WO1996032482A1 - Bacteriocins - Google Patents

Abstract

The present invention relates to a novel anti-microbial agent, more particularly, a novel bacteriocin with nisin-like properties. The bacteriocin is designated lacticin 3147 and has the following properties: a molecular weight of approximately 2.8 kDa; inhibiting activity against lactococci, lactobacilli, enterococci, bacilli, leuconostocs, pediococci, clostridia, staphylococci and streptococci; sensivity to the proteases trypsin, alpha-chymotrypsin, proteinase K and pronase E but not pepsin; heat-stability; activity at acid pH; and the capability of inhibiting nisin-producing bacterial strains.

Description

"BACTERIOCINS"

The present invention relates to a novel anti-microbial agent, more particularly, a novel bacteriocin with nisin-like properties. Recent years have seen major advances in the development of microbial metabolites with antagonistic activities towards spoilage and pathogenic micro-organisms associated with food Although, there now exists in excess of seven thousand known antibiotic compounds of microbial origin, very few have been evaluated for food use. The danger exists that antibiotic-resistant microorganisms of clinical importance will appear with repeated exposure to antibiotics in food, thus compromising the clinical usefulness of the antibiotics (although, of course, clinically important antibiotics are not used for food applications). Additionally, consumer emphasis is now on minimally processed foods which are natural and preservative free Because of this considerable and justifiable resistance to the use of chemical additives and antibiotics as food preservatives, other biological inhibitors produced by microorganisms are currently being investigated for use in foods Of particular interest are those inhibitory substances produced bv the Lactic Acid Bacteria (LAB) which include hydrogen peroxide, diacetyi. bactenocms. as well as secondary reaction products such as hypothiocyanite. Bactenocms are potentially very attractive natural preservatives as they arc produced as normal by-products of microbial metabolism The LAB are industrially important microorganisms and include the genera Laaococcus. Streptococcus. Pedtocυccus. Leuconυstoc. Laciobctallus and Camυbaciermm. They have been used for the production of fermented foods which have been consumed safely for hundreds of years Given their status as "safe" organisms, they are a particularly suitable source for natural antimicrobials such as bactenocms for use in foods

In 1925. the first prototype bacteriocin was discovered by Gratia et al from a strain of Escherchta colt. Bactenocms are always proteins which can be broad or narrow spectrum and are not lethal to the cells which produce them. Bacteria protect themselves from the lethal effects of their own bactenocms by such mechanisms as post-translational modification or the production of an immunity proteιn(s). In any case, bacteπocins are potent antibacterial substances produced by a large and diverse assortment of species. They fomi a heterogeneous group with respect to their producer, inhibitory spectrum, mode of action and chemical properties There are four distinct classes of LAB bactenocins:-

A. Lantibiotics • small peptides of less than 5 kDa which contain unusual amino acids such as lanthionine. β-methyllanthionine and dehydrated residue, e.g. nisin. lacticin 481. camocin U149 and lactocin S

B. Non-Lanthionine containing Peptides - small peptides of 10 kDa or less and can be subdivided as follows: (i) Listena-actws peptides e.g. Pediocin PA-1 and Sakacin A. (ii) Poration complexes consisting of two prote aceous peptides e.g. Lactacin F (iii) Thiol-activated peptides requiring

1

SUBSTTTUTE SHEET (RULE 26) reduced cysteine residues for activity e.g. Lactococc B.

C. Large Heat-Labile Proteins - larger proteins, generally having a molecular weight greater than

31 kDa e.g. Helveticin V-1829.

D Complex Bactenocms - composed of a protein with one or more chemical moieties which may be of a lipid or carbohydrate nature e.g. Pediociπ SJ-1

Lactococci are widely used as starter cultures in the dairy industry and several strains of dairy species can produce bactenocms L. lactts subspecies can produce diplococcin. lactococcm. lactostrepcins or nisin. Diplococcin and lactococcins are small molecular weight proteins, active against other lactococci while nisin is a (antibiotic with a broad spectrum of activity against many Gram positive bactena.

Nisin is the most extensively characterised bacteriocin of the antimicrobial proteins produced by

LAB and has found widespread application in the food industry Nisin was the first "antibiotic" compound to be used on a commercial scale in the food industry It is used to prevent spore outgrowth and toxin production by Clostndtum botu/tnum in processed cheese and cheese spreads In some countries, it has been used to extend the shelf-life of dairy products and to prevent the spoilage of canned foods by thermophiles. Nisin is a pentacychc.

A lantibiotic and it displays an amphipathic character, with a hydrophobic residue (He) at its N-termmus and a hydrophi c residue (Lys) at its C-termmus. It is a peptide of 34 ammo acids and is inactivated by proteases including chymotrypsin. pancrcatin and subulopeptidase. It is insensitive to carboxypeptidasc A. elastase. ercpsin. pepsin and tnpsin. It contains one lanthiomne residue, four β-methyilanthionines. a dchydroalaninc and a dchydrobutyπne

The thioethcr ammo acids, (lanthiomne and β-methyllanthionme) account for the high sulphur content of nisin The unusual ammo acid residues are thought to be responsible for the important functional properties of nisin e g the associated acid tolerance and thermostable properties of nisin are attributed to the stable thioethcr linkages while the specific bactericidal activity is thought to be due to the very reactive double bonds. Nisin has a molecular mass of 3.5 kDa (Gross & Morell.

1971) and often forms dimers and oligomers. Nisin has a very broad spectrum of activity against Gram positive vegetative bacterial cells. The closely related lactococci are especially sensitive but nisin is also inhibitory to several strains of bacilli and clostridia (particularly their spores), lactobacilli. leuconostocs. micrococci. pediococci. streptococci and actinonycetes. Other sensitive strains include Mycobacteriuni tuberculosis.

