MINI-REVIEW

E. Parente A. Ricciardi

Production, recovery and purication of bacteriocinsfrom lactic acid bacteria

Received: 29 December 1998 / Received revision: 23 April 1999 /Accepted: 23 April 1999

Abstract Bacteriocins produced by lactic acid bacteriaare a heterogeneous group of peptide inhibitors whichinclude lantibiotics (class I, e.g. nisin), small heat-stablepeptides (class II, e.g. pediocin AcH/PA1) and largeheat-labile proteins (class III, e.g. helveticin J). Manybacteriocins belonging to the first two groups can besuccessfully used to inhibit undesirable microorganismsin foods, but only nisin is produced industrially and islicensed for use as a food preservative in a partiallypurified form. This review focuses on the production andpurification of class I and class II bacteriocins fromlactic acid bacteria. Bacteriocin production is growthassociated but the yield of bacteriocin per unit biomassis aected by several factors, including the producingstrain, media (carbohydrate and nitrogen sources, cat-ions, etc.) and fermentation conditions (pH, tempera-ture, agitation, aeration and dilution rate in continuousfermentations). Continuous fermentation processes withcell recycle or immobilized cells can result in a dramaticimprovement in productivity over batch fermentations.Several simple recovery processes, based on adsorbingbacteriocin on resins or silica compounds, have beendeveloped and can be used to build integrated produc-tion processes.

Introduction

Mankind has (consciously or unconsciously) exploitedlactic acid bacteria (LAB) for thousands of years in theproduction of fermented foods because of their ability toproduce desirable changes in the taste, flavour and tex-ture and to inhibit pathogenic and spoilage microor-

ganisms. The inhibitory activity of LAB is due to pHdecrease, competition for substrates and to a variety ofantimicrobial compounds, including bacteriocins. Bac-teriocins have been defined as extracellularly releasedprimary or modified products of bacterial ribosomalsynthesis, which can have a relatively narrow spectrumof bactericidal activity, characterized by inclusion of atleast some strains of the same species as the producerbacterium and against which the producer strain hassome mechanism(s) of specific protection (Jack et al.1995). The discovery of nisin, the first bacteriocin usedon a commercial scale as a food preservative, dates backto the first half of this century (Rogers 1928; De Vuystand Vandamme 1994) but research on bacteriocins ofLAB has expanded in the last 15 years, prompted bytheir potential application as natural food preservativesand/or as food grade markers for the development ofcloning vectors (Klaenhammer 1988). Bacteriocins haverecently been grouped into three classes (class I: lanti-biotics; class II: small heat-stable non-lantibiotics; classIII: large heat-labile proteins) on the basis of the se-quence of the mature peptides and prepeptides (Neset al. 1996; Table 1).

Nisin (class I) is licensed for use in foods in more than40 countries and has been produced industrially since1953 (De Vuyst and Vandamme 1994). The largestproducer is Aplin & Barrett Ltd (Dorset, England, asubsidiary of Cultor Food Science, Ardsley, N.Y.,USA), which markets a standardized (1 106 IU g)1)preparation of nisin (Nisaplin). The market price forNisaplin is 810 DM kg)1. In the current industrial pro-cess, pasteurized milk plus added yeast extract is treatedwith a protease and used as a substrate in batch fer-mentation at controlled pH and temperature. An ex-traction process, including a spray-drying step, followsand the resulting powder is standardized with NaCl to1 106 IU g)1. Further purification is costly and notcommercially viable for food use (Trigg, Aplin andBarrett, personal communication). A standardized(minimum 0.9 106 IU g)1) nisin preparation (Chrisin)is also marketed by Chr Hansen (Hrsholm, Denmark).

Appl Microbiol Biotechnol (1999) 52: 628638 Springer-Verlag 1999

E. Parente (&) A. RicciardiDipartimento di Biologia, Difesa e Biotecnologie Agro-Forestali,Universita` della Basilicata, Via Anzio, 10, I-85100 Potenza, Italye-mail: parente@unibas.itTel.: +39-0971-202432Fax: +39-0971-46400

Although several class II bacteriocins have beenshown to be eective in the biopreservation of foods(Stiles 1996), none is licensed or marketed as a food ad-ditive in a partially purified form. Some undefined anti-microbial products (ALTA 2341, Quest BioTechnologyInc., Sarasota, Fla., USA; Microgard, Rhone-Poulenc,Courbevoie, France) are licensed for food use but nodetails are available on their composition and produc-tion, although it has been reported that ALTA 2341 maycontain a class II bacteriocin (Szabo and Cahill 1998).

In this paper the factors aecting the production ofclass I and II bacteriocins from LAB and the techniquesfor their recovery and purification are reviewed. Sincenisin production was recently reviewed by De Vuyst andVandamme (1994), only some aspects of nisin produc-tion and purification will be addressed.

Biosynthesis of bacteriocins from LAB

Bacteriocins are synthesized as pre-propeptides, whichare processed and externalized by dedicated transportmachinery or by the sec-dependent mechanism (Neset al. 1996). Table 1 shows the amino acid sequence ofthe prepeptides of some class I and class II bacteriocins.Cleavage of leader peptides is carried out by specificpeptidases or by a proteolytic domain of the dedicatedABC transporter (Nes et al. 1996), which recognizehighly conserved sequences in the leader peptide. Inaddition, the synthesis of lantibiotics, like nisin, requirespost-translational modification of selected amino acidresidues prior to secretion (De Vuyst and Vandamme1994). In addition to structural and secretion/modifica-tion machinery genes, bacteriocin operons always in-clude genes for specific immunity proteins (Jack et al.1995; Nes et al. 1996) which protect the producer cellsfrom their own bacteriocins. However, little is known ontheir mode of action. The genes nisI and nisFEG havebeen implicated in nisin immunity. nisI is a 32 kDaprotein which is postulated to be lipid modified andextracellularly anchored to the membrane (Kuipers et al.1993). nisF and nisE encode for an ABC transporterwhich shows homology with proteins implicated in re-sistance to subtilin and microcin B17, while the pre-dicted product of nisG is a hydrophobic protein whichmay interact directly with the pore-forming domain ofnisin in a way similar to the immunity proteins ofcolicins (Siegers and Entian 1995). Immunity proteins ofclass II bacteriocins are usually small (51150 aminoacids) and show a low degree of homology, even whenthe bacteriocins are closely related or identical. This maysuggest that they do not interact directly with the bac-teriocins (Nes et al. 1996). However, it has been pointedout that many immunity proteins may actually resembleeach other, at least in structure (Eijsink et al. 1998).Some immunity proteins may integrate in the membraneof the producer strain (Axelsson et al. 1993; Fremauxet al. 1993) but many others do not present transmem-brane helices (Eijsink et al. 1998).

