rees for the English improvementof the tex!.

REFERENCES

Adda J, Gripon JC, Vassal L (1982) The chemistry offlavour and texture generation in cheese. FoodChem 9,115-125

Anders RF, Jago GR (1964a) The effect of fatty acids onthe metabolism of lactic acid streptococci. 1. Inhibi-tion of bacterial growth and proteolysis. J Dairy Res31,81-89

Anders RF, Jago GR (1964b) The effect of fatty acids onthe metabolism of lactic acid streptococci. II. Resis-tance of a variant of Streptococcus cremoris strainC13. J Dairy Res 31,91-94

Anders RF, Jago GR (1970) The effect of fatty acids onthe metabolism of pyruvate in Iactic acid strepto-cocci. J Dairy Res 37, 445-456

Baker RC, Vadehra DV, Gravani RB (1983) Neutraliza-tion of antimicrobial properties of Lauricidin byTweens. J Food Safety6, 1-12

Borst P, Loos JA, Christ EJ, Slater EC (1962) Uncouplingactivity of long-chain fatty acids. Biochim BiophysActa 62, 509-518

Boulanger P, Letellier L (1988) Characterization of ionchannels involved in the penetration of phage T 4DNA into Escherichia coli cells. J Biol Chem 263,9767-9775

Boyaval P, Corre C (1987) Continuous fermentation ofsweet whey permeate for propionic acid productionin a CSTR with UF recycle. Biotechnol Lett9, 801-806

Boyaval P (1992) Production d'acides organiques enbioréacteur à membrane. Ind Aliment Agric 109, 953-958

Buisson V, Kerjean JR, Courroye M (1987) Miniaturisa-tion de la fabrication fromagère. Tech Lait Market1024,17-23

28 P Boyaval et al

Chamba JF, Prost F (1989) Mesure de l'activité acidili-ante des bactéries lactiques thermophiles utiliséespour la labrication des Iromages à pâte cuite. Lait69,417-431

Chen AM, Larkin JW, Trivin WE (1979) Texturai analy-sis 01 cheese. J Dairy Sei 62, 901-907

CNIEL (1994) L'économie laitière en chiffres. CNIEL,Paris

Costilow RN, Speck ML (1951) Inhibitory effect 01 rancidmilk on certain bacteria. J Dairy Sei 34, 1104-1110

Creamer LL, Oison NF (1982) Rheological evaluationof maturing cheddar cheese. J Food Sei 47,631-646

Cummins CS, Johnson JT (1986) Propionibacterium.ln: Bergey's manua/ of systematic bacteri%gy vo/II(Sneath PH, Mair NS, Sharpe HE, Holt JG, eds)Williams and Wilkins, Baltimore USA, 1346-1353

Decaen C, Journet M (1966) Influence saisonnière sur laproduction et la composition du lail. Ann Zootech15,259-277

Deeth HC, Fitzgerald CH, Snow AJ (1983) Agas chro-matographic method lor the quantitative determina-tion 01 Iree lally acids in milk and milk products. N ZJ Dairy Sei Techno/18, 13-20

de Jong C, Badings HT (1990) Determination of Ireelatty acids in milk and cheese. Procedures for extrac-tion, and capillary gas chromatographic analysis. JHigh Res Chromatogr 13, 94-98

Delwich EA (1949) Vilamin requirements 01 the genusPropionibacterium. J Bacterio/58, 293-298

Driessen FM (1983) Lipases and proteinases in milk.Occurrence, heat inactivation and their importance lorthe keeping qualily 01 milk products. Neth Milk DairyJ37,193-196

Dubos RJ (1947) The effect 01 Iipids and serum albu-min on bacterial grow1h. J Exp Med85, 9-22

Duperray F, Jezequel D, Ghazi A, Letellier L, ShechterE (1992) Excretion 01 glutamate Irom Corynebac-terium g/utamicum triggered by amine surfactants.Biochim Biophys Acta 1103, 250-258

Dupuis C, Boyaval P (1993) Esterase activity 01 dairyPropionibacterium. Lait 73, 345-356

Dupuis C, Corre C, Boyaval P (1993) Lipolytic andesterasic activities 01 Propionibacterium freudenre-ichii subsp freudenreichii. App/ Environ Microbio/59,4004-4009

Eddy AA (1958) Aspects 01 the chemical composition01 yeasl. In: The chemistry and bi%gy of yeasts(Cook AH, ed) Academic Press, New York

Frapin D, Dulour E, Haertle T (1993) Probing the lallyacid binding site 01 p -Iactoglobulins. J Protein Chem12,443-449

Galbraith H, Miller TB (1973) Effect 01 long chain Iattyacids on bacterial respiration and amino acid uptake.J App/ Bacterio/36, 659-675

Ghazi A, Delamours L, Shechter E (1986) Absence 01 aunique relationship between active transport 01 lac-tose and proton motive lorce in E coli. FEBS Let!209,325-329

Gomperts BD (1976) The plasma membrane: mode/sfor its structure and function. Academic Press,Orlando, USA.