Staphylococcus pyogenes. S. aureits, S. epidermidis and Listerta monocylogenes (De Vuyst &.

Vandamme. 1994). Nisin does not display activity against Gram negative bacteria, except for three

Neisseria strains (Mattick __ Hirsch. 1947) and three Flavobacter strains (Ogden & Tubb. 1985). nor does it inhibit yeasts or viruses. However. Salmonella subspecies and other Gram negative bacteria can be made sensitive by using a chelating agent m combination with nisin (Stevens et. al.. 1992) The site of action of nisin appears to be the cytoplasmic membrane The outer membrane of Gram negative bacteπa is thought to exclude the bacteriocin making contact with the

membrane Hence, by incorporating a chelatmg agent such as EDTA with nisin. the structure of the outer membrane undergoes alteration, resulting in destabihzation of the hpopolysacchaπde (LPS) layer with a corresponding increase in cell permeability Binding of the bacteπocin to the membrane leads to the aggregation of similar peptides. thus initiating ohgomeπzation Such aggregates adopt a transmembrane onentation so that the hvdrophobic portion is exposed to the core of the membrane and the hvdrophi c part forms the aqueous channel Membrane insertion, pore formation, (both of which require a transmembrane potential) and subsequent depolarization leads to the efflux of small cellular consutuents and destruction of energy metabolism of the cell (Ruhr & Salil. 1985) Tins results in a deficiency of metabolic intermediates and hence, inhibition of synthesis of DNA. RN A. proteins and polysacchandes There appears to be a separate mechanism for the prevention of .pore outgrowth Unlike vegetative cells, bacterial spores never lyse hen treated ith nism The dt

residues of nisin provide possible covalcnt attachment sites for membrane sulflwdπ I groups Fhese residues appear to have no such role in membrane pore foπnation (Morns et al . 1984)

Nisin is encoded

a conjugative transposon which can insert into either plasmid DNA or chromosomal DNA (Horn et al . 1991 ) Anah sis of the nisin opcron indicates that the nisin structural gene, as well as the genes required for processing and maturation arc clustered over a pohcistronic region exceeding 8 5 kb. (Stecn et al . 1991) Nisin has certain disadvantages for industnal application, one being that resistance to the bacteriocin can frequenth occur For example there is a nisin resistance gene (as opposed to an immunity gene) on die coii'ugative lactococcal plasmid pNP4ϋ Its second disadvantage is that nism-producing strains are generally not good acid producers, arc phage sensitive and are ^■lon-proteolytic and therefore are not efficient cheese starter cultures It is an object of the present invention to provide a bacteriocin suitable for food use. particularh one which has a broad spectrum of inhibition and for which there is no detectable spontaneous resistance thereto A further ob'ect is to provide bacteriocin encoding genc(s) which

be more easily conjugally mobilized than the nis -encoding gene It is also an object to provide bacteriocin encoding gene(s) which may be easily conjugally mobilised along with genes which may be attached to them It is a further object to provide bactenoc -encoding gene(s) which are not linked to lactose catabo sm genes There is a further object to provide bacteriocin-producing strains which are effective cheese starters A further object of the mvenuon is to provide phage resistance genes, particularly such genes linked to bactenocm-encoding genes According to the present invention there is provided a bacteriocin characterised by a molecular weight of approximately 2 8 kDa. mhibiUng activity against lactococci. lactobacilli. enterococci. bacilli, leuconostocs. pediococci. clostridia. straphylococci and streptococci, sensitivity to the proteases trypsin. alpha-chymotrypsin. proteinase and pronase E but not pepsin. heat-stability, activity at acid pH. and the capability of inhibiting msin-producing bacterial strains. The bacteriocin is designated lacticin 147 The bacteπocin lacticin 3147-encodιng gene does not cross-hybridise with the nisin-encσding gene.

The invention also relates to L. lactts DPC3147 strain as deposited at the National Collection of Industnal and Marine Bacteria. Aberdeen. Scotland, on 1 1th April 1995 under the Accession No NCIMB 40716. This strain produces the bacteπocin lacticin 3147 The present invention also provides a 63 kDa plasmid encoding the bacteriocin as defined above. The plasmid may be the plasmid pMRCOl which encodes the novel bacteriocin designated lacticin 3147 and lacticin 3147 immunity genes, as deposited in the National Collection of Industrial and Marine Bacteπa. Aberdeen. Scotland on 1 1th April 1995 under the Accession Number NCIMB 40716 Furthermore, the invention provides isolated genes for the production of. and immunity to. lacticin 3147 as deposited in the plasmid pMRCO 1 above

The present invention also relates to the use of lacticin 3147 or a lacUcm 3147 producing host in the treatment of mastitis in cattle, to prevent clostπdial spoilage in cheese, as a food preservative in pasteuπsed cheeses and cheese spreads, as a shelf-life extender in milk and milk products, in the production of alcoholic

particularh* in the brewing industry, in vegetable fermentations, by incoφoration into canned foods and in meat preseπ'ation. Additional uses include incoφoration into oral healthcare products such as toothpaste and mouthwashes and in cosmetic treatments for acne The bacteriocin may be used against Gram negative bacteria which have been treated w ith chelating agents The invention also relates to the use of a plasmid or a gene(s) encoding the lacticin 3147 bacteriocin to confer bacteriocin producing properties on a host such as a bacterium, particularh' a cheese-starter culture. The invention also provides a host such as a bacterium containing a plasmid or a gene(s) as defined above, encoding lacticin 3147 The invention also provides a method of producing lacticin 3147 comprising cultunng a bacterium containing lacticin 3147-encodιng gene(s) and isolating lacticin 3147 from die culture, and a method of conferring lacticin 31 7-producιng properties on a host such as a bacterium, comprising introducing and expressing in the host a plasmid or gene(s) as defined above encoding lacticin 3147

The invention also relates to a food-grade genetic marker system comprising genes for immunity to lacticin 3147 as encoded by plasmid pMRCOl. T e genetic determinants which encode lacticin 3147 immunity may be introduced into a bacterial strain together with any desirable gene(s) which have been linked to them. A bacterial cell which has received the genes can be selected from the general population of cells by plating on medium containing lacticin 3147. since cells containing the lacticin 3147 immunity genes will be able to grow in this environment. Given that spontaneous resistance to lacticin 3147 occurs at a low frequency in lactococcal strains (undetectable for some strains) this marker system should provide all die advantages of well known antibiotics such as Ampicillin and Eπthromycin with none of the negative clinical associations. Since these 3147 genes have originated from a GRAS organism they are considered to be Food Grade.