Many bacteriocin operons are regulated by a quo-rum sensing system (for a review see Nes et al. 1996;Kleerezebem et al. 1997). The extracellular accumula-tion of an induction factor (IF) is sensed by a two-component signal transduction system consisting of amembrane- located histidine kinase (HK) which phos-phorylates a response regulator (RR), which in turninteracts with promoters of structural, biosynthetic andregulatory operons and induces gene expression. Thelantibiotic nisin autoregulates its own production(Kuipers et al. 1995; de Ruyter et al. 1996a; Quiaoet al. 1996). In Lactoccus lactis N8 (a nisin Z producer)both nisZBTCIPRK and nisFEG operons are inducedby nisin (Quiao et al. 1996). In nisin A producers, twoinducible promoters are located upstream of nisA andnisF; a third promoter is located upstream of nisR(Kuipers et al. 1995; de Ruyter et al. 1996a). In manyclass II bacteriocins, the IF, HK and RR are organizedin an autoinducible regulatory operon (Nes et al. 1996)and the IF is a bacteriocin-like peptide which mayhave no inhibitory activity. However, it has recentlybeen found that the IF of the plantaricin operons inLactobacillus plantarum C11, PlnA, has indeed bacte-riocin-like activity (Anderssen et al. 1998) and thatbacteriocin operons in Carnobacterium piscicola LV17Bcan be induced by carnobacteriocin CB2 and by abacteriocin-like IF (CnbS) whose gene is organized inan operon with cnbK (HK) and cnbR (RR) (Quadriet al. 1997).

Induction is not the only factor aecting the expres-sion of bacteriocin operons: carbon source regulationand the level of cell-adhered nisin have been shown toaect nisin synthesis (De Vuyst and Vandamme 1992;Meghrous et al. 1992) and catabolite repression has beenclaimed to operate in the regulation of plantaricin Cproduction (Barcena et al. 1998). There are also a fewreports of induction of bacteriocin production caused bycells and extracts of sensitive strains (Barefoot et al.1994; Sip et al. 1998).

Kinetics of bacteriocin production

Bacteriocin production in LAB is growth-associated: itusually occurs throughout the growth phase; and ceasesat the end of the exponential phase (or sometimes beforethe end of growth: Parente et al. 1997; Lejeune et al.1998). A decrease of bacteriocin titre usually follows.This may be attributed to adsorption on producer cellsor to degradation by specific or non-specific proteases.To our knowledge, the latter has never been proven,while adsorption to cells occurs for most bacteriocins(see for example Meghrous et al. 1992; Yang et al. 1992;Parente et al. 1994; Parente and Ricciardi 1994a; DeVuyst et al. 1996; Lejeune et al. 1998; Chinachoti et al.1997d). Since adsorption of bacteriocins to cells ismaximal at pH 5.56.5 (Yang et al. 1992) and decreasesat low pH, it is not surprising that no reduction ofbacteriocin titre is sometimes observed in fermentations

629

Table1

Amino

acid

sequence

ofprepeptides

ofsomebacteriocinsproduced

bylactic

acid

bacteria.(Lc.=

Lactococcus;

Lb.=

Lactobacillus;

Pc.=

Pediococcus;

Ec.

=Enterococcus;

Leuc.=

Leuconostoc)

Name

Producerstrain

Leader

peptidesequence

Propeptidesequence

Ref.a

ClassI(lantibiotics)

Nisin

ALc.lactissubsp.lactis

(manystrains)

MSTKDFNLDNVSVSKKDSGASPR

ITSISLCTPGCKTGALMGCNMKTATCHCSIH

VSK

1

Nisin

ZLc.lactissubsp.lactis

(manystrains)

MSTKDFNLDNVSVSKKDSGASPR

ITSISLCTPGCKTGALMGCNMKTATCNCSIH

VSK

2

LactocinS

Lb.sakei(L45,148)

MEKTEKKVLDELSLHASAKMGARDVESSMNAD

STPVLASVAVSMELLPTASVLYSDVAGCFKYSAKHHC

3

ClassIIa(pediocin-likebacteriocinswith

strongantilisterialeect)

PediocinAcH

/PA-1

Pc.acidilactici(PA1,H),

Lb.plantarum(W

HE92)

MKKIEKLTEKEMANIG

GKYYGNGVTCGKHSCSVDWGKATTCIINNGAMAWATGGHQGNHKC

4

SakacinA

Lb.sakei(706),

Lb.curvatus(LTH1174)

MANVKELSNTELQTITGG

ARSYGNGVYCNNKKCWVRNGEATQSIIGGMISGWASGLAGM

5

EnterocinA/1146

Ec.faecium(CTC492,DPC1146)

MKHLKILSIK

ETQLIY

GG

TTHSGKYYGNGVYCTKNKCTVDWAKATTCIA

GMSIG

GFGGAIPGKC

6LeucocinA

Leuc.gelidum(U

AL187)

MMNMKPTESYEQLDNSALEQVVGG

KYYGNGVHCTKSGCSVNWGEAFSAGVHRLANGGNGFW

7

ClassIIb(two-peptidebacteriocins)

LafX

(lactacinF)

Lb.johnsoniiVPI11088

NRWGDTVLSAASGAGTGIK

ACKSGGP

WGMAICGVGGAAIG

GYFGYTHNACA

8LafA

(lactacinF)

Lb.johnsoniiVPI11088

RNNWQTNVGGAVGSAMIG

ATVGGTICGP

VAGAHYLPILWTGVTAATGGFGKIR

K8

ClassIIc(sec-dependentsecreted

bacteriocins)

EnterocinP

Ec.faeciumP13

MRKKLFSALIG

IFGLVVTNFGTKVDA

ATRSYGNGVYCNNSKCWVNWGEAKENIA

GIVISGWASGLAGMGH

9

aReferences

1Buchmanetal.(1988)

4Motlaghetal.(1992)

7Hastingsetal.(1991)

2deVosetal.(1993)

5Axelssonetal.(1993)

8Allisonetal.(1994)

3Skaugen

(1994)

6Aymerichetal.(1996)

9Cintasetal.(1997)

630

without pH control (De Vuyst and Vandamme 1992;Yang and Ray 1994).

There have been only a few attempts to model bac-teriocin production. Parente and Ricciardi (1994a) andParente et al. (1994) modelled separately the productionand the adsorption/degradation of enterocin 1146 andlactococcin 140 using the following equations:

dBdt YB=X dX

dtif t < t0 1

dBdt kD B X if t > t0 2

where X is biomass concentration; B is bacteriocin titre;YB/X is the yield of bacteriocin per unit biomass pro-duced; t is the time at which bacteriocin production ismaximum; and kD the specific bacteriocin degradationrate. To model enterocin 1146 in batch and continuousfermentations (Parente et al. 1997), Eqs. (1) and (2) werecombined and bacteriocin degradation was assumed tobe a function of cell concentration only:

dBdt YB=X dX

dt kD X 3

Lejeune et al. (1998) modelled bacteriocin production byLactobacillus amylovorus DCE471 in batch culturesusing a dierent approach:

dBdt kB dX

dtif X < X 0 4

dBdt k0 B if X > X or S 0 5

where S is substrate concentration; kB (=YB/X) is de-fined as the specific bacteriocin production rate; X is thecell dry mass level at which bacteriocin productionceases; k is the specific bacteriocin degradation rate. kwas found to be constant over a wide range of condi-tions while both X and kB varied significantly.