Gray IK (1973) Seasonal variations in the compositionand thermal properties 01 New-Zealand milk lat.J Dairy Res 40, 207-214

Hassinen JB, Durbin GT, Bernhart FW (1950) Hexade-cenoic acid as a grow1h lactor lor lactic acid bacte-ria. Arch Biochem Biophys 25, 91-96

Hassinen JB, Durbin GT, Bernhart FW (1951) The bac-teriostatic ellects 01 saturated lally acids. ArchBiochem Biophys 31, 183-189

Humleld H (1947) Antibiotic activity 01 the lally acid-likeconstituents 01 wheat bran. J Bacterio/54, 513-517

Kabara JJ (1979) Toxicological, bactericidal and lungi-cidal properties 01 lally acids and some derivatives.J Am ail Chem Soc 56,760-767

Kashket ER (1981) Proton motive lorce in growing Strep-tococcus tectis and Staphy/ococcus aureus ce Ilsunder aerobic and anaerobic conditions. J Bacterio/146, 369-376

Knaut T, Mazurek K (1974) Caractère Iipolytique desbactéries propioniques. X/X /nt Dairy Congr NewDe/hi B6, 425-426

Kodicek E, Worden AN (1945) The effect 01 unsaturatedlally acids on Lactobacillus he/veticus and otherGram-positive micro-organisms. Biochem J 39, 78-85

Kurtz FE (1965) The lipids 01 milk: composition and prop-erties. /n : Fundamenta/s of Dairy Chemistry (WebbBH, Johnson AH, eds) Avi Publishing Company, Inc,Wesport USA

Langler JE, Day EA (1966) Quantitative analysis 01 themajor Iree latty acids in Swiss cheese. J Dairy Sei 49,91-93

Law BA, Sharpe ME, Chapman HR (1976) The elfect01 Iipolytic Gram negative psychrotrophs in storedmilk on the development 01 rancidity in Cheddarcheese. J Dairy Res 43, 459-468

Ledeoma av, de Ruiz Holdago AP, Oliver G, de GioriGS, Raibaud P, Galpin JV (1977) A synthetic mediumlor comparative nutritional studies 01 lactobacilli.J App/ Bacterio/42, 123-133

Malik AC, Reinbold GW, Vedamuthu ER (1968) An eval-uation 01 the taxonomy 01 Propionibacterium. CanJ Microbio/14, 1185-1191

Masson C, Decaen C, Rousseaux P, Bowty JL (1978)Variations géographique et saisonnière de la com-position du lait destiné à la labri cation de gruyèrede Comté. Lait 58, 261-273

Moss CW, Dowell VR, Farshtchi D, Raines LJ, CherryWB (1969) Cultural characteristics and latty acid com-position 01 propionibacteria. J Bacterio/97, 561-570

Fatty acids and propionibacteria

Nieman C (1954) Influence of trace amounts of fatlyacids on the growth of microorganisms. BacteriolRev 18,147-163

Poznanski S, Surazynski A, d'Obyrn T (1968) Activitéde fermentation et de protéolyse de certainessouches de bactéries lactiques en présence d'acidegras. Lait 48, 261-274

Pullman ME, Racker E (1956) Spectrophotometric stud-ies of oxidative phosphorylation. Science 123, 1105-1107

Rotlenberg H, Hashimoto K (1986) Fatly acid uncou-pling of oxidative phosphorylation in rat liver mito-chondria. Bioehemistry 25, 1747-1755

Rousseau M (1988) Changes in the microstructure ofSaint Paulin cheese during manufacture studied byscanning electron microscopy. Food Microstruet 7,105-113

Rousseau M, Le Gallo C (1990) Etude de la structurede l'Emmental au cours de la fabrication, par latechnique de microscopie à balayage. Lait 70, 55-66

Somkuti GA, Johnson TL (1990) Cholesterol uptake byPropionibacterium freudenreichii. Curr Microbio/20,305-309

Stadhouders J, Veringa HA (1973) Fat hydrolysis bylactic acid bacteria in cheese. Neth Milk Dairy J 27,77-91

Tarassuk NP, Smith FR (1940) Relation of surface ten-sion of rancid milk ta its inhibitory effect on the growthand acid fermentation of Streptococcus lactis. J DairySei 23, 1163-1170

Trouvé E, Maubois JL, Piot M, Madec MN, Fauquant J,Tabard J, Brinkman G (1991) Rétention de différentesespèces microbiennes lors de l'épuration du lait parmicrofiltration en flux tangentiel. Lait 71, 1-13

Williams VR, Fieger EA (1946) Oleic acid as a growthstimulant for Lactobacillus casei. J Biol Chem 166,335-343

Williams WL, Brognist HP, Snell EE (1947) Oleic acidand related compounds as growth factors for lacticacid bacteria. J Biol Chem 170, 619-630

Woo AH, Kollodge S, Lindsay RC (1984) Quantificationof major free fatly acids in several cheese varieties.J Dairy Sei 67, 874-878

Wynne ES, Foster JW (1950) Studies on the effects ofC18 unsaturated fatty acids on growth and respira-tion of Mierococcus pyogenes var aureus. J InfectDis 86, 33-37

29