The invention also provides phage resistance gene(s). particularly gene(s) encoding total resistance to the small isometric-headed phage 712 and burst size limitation for the prolate-headed phage c2 The phage resistance gene(s) may be encoded by the plasmid pMRCOl. deposited as described above. Further the invention relates to a host such as a bacterium containing a plasmid or genc(s) as defined above encoding phage resistance. Also provided is a method of conferring phage resistance on a host such as a bacterium and particularly a cheese starter culture, comprising introducing and expressing therein a plasmid or gcne(s) as defined above encoding phage resistance The invention also relates to the use of a plasmid or phage resistance gene(s) to confer phage resistance on a host such as a bacteπum. particularly a cheese-starter culture

In a further aspect the invention provides L. locus strain DPC2949 as deposited at the National Collection of Industrial and Marine Bacteria. Aberdeen. Scotland on 1 1th April 1 95 under the Accession No NCIMB 40715. the bacterium-encoding gene(s) thereof and the bacteriocin produced thereby The invention also relates to methods of preventing late gas-blowing or of controlling non-staner lactic acid bacteria in Cheddar cheese production comprising either the addition of L lactts DPC2949 or the bacteriocin produced thereby to the initial cheese starter culture.

The invention also provides bactenocm-encoding genes, bacteriocin immunity genes, phage resistance genes, plasmids and strains substantially homologous to those defined above also encoding bacteriocin production and immunity and phage resistance. Such homologous genes, plasmids and strains arise because of the degeneracy of the genetic code, the possibility of replacing one ammo-acid with another w lt out affecting the funtional characteristics of a protein and the possibility of deleting a non-essential portion of a gene or ammo-acid sequence hile still producing a functional end-product. The invention also provides genes as defined above linked to other DNA sequences. Since the biological activity of the bacteriocin of the present invention is similar to that of nisin. it could be used for similar applications. Unlike Northern Europe, many fermented dairy products in Southern Europe are still being made using traditional methods without the use of commercial starters. One source of these strains in Ireland is Buttermilk plants commonly used by Irish housewives in the sounng of bread for breadmakmg. These Buttermilk plants, also known as kefir grains, are creamy white in colour and resemble cauliflower florets. They are resilient and difficult to break up due to the presence of kefir in. a water soluble polysaccharide produced by the lactobacilli in the plant. They are composed of lactococci. leuconostocs. lactobacilli. yeasts and acetic acid bacteria which are held together in a matrix and are recoverable at the end of the fermentation process as a solid mass. Some isolates produce a bacteπocin which may be important in maintaining the integrity of the grain as it is assumed that all other organisms in the grain are resistant to it (Rea & Cogan. 1994).

The present invention will now be described in greater detail with reference to the accompanying drawings in which. -

Figure 1. Cross-Sensitivity Study a I. lactis DPC3147: b. NCD0496. c. DPC2949. d DPC3220, e. DPC33(1) were grown on GM17 agar plates and subsequently overlaid with A DPC3147 and B. NCD0496. Zones where no growth has occurred indicate inhibition of the indicator due to bactenocιn(s) produced by the producer

Figure 2. Protease Sensitivity Assay. Filtered cell-free bacteriocin solution and enzyme solution (A represents control, no enzyme. B. alpha-chymotrypsin: C. trypsin. D. pronase E and E represents proteinase K) were spotted 1 cm apart on GM 17 agar plates and subsequently overlaid with the indicator /., lactis strain HP.

Figure 3 Polymerase Chain Reacuon. The 166 bp PCR amplified fragment from gcnomic DNA isolated from L. lactis NCD0496 is indicated in lane 1. No amplified products arc evident where DNA isolated from strains DPC3 I47. MG1614. and DPC2949 were used as templates Lane 3 corresponds to DNA molecular weight markers. Figure 4 DNA Hybridization. Hind III restricted genome DNA from L lactis DPC3147.

MG1614. and NCD0496 is shown in lanes 1.2 and 3 respectively Lane 4 corresponds to DNA molecular weight markers The nts A gene probe, hybridized to a 3.5 kb fragment on the NCD0496 genome (lane 3') No hybridizing DNA is observed in DNA isolated from DPC3 147 (lane I ) or from MG1614 (lane 2) Figure 5 Growth of L. iactts DPC3147 in GM 17 and TYT30 media.

Figure 6 Effect of heating at 60"C to I2FC for 10 minutes on the stability of lacticin 3147 at pH5. pH7. pH9. 100% activity is the activity at pH 5. 7 and 9 with no heating Figure 7 Overlaid SDS-PAGE gel. Preparations of lacticin 3147 (A. lanes 2. 3 and 4. B. lane 2) and of lactococcm A (B. lane 3) were electrophoresed on a 10% SDS-PAGE gel. This was overlaid with an active culture of L. lactis HP. Inhibition of the indicator is seen as zones in the indicator lawn. Molecular weight markers (2.5 kDa to 17 kDa) are indicated in A. lane 1. Figure 8. Analysis of the plasmid complements of a putative transconjugant (lane 2) obtained from a mating between L. lactis DPC3147 (lane 1 ) and the piasmid-free strain MG 1614. The transconjugants contain a 63 kDa plasmid acquired from the DPC3147 donor which is not evident in the plasmid-free MG 1614 strain.