The models described by Eqs. (1)(5) are empiricaland do not take induction into account, but this mayhave no consequence under practical conditions. In fact,even if the expression of nis operons is tightly regulatedand no transcription is obtained in the absence of nisin,the level needed for full induction in Lc. lactis is very low(0.50 lg l)1 = 0.02 IU ml)1; de Ruyter et al. 1996b). InCb. piscicola LV17, the use of very low inocula (

riocin production has been observed for nisin (Kim et al.1997) and amylovoryn L471 (De Vuyst et al. 1996; Le-jeune et al. 1998): In an elegant study, Kim et al. (1997)demonstrated that the ceiling for nisin production isaected by both nutrient availability and nisin inhibi-tion.

Media

Bacteriocin production is deeply aected by type andlevel of the carbon, nitrogen and phosphate sources,cations, surfactants and inhibitors.

Bacteriocins can be produced from media containingdierent carbohydrate sources. Nisin Z can be producedfrom glucose, sucrose and xylose by Lc. lactis IO-1(Matsusaki et al. 1996; Chinachoti et al. 1997b, d) butbetter results were obtained with glucose (4000 IU ml)1)compared to xylose (3000 IU ml)1). Glucose, followedby sucrose, xylose and galactose were the best carbonsources for the production of pediocin AcH in an un-buered medium (Biswas et al. 1991). Sucrose wasfound to be a better carbon source than glucose forenterocin 1146 production; fructose or lactose result incomparable levels of biomass but low levels of bacte-riocin (Parente and Ricciardi 1994b). Catabolite re-pression has been used to explain the production ofplantaricin C at higher dilution rates (0.10.12 h)1) withsucrose and fructose compared to glucose (0.055 h)1) incontinuous cultures (Barcena et al. 1998).

Both YB/X and final bacteriocin concentration areaected by the initial carbohydrate concentration inbatch fermentations at controlled pH. Using Lc. lactissubsp. lactis NIZO22186 in sucrose media (De Vuystand Vandamme 1992), maximum nisin titres were ob-tained with 30 g l)1 sucrose. Even if more biomass wasproduced, YB/X decreased from 19.1 to 10.9 mg g

)1 assucrose concentration increased from 10 to 40 g l)1.This was explained by carbon source regulation of thesynthesis or activity of prenisin-modifying enzymes.Substrate and product inhibition of growth resultedin reduced enterocin 1146 production rate at high S0(Parente et al. 1997). Maximum amylovorin L471activity did not increase when glucose was increasedfrom 20 to 60 g l)1, but bacteriocin degradation startedsooner.

Since LAB are nutritionally fastidious microorgan-isms, growth and bacteriocin production are often lim-ited by organic nitrogen sources rather than by thecarbon substrate. Kim et al. (1997) found that maxi-mum nisin concentration increased with increasing or-ganic nitrogen content. However, at any given specificgrowth rate, the specific nisin production rate decreasedwith the increasing amount of nitrogen source. The typeof nitrogen source also aects bacteriocin production.De Vuyst and Vandamme (1993) compared the eect oforganic nitrogen sources (at 10 g l)1) on nisin produc-tion in a complex medium: best results (2500 IU ml)1,with YB/X 64 mg g

)1) were obtained with cotton-seed

meal but high nisin yields (>2000 IU ml)1) were alsoobtained with yeast extract and fish meal. Lc. lactissubsp. lactis ATCC11454 produced 1.5 times more nisinin a filtered stillage-based medium (a raw nitrogensource) compared to LTB broth (Vant Hul and Gib-bons 1996). The dierential eect of nitrogen sourceswas confirmed by factorial experiments for enterocin1146 (Parente and Hill 1992; Parente and Ricciardi1994b) and lactocin D (Parente and Hill 1992).

Both anions (phosphate) and cations (Mg2+, Ca2+)aect bacteriocin production, but their eect may bestrain specific. Inorganic phosphate improved nisinproduction with Lc. lactis subsp. lactic NIZO22186 (DeVuyst and Vandamme 1993): best results (3500 IU ml)1)were obtained in batch fermentations at pH 6.8 with50 g l)1 K2HPO4. However, the stimulatory eect ofphosphate was not confirmed for strain IO-1 (Matsusakiet al. 1996). Mg2+ has been shown to increase pediocinAcH production (Biswas et al. 1991), to improve nisinproduction and significantly decrease cell adhered nisinin Lc. lactis subsp. lactis ATCC11454 (Meghrous et al.1992) but it did not improve nisin production with strainIO-1 (Matsusaki et al. 1996). Adding 0.1 mol l)1 CaCl2resulted in increased maximum nisin Z concentrationand specific production rate, but did not aect growth orlactate production from xylose and glucose in batchfermentations at controlled pH (Matsusaki et al. 1996;Chinachoti et al. 1997b). This was explained by Ca2+

activation of increased immunity of the leader peptidase(NisP) (by protecting the integrity of the cytoplasmicmembrane) or by nisin Z displacement from the cellsurface.

Tween 80 has been found to stimulate the productionof some bacteriocins (Parente and Hill 1992; Daba et al.1993; Matsusaki et al. 1996). However, Tween maysimply have the eect of preventing bacteriocin ad-sorption on polypropylene and glass surfaces (Joostenand Nunez 1995) thus increasing apparent bacteriocintitres. Adding ethanol (1% v/v) improved lactocin S(Mrtvedt-Abildgaard et al. 1995) and amylovorin L471(De Vuyst et al. 1996) production. The stimulatory eectwas attributed to influences on gene expression, pre-vention of bacteriocin aggregation and to increased YB/Xunder stress conditions.

Eect of fermentation conditions

Since pH control improves the growth of LAB, it alsoresults in improved bacteriocin production. However,the optimal pH for bacteriocin production is usually5.56.0 (Meghrous et al. 1992; Kaiser and Montville1993; Parente et al. 1994; Parente and Ricciardi 1994a;Matsusaki et al. 1996; Chinachoti et al. 1997b), oftenlower than the optimal pH for growth. For enterocin1146 (Parente and Ricciardi 1994a) the optimal pH is aresult of higher YB/X and lower kD. A few bacteriocinsare produced only at low pH (5.0) (Biswas et al. 1991;Yang and Ray 1994; Mrtvedt-Albilgaard et al. 1995;

632

Tahara et al. 1996; Barcena et al. 1998). For pediocinAcH this was attributed to the low pH needed for post-translational processing of the bacteriocin. However,this eect may be strain or species-dependent sincepediocin AcH is produced at pH 6.0 by Lb. plantarumWHE2 (Ennahar et al. 1996). Optimal pH may also beaected by the culture medium: nisin Z production withstrain IO-1 was optimal at pH 6.0 in xylose media(Chinachoti et al. 1997b) and at pH 5.5 in glucose media(Matsusaki et al. 1996).

Growth at optimal temperature usually results inoptimal bacteriocin production (Meghrous et al. 1992;Daba et al. 1993; Matsusaki et al. 1996; Chinachotiet al. 1997b; Lejeune et al. 1998) but temperature stressand growth at sub-optimal temperature may result in anincrease of YB/X (Lejeune et al. 1998).

Some LAB produce more than one bacteriocin. In-terestingly, the optimal pH and temperature for theproduction of the two bacteriocins of Leuconostoc me-senteroides subsp. mesenteroides FR52 (mesenterocin52 A and 52B; Krier et al. 1998) were dierent, thusallowing manipulation of the ratio of the two bacterio-cins by changing growth conditions.

Agitation and aeration aect bacteriocin produc-tion. In fermentations at pH 5.5 using glucose mediawith Lc. lactis IO-1, maximum nisin Z concentration(3940 IU ml)1) and YB/X (68.5 mg g

)1) were obtained at320 rpm (Chinachoti et al. 1997d) and only a small de-crease of nisin concentration and yield (3410 IU ml)1,YB/X 68.5 mg g

)1) was obtained at 1000 rpm. On theother hand agitation at >540 rpm resulted in inhibitionof growth and nisin production in xylose media. Aera-tion (1040 ml min)1) significantly reduced nisin pro-duction, perhaps because of chemical degradation.Chemical degradation and eects on gene expressionwere also used to explain the suppression of lactocin Sproduction in aerobic conditions (Mrtvedt-Abildgaardet al. 1995). As observed for other stress factors, fer-mentation in the presence of oxygen resulted in an in-creased YB/X for amylovorin L471 (De Vuyst et al.1996), despite lower final bacteriocin titres.

Since bacteriocin production is growth associated,bacteriocin production rates should improve in contin-uous fermentations where high growth rates can bemaintained. Batch and continuous processes for selectedbacteriocins are compared in Table 2. As predicted byEq. (4) a linear relationship is observed between dilutionrate (D) and specific enterocin 1146 production rate(Parente et al. 1997) with 0.10 < D < 0.60 h)1. A sim-ilar relationship can be calculated for bavaricin MN(Kaiser and Montville 1993) and for divercin (Bhugaloo-Vial et al. 1997). On the other hand, plantaricin C pro-duction in continuous culture is obtained only at lowdilution rates (Barcena et al. 1998). A more complexpattern was observed for nisin production in continuousculture with Lc. lactis subsp. lactis ATCC 11454 (Meg-hrous et al. 1992): in both lactose-limited and non-lim-ited cultures, maximum nisin concentration, yield andspecific production rate were obtained at D = 0.25 h)1

and sharply decreased at higher and lower D (Table 2;Meghrous et al. 1992). Both the level of cell-adherednisin and the specific lactose uptake rate aected nisinproduction.

Increase in D results in a decrease of biomass andbacteriocin concentration, a high substrate concentra-tion in the euent, and eventually wash-out of the cul-ture. Therefore, continuous fermentation with cellrecycle or with immobilized cells has been tested by someauthors (Table 2). Continuous fermentation coupledwith cell recycle with Lc. lactis subsp. lactis IFO12007led to an increase in nisin titre and volumetric nisinproductivity compared to batch fermentations (Tanigu-chi et al. 1994). Decrease in nisin productivity at highsubstrate concentration in the feed was attributed tohigh lactic acid concentrations which inhibited growthand nisin production. Continuous culture with cell re-cycle using a ceramic membrane can be used to producenisin Z at high dilution rates (Chinachoti et al. 1997c)with a slight improvement in productivity comparedto batch or continuous fermentation without cell recy-cle. Choice of the membrane type proved to be critical,since a hollow fibre polyolefin membrane completelyadsorbed nisin. Retention of bacteriocin with ceramicmembranes was also observed by Bhugaloo-Vial et al.(1997).

Wan et al. (1995) compared nisin, brevicin 286 andpediocin PO2 production with free or Ca-alginate im-mobilized cells. In repeated batch fermentations, im-mobilized cells produced less bacteriocin than free cellsperhaps because of diusional limitations in alginategels. However, in continuous fermentation at high di-lution rate (3 h1) the titre of bacteriocin was higher(brevicin) or equal to (nisin and pediocin PO2) thoseobtained in batch fermentations, with significantly im-proved productivity. Using Ca-alginate immobilizedcells of Cb. divergens LV41 in a plug flow reactor op-erated at D = 2 h)1 resulted in a dramatic improvementof productivity compared to batch and continuous fer-mentations with free cells (Table 2; Bhugaloo-Vial et al.1997; Boyaval et al. 1998).

Removal of phosphate and addition ofCaCl2 is neededtomaintainCa-alginate gel stability in repeated batch andcontinuous fermentations. Adsorption of cells on a re-generable photo-crosslinked gel (ENTG-3800) resulted innisin titres (up to 3000 IU ml)1) and yields (2.5 106

IU g)1 CDW) higher than from free cells in repeatedbatch fermentations, with excellent support stability(Chinachoti et al. 1997a). In continuous fermentations,immobilization allowed the use of higher D compared tofree cell systems and, although nisin titre was lower thanwith free cells, nisin productivity increasedup to 3.89 105

IU l)1 h)1 when D was increased to 0.3 h)1.

Recovery and purication of bacteriocins

The recovery and purification of bacteriocins on a lab-oratory scale has been recently reviewed (Carolissen-

633

Table2Comparisonofbatchandcontinuousprocessesfortheproductionofselected

bacteriocinsfromlacticacidbacteria.Bbacteriocintitre;YB/Xbacteriocinyieldperunitbiomass;

r B=

volumetricproductivity;Lb.=

Lactobacillus;Ec.=

Enterococcus;Cb.=

Carnobacterium;Lc.=

Lactococcus;Pc.=

Pediococcus;n.a.notavailable;Ddilutionrateh

)1;Ca-alg

calcium

alginateimmobilized

cells;MFmicrofiltration;imm.immobilized

cells(onENTG-3800)

Bacteriocin

Strain

Process

B (106AU

orIU

l)1)

YB/X

(106AU

g)1)

r B (106AU

l)1h

)1)

Reference

BavaricinMN

Lb.bavaricusMN

batch

3.2

n.a.

0.12

Kaiser

andMontville(1993)

cont.D=0.058

6.4

n.a.