Figure 9 pH profiles during Cheddar cheese manufacture using die DPC3147 strain as cheese starter.

Figure 10. Growth of starter and NSLAB during πpening of Cheddar cheese using the DPC3147 strain as cheese starter.

Figure 11. pH profiles during Cheddar cheese manufacture using a transconjugant of strain 303 which produces lacticin 3147 as cheese starter.

Figure 12. Growth of starter and NSLAB during ripening of Cheddar cheese using a transconjugant of strain 303 which produces lacticin 3147 as cheese starter.

Figure 13 A. Residual Lacticin 3147 activities in Cheddar cheese made with bacteπocin-producing strains (DPC3147, DPC3256 and DPC3204) sampled after 2 weeks. 4 months and 6 months. The commercial strainDPC4268 was used as the control starter.

Figure 13B Residual lacticin 3147 activity in Cheddar cheese made with the bacteπocin-producing transconjugant DPC4275 Agaui. DPC4268 was used as the control

Figure 14A AnUmicrobial effectiveness of lacticin 3147 requires the presence of Tween AU/i.il) lias been added. Figure 14B. Effect of adding lacticin 3147 (10.240 AU/ml) to stationaπ cells of Streptococcus dλ.galactiae.

Materials and Methods A complete list of strains, their growth media and temperature of incubation used throughout this study is included in Table 1 Both L lactis DPC3147 and DPC3220 were isolated from kefir grains. whereas L. lactis strains NCD0496 and NCD0497. known nisin producers, were obtained from the

National Collection of Dairy Organisms (NCDO), National Institute for Research and Dairying.

Shinfield. Reading. Berkshire. RG2 9AT. England. The source of each of the indicator organisms tested are also listed in Table 1 L locus cells were routinely propagated at 30T in M 17 (Difco laboratories. Detroit. USA) supplemented with 0.5% glucose or lactose Other media used in this itudy. as indicated in Table 1. include MRS (Difco Laboratories). BHl (Oxoid Ltd . Hampshire.

England). TYP (Tryptone 16 g/1. Yeast Extract 16 g/1, NaCl 2.5 g/l. K₂HP0₄ 2.5 g/l). TYT30

(Tryptone 2.5 g/1. Yeast Extract 5 g/1. Tween 20 1 g/1, Glucose 10 g/l. β-gh ccrophosphate 19 g/1.

MgS0₄7H_:0. 0.25 g/1, MnS0 4H₂0 0.05 g/1), TSA, (Becton Dickinson Microbiology Systems.

Maryland. USA), Baird-Parker media (Merck. Darmstadt. Geπnany). RSM (Reconstituted Skim Milk) and pasteurized whole m lk.

To test the sensitivity of a strain to DPC3147, to L. lactis NCD0496 and to L. lactis DPC3220. 10 μl aliquots of a fresh overnight culture of each were first spotted on GM17 agar plates and incubated overnight at 30°C. These plates were then gently overlaid with 3 ml of soft agar seeded with 100 μl of the test strain, so as not to disturb the grown producer. The sensitivity of a strain to each bacteriocin producer was scored according to the diameter of the zone of inhibition surrounding that producer. The scoring system adopted is as follows: 0 to 1 mm (-). 1 to 5 mm (+). 5 to 15 mm (++).

7

SUBSTITUTE SHEtΞT (RULE 26) over 15 mm (++^■+) All strains were stocked in 40% ghcerol and stored at -80"C Working cultures were stored at 5°C and transferred periodically

Plates containing bacteπocin were prepared as follows A cell-free, sterile bacteriocin solution was obtained from an overnight culture of L lactis DPC3147 grown in GM 17 by first centπfuging at 8000g (Sorvall RC-5B) for 10 mm The resulting supernatant was then sterilized by filtering with Mil pore HVLP filters (0 45 um) and tempered to 45°C An equal volume was then added to double strength GM 17 and the plates were poured Where needed, streptomvein was added to agar at a concentration of 500 μg/ml To identify' lactose metabolizing bacteria. Lactose Indicator Agar (LI A) was used (Tryptone 20 g/1. Yeast Extract 5 g/1, Lactose 10 g/1. NaCl 4 g/1. Sodium Acetate-anhydrous 1 5 g/l. Ascorbic Acid 0 5 g/1. Gelatin 2 5 g/1. Bromocresol Puφle (0 004%). Agar Bacteriological 15 g/1 Sucrose Indicator Agar plates were prepared similarly except that sucrose was substituted for lactose as the carbon source Measurement of Bacteriocin Activity:- Bacteriocin ac it*. was estimated using an agar well-diffusion assav essenuallv as described in Parente and Hill ( 1992) In the agar well diffusion assay, molten agar (GM17 for lactococci) was cooled to 48^UC and inoculated with overnight cultures of the appropriate indicator strain (200 μl in 25 ml agar) The inoculated medium was rapidh dispensed in sterile Petπ-dishcs and. after solidification, dried for 30 mm under a laminar flow hood Wells of a uniform diameter were bored in the agar and sealed with 15 μl of tempered soft agar Sterile culture supernatant fluids (50 μl) ere then dispensed in the wells and the plates were incubated overnight at 30"C These cell-free solutions of the bacteriocin were obtained as above The difference between the area of the zone of inhibition (in mnι^:) and the area of the well was measured to estimate bacteriocin activih Protease Sensitivity Assays:- The following enzymes were dissolved in sterile distilled H-0 (SDW) to a final concentration of 50 mg/ml Trypsin (type II. Sigma Chemical Co . Poole. Dorset. England), alpha-chymotnpsin (type II. Sigma). Proteinase (Sigma). Pronase E