0.37

cont.D=0.205

6.4

n.a.

1.31

Enterocin1146

Ec.faeciumDPC1146

batch

2.8

2.0

0.28

Parenteetal.(1997)

cont.D=0.14

3.2

1.9

0.45

cont.D=0.56

1.8

1.9

1.01

Divercin

Cb.divergensV41

batch

100.0

n.a.

2.2

Bhugaloo-Vialetal.(1997)

cont.D=0.03

200.0

78.1

6.1

cont.D=0.2

210.0

78.1

40.0

cont.+Ca-alg

D=2

5.0

n.a.

200.0

cont.+MFD=0.4

2.0

n.a.

0.8

PlantaricinC

Lb.plantarumLL441

cont.D=0.055

3.2

2.1

0.18

Barcenaetal.(1998)

cont.D=0.12

0.4

0.2

0.05

cont.D=0.25

0.1

0.06

0.03

Brevicin

286

Lb.brevisVB286

batch

24.8

n.a.

1.6

Wanetal.(1995)

cont.Ca-alg

D=3

38.4

n.a.

115.2

Nisin

Lc.lactisAFISC2011

cont.Ca-alg

D=3

0.13

n.a.

0.38

PediocinPO2

Pc.acidilacticiPO2

cont.Ca-alg

D=3

0.25

n.a.

0.75

Nisin

Lc.lactisATCC11454

cont.D=0.1

0.08

0.043

0.008

Meghrousetal.(1992)

cont.D=0.25

0.18

0.097

0.044

cont.D=0.4

0.06

0.033

0.024

Nisin

Lc.lactisIFO12007

batch

0.12

0.035

0.019

Taniguchietal.(1994)

cont.+

MFD=0.5

0.15

n.a.

0.073

cont.+

MF+BD=0.5

0.14

n.a.

0.071

Nisin

Lc.lactisIO

-1batch

3.15

2.25

0.263

Matsusakietal.(1996)

cont.D=0.1

2.81

2.14

0.281

Chinachotietal.(1997c)

cont.+imm.D=0.1

2.16

n.a.

0.216

cont.+imm.D=0.3

1.30

n.a.

0.389

cont.+MFD=0.1

2.75

n.a.

0.275

cont.+MFD=0.3

2.00

n.a.

0.60

batch+MF+SepPakC8

2.4(+

1.4)

2.50

0.33

634

Mackay et al. 1997). The cationic and hydrophobic na-ture of bacteriocins is used for their recovery fromcomplex fermentation broths which contain high levelsof peptides (1030 g l)1 compared to a bacteriocinconcentration of 10100 mg l)1). Laboratory purifica-tion protocols usually include an ammonium sulphateprecipitation step, followed by various combinations ofion-exchange and hydrophobic interaction chromatog-raphy, with a final RP-HPLC purification step. Chro-matography may be replaced by preparative isoelectricfocusing (Venema et al. 1997) and immunoanitychromatography has been reported to allow one-steppurification of nisin A (Suarez et al. 1997). Althoughthese procedures may provide excellent results in termsof yield and purification (Cintas et al. 1998), they areunsuitable for large scale bacteriocin recovery and pu-rification.

Several protocols based on adsorption/desorption oron phase partitioning have been developed for largescale recovery and purification of bacteriocins.

Bacteriocins can be recovered by adsorption onproducer cells at pH 6.06.5, followed by cell separationand desorption at pH 2.0 and 0.1 mol l)1 NaCl. Thismethod was very eective for pediocin AcH, nisin, sa-kacin A and leuconocin Lcm1 (Yang et al. 1992) butrecovery may be limited for other strain/bacteriocincombinations (Daba et al. 1994). Vortex flow filtrationsystems may replace centrifugation (Vant Hul andGibbons 1996) and may be more amenable to large scalerecovery of cells. However, when bacteriocin concen-tration is very high, the ability of cells to adsorb nisincan be exceeded, and recovery of bacteriocin with thecell fraction may only be partial.

Bacteriocins can be adsorbed on HIC and cationexchange resins. The lantibiotics nisin and carnocinUI49 have been purified nearly to homogeneity with asimple two step protocol based on adsorption on HICand cation exchange resins (Stoels et al. 1993). Re-covery was higher than 100% (attributed to removal ofinterferences in the bioassay due to the purification) andpurification was 245 and 60fold for carnocin and nisin,respectively. Chinachoti et al. (1997c) found that Sep-Pak C8 cartridges were the best adsorbent for nisin Z.SepPak C8 were used in an integrated fermentationsystem with a batch fermentation coupled to microfil-tration; permeate was circulated on the cartridge andreturned to the fermenter: growth of the producer strainwas improved and a higher nisin productivity was ob-tained (Table 2).

Ingestible porous silica compounds can also be usedto adsorb bacteriocins (nisin, pediocin PO2, brevicin 286and piscicolin 126) from fermentation broths (Wan et al.1996). Best results were obtained with Micro-Cel E.Adsorbed bacteriocins were still active on target or-ganisms (Wan et al. 1996) and could be desorbed with0.1% sodium dodecyl sulphate (Coventry et al. 1996),obtaining 110130 fold purification. Unfortunately, re-moval of SDS (by cold precipitation) was only partial(6070%).

Boyaval et al. (1998) developed a simple, two-steppurification system based on detergent (Triton 114),phase partitioning and adsorption/desorption on a cat-ion exchange resin. Upon addition of 2% detergent tosupernatants, divercin V41 accumulated in the detergentphase and was recovered with a purity >95% afteradsorption on a cation exchange resin, washing andelution with NaCl 0.7 mol l)1. The method was claimedto be eective in the recovery of mesenterocin Y105 andnisin.

Conclusions

Although many bacteriocins produced by LAB havebeen shown to be very eective against pathogenic andspoilage microorganisms both in vitro and in vivo onlynisin is currently licensed for food use in a partiallypurified form. There might be several reasons for this.Although some class II bacteriocins may be more ef-fective than nisin against some food-borne pathogens(such as Staphylococcus aureus and Listeria monocyto-genes, Cintas et al. 1998; Eijsink et al. 1998), definitiveevidence on their eectiveness and stability in foods isstill scarce (Stiles 1996). In some cases it may be simplerand cheaper to use bacteriocin-containing food ingre-dients (Zottola et al. 1994) or to produce the bacteriocinin situ: protective cultures of Lb. plantarum WHE92(Ennahar et al. 1996, ALC 01 Visbyvac, from Wiesby,Niebull, Germany) and Lb. sakei BJ-33 (Andersen 1995,Bactoferm B-2, from Hansen, Hrsholm, Denmark) aremarketed for use in cheese and meat products, respec-tively. However, in several food products the growth ofLAB may be undesirable and direct addition of bacte-riocins may be a viable alternative. This would requireapproval by regulatory agencies (Fields 1996): althoughLAB are GRAS organisms and bacteriocins are likely tobe present in all natural, fermented foods (Stiles 1996).Although bacteriocin containing products (ALTA 2341)or bacteriocinogenic cultures (see above) are alreadybeing marketed, the cost of approval of bacteriocins fordirect use as food additives may not be justified by theassets.

Low yields and high costs may be further bottleneckslimiting the commercial production of bacteriocins otherthan nisin. While there is no information on nisin yieldwith industrial strains, the maximum yield reported fornisin (1 mg=40,000 IU) is currently around 100 mg l)1

(Chinachoti et al. 1997b). No figures are available forclass II bacteriocins, although in a recent paper (Cintaset al. 1998) yields of 0.10.25 mg l)1 were reported.Further improvement in bacteriocin yield and produc-tivity seem possible: by careful optimization the yield ofgallidermin has been raised from 10 to 720 mg l)1 (Jacket al. 1995). Although low cost substrates can be suc-cessfully used as carbon (for example whey permeate) ornitrogen (cotton seed meal, filtered stillage) sources, theneed for rich, complex media increases costs and com-plicates bacteriocin recovery; and the choice of ingredi-

635

ents should be carefully optimized. The advantages ofintegrated production and recovery processes (batch orcontinuous fermentation with cell recycle and continu-ous adsorption of bacteriocins on resin) similar to thosedeveloped for the lantibiotic epidermin produced byStaphylococcus epidermidis (Horner et al. 1989) havebeen demonstrated for nisin (Chinachoti et al. 1997c).Appropriate exploitation of both cell biomass and bac-teriocin can result in further improvement of the prof-itability of the process: biomasses of the pediocin AcHproducer Lb. plantarum WHE92 are already used asprotective cultures in cheese.

Genetic manipulations may improve the value ofexisting bacteriocins or allow their production in heter-ologous hosts. Improvement of nisin solubility and sta-bility can be obtained by protein engineering:replacement of Asn-27 or His-31 with lysine in nisin Zimproves solubility and stability without aecting anti-microbial activity (Rollema et al. 1995). Nisin was ex-pressed in the subtilin producer Bacillus subtilisATCC6633 (which has simpler nutritional requirementsthan LAB) but no data are available on the yields andproductivities (Rintala et al. 1993). Heterologous ex-pression of class II bacteriocins in other bacteriocin-producing strains has been demonstrated for lactacin F(Allison et al. 1995). Finally, increase in the expressionof immunity genes may raise the ceiling for bacterio-cin production (Kim et al. 1998).

References

Allison GE, Fremaux C, Klaenhammer TR (1994) Expansion ofbacteriocin activity and host range upon complementation oftwo peptides encoded within the lactacin F operon. J Bacteriol176: 22352241

Allison GE, Worobo RW, Stiles ME, Klaenhammer TR (1995)Heterologous expression of the lactacin F peptides by Carno-bacterium piscicola LV17. Appl Environ Microbiol 61: 13711377

Andersen L (1995) Biopreservation with FloraCarn L-2.Fleischwirtschaft 75: 13271329

Anderssen EL, Bao Diep D, Nes IF, Eijsink VGH, Nissen-Meyer J(1998) Antagonistic activity of Lactobacillus plantarum C11:two new two-peptide bacteriocins, plantaricin EF and JK, andthe induction factor plantaricin A. Appl Environ Microbiol 64:22692272

Axelsson L, Holck A, Birekland S-E, Aukrust T, Bloom H (1993)Cloning and nucleotide sequence of a gene from Lactobacillussake LB706 necessary for sakacin A production and immunity.Appl Environ Microbiol 59: 28682875

Aymerich T, Holo H, Havarstein LS, Hugas M, Garriga M, Nes IF(1996) Biochemical and genetic characterization of enterocin Afrom Enterococcus faecium, a new antilisterial bacteriocin in thepediocin family of bacteriocins. Appl Environ Microbiol 62:16761682

Barcena BJM, Sineriz F, Gonzalez de Llano D, Rodrguez A,Suarez JE (1998) Chemostat production of plantaricin C byLactobacillus plantarum LL41. Appl Environ Microbiol 64:35123514

Barefoot SF, Chen Y-R, Hughes TA, Bodine AB, Shearer MY,Hughes MD (1994) Identification and purification of a pro-tein that induces production of the Lactobacillus acidophilusbacteriocin lactacin B. Appl Environ Microbiol 60: 35223528

Bhugaloo-Vial P, Grajek W, Dousset X, Boyaval P (1997) Con-tinuous bacteriocin production with high cell density bioreac-tors. Enzyme Microb Technol 21: 450457

Bhunia AK, Bhowmik TK, Johnson MG (1994) Determination ofbacteriocinencoding plasmids of Pediococcus acidilacticistrains by Southern hybridization. Lett Appl Microbiol 18:168170

Biswas SR, Ray P, Johnson MC, Ray B (1991) Influence of growthconditions on the production of a bacteriocin, pediocin AcH, byPediococcus acidilactici H. Appl Environ Microbiol 57: 12651267

Boyaval P, Bhugaloo-Vial P, Dues F, Metivier A, Dousset X,Marion D (1998) Reacteurs a` haute densites cellulaires pour laproduction de solutions concentrees de bacteriocines. Lait 78:129133

Buchman GW, Banerjee S, Hansen JN (1988) Structure, expressionand evolution of a gene encoding the precursor of nisin, a smallprotein antibiotic. J Biol Chem 263: 1626016266

Carolissen-Mackay V, Arendse G, Hastings JW (1997) Purificationof bacteriocins of lactic acid bacteria: problems and pointers.Int J Food Microbiol 34: 116

Chinachoti N, Sonomoto K, Ishikazi A (1997a) Immobilizationof Lactococcus lactis IO-1 for nisin production. J Fac AgricKyushu Univ 42: 151169

Chinachoti N, Matsusaki H, Sonomoto K, Ishikazi A (1997b)Utilization of xylose as an alternative carbon source for nisin Zproduction by Lactococcus lactis IO-1. J Fac Agric KyushuUniv 42: 171181

Chinachoti N, Endo N, Sonomoto K, Ishikazi A (1997c) Biore-actor systems for ecient production and separation of nisin Zusing Lactococcus lactis IO-1. J Fac Agric Kyushu Univ 43:421436

Chinachoti N, Zaima T, Matsusaki H, Sonomoto K, Ishikazi A(1997d) Relationship between nisin Z fermentative productionand aeration condition using Lactococcus lactis IO-1. J FacAgric Kyushu Univ 43: 437448

Cintas LM, Casaus P, Havarstein LS, Hernandez PE, Nes IF(1997) Biochemical and genetic characterization of enterocin P,a novel sec-dependent bacteriocin from Enterococcus faeciumP13 with a broad antimicrobial spectrum. Appl Environ Mi-crobiol 63: 43214330