pc XIV Sigma). Catalase (Sigma) Pepsin was dissolved in 0 02N HC1. also to a final concentration of 50 mg/ml All enzyme solutions were filter sterilised using disposable filters (Rotrand/Red πm 0 2 um Schleicher & Schueil) 20 μl a quots of filtered cell-free bacteriocin solution and 20 μl of each enzΛ-me solution were spotted 1 cm apart on GM 17 agar plates and dried for 30 mm All plates were incubated overnight at 30°C The plates move subsequently overlaid with the indicator organism Growth of the producer was evident as a zone of inhibition Controls included plates with spots of either bacteπocin solution or enzyme solution Effect of pH and temperature on bacteriocin stability:- A cell-free bacteriocin was exposed to vanous pH/temperature treatments Samples were heated to 60. 70. 80, 90, 100, 1 10 and 121 °C at

θ

SUBSTTTUTE SHEET (RULE 26) pH5. pH7 and pH9 for 10 nun. Following treatment, all solutions were rapidly cooled and bacteriocin activity was assayed by the well diffusion method.

Estimation of Molecular Weight:- Samples containing 3147 bacteriocin were loaded on urea-SDS polyacrylamide gels, prepared as described by Swank and Munkres (1971). Molecular weight markers for peptides ranging from 2.5 to 17 kDa (Sigma) were used as a standard A sample of lactococcm was also included on the gel as another indicator of molecular weight. After electrophoresis at 22 mA for 16 h the gels were divided. One half containing the sample and molecular weight markers was stained according to procedures as recommended by the manufacturers (Hoeffer Scientific Instruments). This essentially involved staining overnight with 0 125% Coomassie Brilliant Blue R250 stain The following day. the gel was destained with "Destam Solution 1" [Methanol 50% (v/v). Acetic Acid 10% (v/v)] for 5 h and then placed in "Destain Solution 2" [Methanol 5% (v/v). Acetic Acid 7% (v/v)] Tlie other half of the gel was fixed immediately for 2 h in a solution of 20% (v/v) isopropanol and 10% (v/v) acetic acid It was dien washed in several volumes of dcionised water for 4 to 6 h The gel was placed in a large sterile glass Petn-dish and overlaid with 30 ml of soft tempered agar seeded with 800 μl indicator strain The plate was incubated overnight at 30"C and examined for a zone of inhibition (Bhunia et. al.. 1987) Isolation of DNA:- Plasmid DNA from lactococci was isolated by the Anderson & McKay method ( 1983) as follows. Actively growing lactococcal cells were collected as with the previous method by centnfugation for 5 sec. The pelleted cells were then resuspended in 379 μl of solution containing 6 7% sucrose. 50 mM Tris. 1 mM EDTA (pH8) and heated to 37"C Lysozymc (Sigma), dissolved in 25 mM Tris pH8 was added (96.5 μl). and incubated at 37T for 5 mm. 48 2 μl 0 25 M EDTA. 50 mM Tris pH8 was then mixed in by gentle vortexing. followed by addition of 27 6 μi of the lysis solution (20% SDS. 50 mM Tris. 20 M EDTA pH8) Tlie resulting mixture was then incubated for 5 to 10 mm at 37^UC to complete lysis after which it was completely clear Tlie lysatc was rigorously vortexed for 30 sec and 3 N NaOH (27.6 μl) added. This was mixed in gently

inversion for 10 min. after which. 49 6 μl of 2 M Tris pH7 was added This was mixed for a further 3 min by inversion. 71.7 μl of 5 M NaCl was then added and the resulting mixture vortexed Tins was extracted once with phenol and once with chloroform isoamyl alcohol (24: 1) The DNA was then precipitated on addition of 600 μl isopropanol. Following centrifugation for 10 mm. the precipitated DNA was washed once in 70% ethanol. dried and resuspended in 30 μl SDW. all of which was then loaded on the gel.

.Genomic DNA was extracted from lactococci according to a modification of the method outlined by Hoffman & Winston (1987) which uses shearing with glass beads to lyse the cells. The strains from which DNA was extracted were grown overnight in 5 ml volumes of tlie appropriate medium Tlie cells were collected by centrifuging 2 ml volumes of the cultures for 5 sec. Tlie supernatant was discarded and the remaining pellet was vortexed briefly The cells were resuspended in 0 2 ml of sterile "Extraction Solution" which consists of 2% Triton X 100. 1% SDS. 1 0 mM NaCl. 10 mM Tris (pH8) and 1 mM EDTA Phenol-chloroform (0.2 ml) was then added, followed by 0 3 g of acid-washed glass beads of diameter 0 45 to 0.52 mm (Sigma) The suspension was vigorously vortexed for 2 mm and then microcentrifuged for 5 m . The resulting upper aqueous phase was gently transferred into a sterile micro-centrifuge tube to which 20 μl of 3 M sodium acetate was added After a gentle vortex. 600 μl absolute ethanol (-20°C) was added and the mixed suspension centπfuged for 10 mm Then the pellet was washed in 70% alcohol, dried and finally resuspended in 50 μl SDW To estimate DNA concentration, a 5 μl sample was electrophoresed on a 0 7% agarose gel with ethidium bromide staining Alternatively, lactococcal genomic DNA was isolated by a modification of the Anderson & McKay procedure This was performed identically to the method outlined above except that the alkaline dcnatu ration step was omitted

Polymerase Chain Reaction:- Amplification of lactococcal DNA was perfoπned by the following method using a Perkin Elmer DNA Thermal Cycler PCR reactions of 100 μl were set up which contained one-tenth volume 10X buffer (Bioline). 5 mM MgCk 200 μM of each of the dNTPs. 1 μM of pπmer(s) and 1.25 units of Taq DNA