Cintas LM, Casaus P, Fernandez MF, Hernandez PE (1998)Comparative antimicrobial activity of enterocin L50, pediocinPA-1, nisin A and lactocin S against spoilage and foodbornepathogenic bacteria. Food Microbiol 15: 289298

Coventry MJ, Gordon JB, Alexander M, Hickey MW, Wan J(1996) A food grade process for isolation and partial purifica-tion of bacteriocins of lactic acid bacteria that uses diatomitecalcium silicate. Appl Environ Microbiol 62: 17641769

Daba H, Lacroix C, Huang J, Simard R (1993) Influence of growthconditions on production and activity of mesenterocin 5 by astrain of Leuconostoc mesenteroides. Appl Microbiol Biotechnol39: 166173

Daba H, Lacroix C, Huang J, Simard RE, Lemieux L (1994)Simple method of purification and sequencing of a bacteriocinproduced by Pediococcus acidilactici UL5. J Appl Bacteriol 77:682688

De Vuyst L (1994) Nisin production variability between naturalLactococcus lactis subsp. lactis strains. Biotechnol Lett 16: 287292

De Vuyst L, Vandamme EJ (1992) Influence of the carbon sourceon nisin production in Lactococcus lactis subsp. lactis batchfermentation. J Gen Microbiol 138: 571578

De Vuyst L, Vandamme EJ (1993) Influence of the phosphorus andnitrogen source on nisin production in Lactococcus lactis subsp.lactis batch fermentations using a complex medium. ApplMicrobiol Biotechnol 40: 1722

De Vuyst L, Vandamme EJ (1994) Nisin, a lantibiotic produced byLactococcus lactis subsp. lactis: properties, biosynthesis, fer-mentation and applications. In: De Vuyst L, Vandamme EJ(eds) Bacteriocins of lactic acid bacteria. Blackie, London, pp151221

636

De Vuyst L, Callewaert R, Crabbe K (1996) Primary metabolitekinetics of bacteriocin biosynthesis by Lactobacillus amy-lovorus and evidence for stimulation of bacteriocin productionunder unfavourable growth conditions. Microbiology 142:817827

Eijsink VG, Skeie M, Middelhoven PH, Brurberg MB, Nes IF(1998) Comparative studies of Class IIa bacteriocins of lacticacid bacteria. Appl Environ Microbiol 64: 32753281

Ennahar S, Aoude-Werner D, Sorokine O, Van Dorsselaer A,Bringel F, Hubert J-C, Hasselmann C (1996) Production ofpediocin AcH by Lactobacillus plantarumWHE92 isolated fromcheese. Appl Environ Microbiol 62: 43814387

Fields FO (1996) Use of bacteriocins in foods: regulatory consid-erations. J Food Prot 1996 [Suppl]: 8286

Fremaux C, Ahn C, Klaenhammer TR (1993) Molecular analysisof the lactacin F operon. Appl Environ Microbiol. 59: 39063915

Hastings JW, Sailer M, Johnson K, Ray L, Vederas JC, Stiles ME(1991) Characterization of leucocin A-UAL187 and cloning ofthe bacteriocin gene from Leuconostoc gelidum. J Bacteriol 173:74917500

Horner T, Zahner H, Kellner R, Jung G (1989) Fermentation andisolation of epidermin, a lanthionine containing polypeptideantibiotic from Staphylococcus epidermidis. Appl MicrobiolBiotechnol 30: 219225

Jack RW, Tagg JR, Ray B (1995) Bacteriocins of Gram-positivebacteria. Microbiol Rev 59: 171200

Joosten HMLJ, Nunez M (1995) Adsorption of nisin and enterocin4 to polypropylene and glass surfacec and its prevention byTween 80. Lett Appl Microbiol 21: 389392

Kaiser AL, Montville TJ (1993) The influence of pH and growthrate on the production of the bacteriocin, bavaricin MN, inbatch and continuous fermentations. J Appl Bacteriol 75: 536540

Kim WS, Hall RJ, Dunn NW (1997) The eect of nisin concen-tration and nutrient depletion on nisin production of Lac-tococcus lactis. Appl Microbiol Biotechnol 48: 449453

Kim WS, Hall RJ, Dunn NW (1998) Improving nisin productionby increasing nisin immunity/resistance genes in the producerorganism Lactococcus lactis. Appl Microbiol Biotechnol 50:429433

Klaenhammer TR (1988) Bacteriocins of lactic acid bacteria. Bio-chimie 70: 337349

Kleerebezem M, Quadri LEN, Kuipers OP, Vos WM de (1997)Quorum sensing by peptide pheromones and two-componentsignal transduction systems in Gram-positive bacteria. MolMicrobiol 24: 895904

Krier F, Revol-Junelles AM, Germain P (1998) Influence of tem-perature and pH on production of two bacteriocins by Le-uconostoc mesenteroides subsp. mesenteroides FR52 duringbatch fermentation. Appl Microbiol Biotechnol 50: 359363

Kuipers OP, Beerthuyzen MM, Siezen RJ, Vos WM de (1993)Characterization of the nisin gene cluster nisABTCPIR ofLactococcus lactis, requirement for expression of the nisA andnisI genes for development of immunity. Eur J Biochem 216:281291

Kuipers OP, Beerthuizen MM, Ruyter PGGA de, Luesink EJ, VosWM de (1995) Autoregulation of nisin biosynthesis in Lac-tococcus lactis by signal transduction. J Biol Chem 270: 2729927304

Lejeune R, Callewaert R, Crabbe K, De Vuyst L (1998) Modellingthe growth and bacteriocin production by Lactobacillus amy-lovorus DCE 471 in batch cultivation. J Appl Bacteriol 84: 159168

Matsusaki H, Endo N, Sonomoto K, Ishikazi A (1996) Lantibioticnisin Z fermentative production by Lactococcus lactis IO-1:relationship between production of the lantibiotic and lactateand cell growth. Appl Microbiol Biotechnol 45: 3640

Meghrous J, Huot M, Quittelier M, Petitdemange H (1992) Reg-ulation of nisin biosynthesis by continuous cultures and byresting cells of Lactococcus lactis subsp. lactis. Res Microbiol143: 879890

Mrtvedt-Abildgaard C, Nissen-Meyer J, Jelle B, Grenov B,Skaugen M, Nes I (1995) Production and pH-dependent bac-tericidal activity of lactocin S, a lantibiotic from Lactobacillussake. Appl Environ Microbiol 61: 175179

Motlagh AM, Bhunia AK, Szostek F, Hansen TR, Johnson MG,Ray B (1992) Nucleotide and aminoacid sequence of pap-gene(pediocin AcH production) in Pediococcus acidilactici H. LettAppl Microbiol 15: 4548

Nes IF, Bao Diep D, Havarstein LS, Brurberg MB, Eijsink V, HoloH (1996) Biosynthesis of bacteriocins of lactic acid bacteria.Antonie van Leeuwenhoek 70: 113128