After overlaying each tube with 100 μl sterilised paraffin oil. 1 μl DNA (isolated as previous!*, described) was added to the reaction The Taq added during the first temperature cycle ("Hot Start") and the DNA was amplified for

ch cycle involved I mm dcnaturation at 93°C. followed

an annealing step at 55°C for 1 nun and an extension step of 72"C for 1 min Of the final reaction mixture. 10% was analyzed on 1 8% (w/v) agarose (Sigma) gels with ethidium bromide staining Restriction of DNA:- The isolated genomic DNA from the three L lactis strains. DPC3 147 (lacticin 3147 producer). MG 1614 (Bac ) and NCD0496 (nisin producer) were restricted with Hind III enzyme Restrictions were carried out 10X "Cuts-All" buffer containing 200 mM Tris-HCI (pH7.5). 70 mM MgCk 1 M KC I and 20 mM β-mercaptoethanol and the entire reaction mix was incubated overnight at 37"C The DNA samples were run on a 1 8%(w/v) agarose gel at 25V overnight Tlie DNA fingcφπnt thus obtained, was examined under ultra-violet light to ensure that sufficient DNA had been restricted before proceeding to the next step. Southern Blotting:- After electrophoresis. the restricted DNA was transferred to a Hybond N^* nylon membrane by capillary blotting Tlie procedure used was that described by Ma atis et al. (1989) Initially, the gel was soaked for 10 min in several volumes of 0.2 N HC1 and rinsed briefly in.deionised water. The DNA was then denatured by soaking the gel for 45 min in several volumes of 1.5 M NaCl. 0.5 N NaOH with constant gentle agitation. Again, the gel was rinsed in deionised water. To neutralise die gel. it was soaked twice in 1 M Tris (pH7.4). 1.5 M NaCl for 30 min with gentle agitation After neutralisation, the DNA was transferred to a Hybond N^* nylon membrane by

10

SUBSTTTUTE SHEET (RULE 26) capillary action. Tlie membrane, which had been previously immersed in 10X SSC transfer buffer (NaCl 87.65 g/1, sodium citrate pH7.0 44.1 g/1) for 5 mm was placed directly over the gel and transfer was allowed proceed for 24 h. After this time, the membrane was removed, dried and baked for 2 h at 80°C. Hybridization:- Hybridizations were performed according to the "Enhanced Chemiiuminescence" ECL Gene Detection System a Tecline Hybridiser HB-1 All hybridizations were carried out at 42°C as suggested by the manufacturer. The nylon membrane containing the transferred DNA was placed in hybridization buffer (supplied with tlie ECL kit) at 42°C and a pre-hybπdization was earned out for 15 min. Tlie labelled DNA probe was then added and incubation was allowed to proceed overnight at 42^UC Tlie membrane was then removed from the hybridization solution and washed twice at 42°C for 20 nun with primary wash buffer which consists of Urea 36% (w/v). SDS 0 4% (w/v) and 20X SSC 2.5% (v/v). It was then washed for 5 mm at room temperature in secondary wash buffer I e 20X SSC 10% (v/v) Amplified DNA (20μl) from L lactis NCD0496 was run on a 0 6% (w/v) low melting point Sea Plaque Agarose (FMC) gel Tlie 166 bp nisin probe DNA amplified from NCD0496 DNA was carefully cut from the gel Labelling this DNA was achieved as follows 20μl DNA labelling reagent was added to an equivalent volume of cooled to 37°C denatured DNA and mixed thorough h 20μl glutaraldehyde solution was then added and the mixture was incubated for 10 m at 37^UC Labelling reagent and glutaraldehyde solutions were both supplied with the ECL kit Detection of hybridization signals was perfoπncd according to the manufacturers recommendations

Conjugation:- Conjugations were perfoπned as follows A conjugal mating was set up using L lactis DPC3147 (bac^*. bac'. strep") as the donor strain and the plasmid free strain MG 1614 (bac . bac'. strep') as the recipient Both strains were grown to mid-log phase (OD₆,«,-_mϋ.5 to 1 ) Tlie '3PC3147 strain was cultivated in GM17 containing pronase E (50 mg ml). while MG1614 was grown in GM 17 supplemented with streptomycin (500 μg/ml). 1 l ahquots of these cultures were then centπfuged in a microfuge for 30 sec and the resultant pellets ere washed once in 1 ml volumes of GM 17. Tlie pellets obtained were resuspended in 25 μl GM 17. mixed and spotted on a non-selective GM17 agar plate. Donor and recipient controls were prepared in a similar manner Following overnight incubation at 30°C. the cultures were resuspended in GM I7 broth supplemented with 40% glycerol for long-teπn storage at -80°C. A serial dilution was carried out on an aliquot of the mating mix which was then plated on selective media. The conjugation frequency was estimated as the number of transconjugants (appearing on selection plates) divided by the number of donor cells. Putative transconjugants were checked for lactose metabolizing activity by streaking on L1A plates. To determine the molecular weight of the plasmid encoding bacteriocin production, plasmid DNA from the transconjugants was isolated by the Anderson and McKay method and was run on a 0.7% agarose gel. L. lactis MG1614 was used as a negative control and the plasmid profile of L. lactis subsp. lactis DRC3 was used as the molecular weight standard

Preparation of Inocula for Cheese-making Trials: d e selected strains were stored at -80°C. and were sub-cultured once in LM 17 broth. At 30°C overnight and twice in RSM before use. Bulk starters were cultivated in 10% RSM which had been pre-heated to 90°C for 30 mm and cooled to 21°C before inoculation All die strains were grown separately at 21°C for 16 h. mixed in equal proportions and inoculated at the levels shown in Table 2

Cheesemaking: milk was pasteurised (72T x 15s) and cooled to 30°C. Cheese was made in circular jacketed stainless steel 500 1 vats. The milk was inoculated widi cultures at the levels summarized in Table 1. Filter- sterilised rennet (Chr. Hansens Laboratories. 31ml. diluted in 500 ml sterile distilled water) was added 30 m after addition of the starter and the coagulum was cut approximately 40 min later The curds and whey were cooked to 38^ϋC and pitched at pH 6 2 The cheddared curd was milled at approximately pH5.2. salted at a level of 27g/kg and pressed in 18 kg moulds overnight at approximate!) 412 kPa Cheeses were vacuum packed and ripened for 12 niond_s at 8°C

Analysis of Cheese: Bacteriological analysis. At intervals, cheeses were ascpticaih sampled Starter cells were enumerated on LM 17 agar for 3 days after incubation at 3ϋ^uC. lactobacilli on LBS agar after 5 days at 30°C. enterococci on Kanamycin acscuhn agar (Oxoid) after 24 h at 37^υC and co form on VRB agar after 24 h at 3ϋ"C. Microbiological analyses were single estimations at each sampling time

Gross composition Grated cheese samples (2 weeks old) were analysed for salt. fat. protein and moisture Tlie pH was measured on a paste prepared by macerating 10 g of grated cheese in 10 nil of distilled water All values are the average of duplicate analyses Proteo sis Proteolysis of cheese was monitored by measuring the percentage of total nitrogen soluble in water at pH 4 6 (WSN) or in 5% phosphotungstic acid (PTA-N)

Free amino acids Free ammo acids were measured in the 12% TCA-soluble fraction of tlie WSN on a Beckman 6300 Amino Acid Analyser.

Sensory Evaluation: Cheeses were assessed for flavour at 3. 6. 9 and 12 months by a sensory evaluation panel based on a score of 0-8 (0-1 unacceptable. 2-3 poor. 4-5 acceptable. 6-7 good. 8 excellent).

Assaying bacteriocin directly in cheese samples.- Bacteriocin activity was assayed from cheese samples as follows. Cheese samples were initially macerated in equal volumes of distilled water in a stomacher (Lab-Blender 400) for 15 min and heated to 80 °C for 10 min. Then ahquots of 50 μl were dispensed in wells and bacteriocin activity calculated as die difference in area of the zone of inhibition (in mm2) and the area of die well (as outlined previously).

Transfer of pMRCOl into lactococcal starter strains:- Initially a conjugation was set up widi L. lactis DPC3147 (Lac+. Bac- and Streps) as the donor strain and the plasmid-free antibiotic sensitive L. lactis MG 1363 as the recipient. Bodi strains were grown from an overnight culture for 4hr at 30 °C in IJGM 17. The conjugation was carried out in a ratio of 20: 1 of recipient to donor. 1ml of recipient (MG1363) and 50 μl of donor (DPC3147) were centπfuged Tlie donor (Bac+) was washed with LM 17 broth and resuspended in 50 μl of LM 17. Tlie donor, recipient and mating mix were spotted onto non-selective GM 17 agar plates and allowed to dry. Following an overnight incubation at 30 °C. the cultures were harvested from the agar plates and resuspended in 500 μl of GM17 supplemented with 40% glycerol (for long-term storage at -80 °C) A serial dilution of an aliquot of the mating mix was plated onto selective media (LI A containing lacticin 3 147) The conjugation frequency was estimated by dividing the number of transconjugants appearing on tlie selection plates by the number of donor cells Transconjugants were not readily visible in this case because the donor Lac+ colonies tum the LIA plate yellow and the Lac- colonies are masked To overcome this problem colonies were picked off at random and spotted onto LIA and observed for lactose utilisation Tlie transconiugant resulting from tins mating was used as a Food Grade donor on further matings with the commercial lactococcal starters These matings were earned out similarly to above, but at a ratio if I 1 (with the exception of strain AM2) In these matings transconjugants were observed as Lac+ colonies in a Lac- background Putative transconjugants were picked from the mating plates and streaked for purity Tlie-.' were tested for lacticin 3 147 production - which is usually an indicator of the success of the mating and plasmid profiles were prepared to compare the recipient to the transconjugant.

Piiage resistance:- Tlie phage resistance of a culture was determined by comparing the transconjugant to the parent strain for resistance to a phage homologous to the parent Plaque assays were carried out as follows 0 25 ml of an overnight culture. 0 1 ml of 1 M CaC12 and 0 I ml of the appropriate phage dilution were added to 3 ml of L/GM 17 sloppy agar (0 7% agar) The contents were mixed, poured onto M 17 agar and incubated at 30 ^UC for 18 hr.

Detection of Diacetyi, Citrate and Acetolactate. Tlie assays were conducted on cells grown in 10 % RSM (+ 0.5 % tryptone) at 21 T and 30°C for 16 hr and 18 hr (respectively) according to the methods for detection of diacetyi. citrate and acetolactate as described in Prill & Hammer ( 1938). Marier & Boulet ( 1958) and Jordan & Cogan ( 1995) respectively Each assay was carried out in triplicate and the average presented in Table 5.