Parente E, Hill C (1992) A comparison of factors aecting theproduction of two bacteriocins from lactic acid bacteria. J ApplBacteriol 73: 290298

Parente E, Ricciardi A (1994a) Influence of pH on the productionof enterocin 1146 during batch fermentation. Lett Appl Mi-crobiol 19: 1215

Parente E, Ricciardi A (1994b) Eect of nitrogen and carbohy-drate sources on lactic acid and bacteriocin production byEnterococcus faecium DPC1146. Agro-Industry Hi-Tech 5:3539

Parente E, Ricciardi A, Addario G (1994) Influence of pH ongrowth and bacteriocin production by Lactococcus lactis subsplactis 140NWC during batch fermentation. Appl MicrobiolBiotechnol 41: 388394

Parente E, Brienza C, Ricciardi A, Addario G (1997) Growth andbacteriocin production by Enterococcus faecium DPC1146 inbatch and continuous culture. J Ind Microbiol Biotechnol 18:6267

Quadri LEN, Kleerebezem M, Kuipers OP, Vos WM de, RoyKL, Vederas JC, Stiles ME (1997) Characterization of a lo-cus from Carnobacterium piscicola LV17B involved in bac-teriocin production and immunity: evidence for globalinducermediated transcriptional regulation. J Bacteriol 19:61736171

Quiao M, Ye S, Koponen O, Ra R, Usabiaga M, Immonen T, SarisPEJ (1996) Regulation of the nisin operons in Lactococcus lactisN8. J Appl Bacteriol 80: 626634

Quiao M, Omaetxebarria MJ, Ra R, Oruetxebarria I, Saris PEJ(1997) Isolation of a Lactococcus lactis strain with high resis-tance to nisin and increased nisin production. Biotechnol Lett19: 199202

Rintala H, Graee T, Paulin L, Kalkkinen N, Saris PEJ (1993)Biosynthesis of nisin in the subtilin producer Bacillus subtilisATCC6633. Biotechnol Lett 15: 991996

Rodriguez JM, Cintas LM, Casaus P, Suarez A, Hernandez P(1995) PCR detection of the lactocin S structural gene in bac-teriocin-producing lactobacilli from meat. Appl Environ Mi-crobiol 61: 28022805

Rogers LA (1928) The inhibiting eect of Streptococcus lactis onLactobacillus bulgaricus. J Bacteriol 16: 321325

Rollema HS, Kuipers OP, Both P, Vos WM de, Siezen RJ (1995)Improvement of solubility and stability of the antimicrobialpeptide nisin by protein engineering. Appl Environ Microbiol61: 28732878

Ruyter PGGA de, Kuipers OP, Beerthuyzen MM, Alen-Boerr-trigter I van, Vos WM de (1996a) Functional analysis of pro-moters in the nisin gene cluster of Lactococcus lactis. J Bacteriol178: 34343439

Ruyter PSGGA de, Kuipers OP, Vos WM de (1996b) Con-trolled gene expression systems for Lactococcus lactis withthe food-grade inducer nisin. Appl Environ Microbiol 62:36623667

Saucier L, Poon A, Stiles ME (1995) Induction of bacteriocin inCarnobacterium piscicola LV17. J Appl Bacteriol 78: 684690

Siegers K, Entian K-D (1995) Genes involved in immunity to thelantibiotic nisin produced by Lactococcus lactis 6F3. Appl En-viron Microbiol 61: 10821089

Sip A, Grajek W, Boyaval P (1998) Enhancement of bacteriocinproduction by Carnobacterium divergens AS7 in the presence ofa bacteriocin-sensitive strain Carnobacterium piscicola. Int JFood Microbiol 42: 6369

637

Skaugen M (1994) Lactocin S: structure determination and geneticanalysis. Doctoral thesis, Agricultural University of Norway,As, Norway

Stiles ME (1996) Biopreservation by lactic acid bacteria. Antonievan Leeuwenhoek 70: 331345

Stoels G, Sahl H-G, Gudmundsdottir A (1993) Carnocin UI49, apotential biopreservative produced by Carnobacterium piscico-la: large scale purification and activity against various Gram-positive bacteria including Listeria sp. Int J Food Microbiol 20:199210

Suarez AM, Azcona JI, Rodrguez JM, Sanz B, Hernandez P(1997) One-step purification of nisin A by immunoanitychromatography. Appl Environ Microbiol 63: 49904992

Szabo EA, Cahill ME (1998) The combined eects of modifiedatmosphere, temperature, nisin and ALTA 2341 on the growthof Listeria monocytogenes. Int J Food Microbiol 43: 2131

Tahara T, Oshimura M, Umezawa C, Kanatani K (1996) Isola-tion, partial characterization, and mode of action of acidocinJ1132, a two-component bacteriocin produced by Lactobacil-lus acidophilus JCM 1132. Appl Environ Microbiol 62: 892897

Taniguchi M, Hoshino K, Urasaki H, Fuji M (1994) Continuousproduction of an antibiotic polypeptide (nisin) by Lactococcuslactis using a bioreactor coupled to a microfiltration module.J Ferment Bioeng 77: 704708

Vant Hul JS, Gibbons WR (1996) Concentration and recovery ofthe bacteriocin nisin from Lactococcus lactis subsp. lactis.Biotechnol Appl Biochem 24: 251256

Venema K, Chikindas ML, Seegers JFML, Haandrikman AJ,Leenhouts KJ, Venema G, Kok J (1997) Rapid and ecientpurification method for small, hydrophobic, cationic bacterio-cins: purification of lactococcin B and pediocin PA-1. ApplEnviron Microbiol 63: 305309

Vos WM de, Mulders JWM, Siezen RJ, Hugenholtz J, Kuipers O(1993) Properties of nisin Z and distribution of its gene, nisZ, inLactococcus lactis. Appl Environ Microbiol 59: 213218

Wan J, Hickey MW, Coventry MJ (1995) Continuous productionof bacteriocins, brevicin, nisin and pediocin using calcium al-ginate immobilized bacteria. J Appl Bacteriol 79: 675679

Wan J, Gordon J, Hickey MW, Mawson RF, Coventry MJ (1996)Adsorption of bacteriocins by ingestible silica compounds.J Appl Bacteriol 81: 167173

Yang R, Ray B (1994) Factors influencing production of bacte-riocins by lactic acid bacteria. Food Microbiol 11: 281291

Yang R, Johnson MC, Ray B (1992) Novel method to extract largeamounts of bacteriocins from lactic acid bacteria. Appl EnvironMicrobiol 58: 33553359

Zottola EA, Yezzi TL, Ajaio DB, Roberts RF (1994) Utilization ofCheddar cheese containing nisin as an antimicrobial agent inother foods. Int J Food Microbiol 24: 227238

638