Concentration and partial purification of lacticin 3147 for incorporation into teat seals:- TY (Tryptone 2.5g/L. Yeast Extract 5g L, Glucose 10 g/L, b-glycerophosphate 19 g/L, MgS04.7H20 0.25g/L. MnS04 4H20 0.05 g/L, pH 6.75) broth was prefiltered (15 filter/litre) with HA (Millipore) filters to remove proteins in tlie media which would bind to the filters. L. lactis DPC 3147 was then grown overnight in the filtered TY-broth. The culture was centrifυged at 10.000 φm for 15 minutes and then filtered through HVLP filters (Millipore). Tlie bacteriocin was then bound to HA filters (8 filters/ litre) and subsequently harvested from the filters using acetone/ 5mM Phosphate-Buffer. pH 7.0 (2.1) The mix was subsequently centπfuged and the acetone removed by evaporation The resultant bacteπocin preparation was frecze-dried and dissolved in sterile distilled water. This was then assayed as above. To add bacteriocin into die teat seal 17.000 units of the bacteπocin prepared as above was added to die seal in a steπle petπ disc. On mixing, die bacteriocin and seal formed an emulsion which was then aseptically transferred to a syringe (Cross

Vetpharm)

Effectiveness of lacticin 3147 on Streptococcus dysgalactiae M - 10 μl from a overnight culture was added to 490 μl of steπle 50 mM phosphate- buffer. pH 7 0 500 μl lacticin 3147 ( 10.000 AU/ml) was then added and plate counts were earned out after 0. 15. 30. 60. 90 and 120 minutes at 37°C to assess cell viability

Oral Streptococci:- Four caπogenic streptococcal strains isolated from infected patients were grown overnight at 30 °C in Brain Heart Infusion broth These strains were obtained from Dr Ger Fitzgerald. University College Cork. The relative seπsitivit*. of each of these strains to the bacteriocin was then determined in companson to L. laciis HP Results

A number of lactococci which exhibited antimicrobial activities were isolated from kefir grains Protease sensitivity assays demonstrated that the antimicrobials were bactenocms. since t e could readily be degraded by proteinase K On die basis of cross-sensitivity assays, these bacteriocin producers could be classified into different groups Strains from the first group. DPC3147.

DPC3 I53. DPC32I5. DPC3400. DPC3204 and DPC3244 exhibited cross-immunitv indicating that they all produced the same or a very similar bacteriocin (see Figure I ) Likewise, strains from the second group. DPC3220 and DPC33( 1) also exhibited cross-immuπitv to each other, but were sensitive to the bacteπocιn(s) produced by the first group Strain DPC2949 defined a third set of producers which exhibited cross-sensitivity to members of the first group

Subsequently. L. lactis DPC3147. L. laciis DPC2949 and L. lactis DPC3220 were chosen as representatives of each group for further study Initially, these strains were tested for their ability to inhibit a wide range of organisms. Tlie results of these experiments are given in Table 1. It can be seen that the bacteπocin-producers. DPC3147. DPC2949 and DPC3220 do no inhibit themselves but do however inhibit one another. This indicates that die strains are at least distinct from one another. To reduce tiie possibility diat they may be previously well characterised bactenocin-producing strains, cross-sensitivity studies were carried out to some well known bacteπocin-producers which were. L. lactis CNRZ481. the producer of lacticin 481 (Piard et. al. 1991). L. lactis NCD0496 (Figure 1) and NCD0497, nisin producers, and L. lactis subsp. cremons 9B4. the producer of lactococcin A. B and M. Each of the bacteriocin-producing strains in question inhibit these four previously characteπzed strains very well, with DPC3147 being particularly effective against them. In addition, these four strains effectively inhibited DPC3 147. DPC2949 and DPC3220. Based on these observaUons. it thus appeared that none of the strains DPC3147. DPC2949 and DPC3220 produced either nisin. lacticin 481 nor lactococc A. B and M Spectrum of Inhibition: The range of organisms inhibited by each of the strains was examined. Given the extensive application which nisin has found in the food industry, the nisin producer NCD0496 was included in this study for comparative puφoses. The relative sensitivities of 54 strains to L. lactis DPC3147. NCD0496. DPC2949 and DPC3220 are presented in Table 1 All four producers were very effective in inhibiting other lactococci which included a number of cheese-making strains The range of inhibition exhibited by DPC3220 however, appears limited to lactococci. and as such is described as having a narrow spectrum In contrast, the bacteriocin produced by DPC3 147 has a very broad spectrum of inhibition which closely resembles that of the nisin producer. NCD0496 Without exception, all Gram positive indicator bacteria tested, clcding lactococci. lactobacilli. enterococci. bacilli, leuconostocs. pediococci. clostridia. staphylococci ;_ιd streptococci were sensitive to it Notably, the two clostπdial strains tested were particularh sensitive to DPC3147 Indeed. C spυrogenes was so sensitive that the entire overlay was inhibited and no actual zone of inhibition could be measured C lyrobuiyricum was also found to be extremely sensitive The Listeria strains tested, including the pathogenic strain L monυcyiυgenes NCTC5348 were also sensitive to bactenocιn(s) produced by the DPC3147 strain Overall, the biological activity of the 3147 bacteriocin closely resembled that of nisin Interestingly, it was found that DPC3147 was significantly more active than the nisin producer against S ther ophilus HA Tlie closely related lactococci were all very sensitive

In contrast. L. lactis DPC2949 has an intcπncdiate spectrum of inhibition h ing somewhere between that observed for strains DPC3147 and DPC3220 Like DPC3220. this strain was effective in inhibiting all of the lactococci tested, but in addition, inhibited most of the Laciυbσctllus and Leitconosioc species. Thus, its biological activity was quite unlike that of lacticin 481 producers However, as stated above, the DPC2949 strain and L. lactis CNRZ481 exhibited cross-sensitivity Interestingly, this strain also had slight activity against Eschertchia cult and J'seudυmonas aerogmosa but it did not show any inhibition against enterococci. Listeria or staphylococci Relationship with Nisin: Although the cross-sensitivity studies suggest that DPC3 147 is not a nisin producer, even though their biological activity appears remarkably similar, a number of experiments aimed at probing the relationship between the two bactenocms were then performed. Protease sensitivity assays demonstrated that the 3147 bacteriocin is sensitive to trypsin. alpha-chymotrypsin. proteinase K and pronase E but not to pepsin (Figure 2). In contrast, nisin is not sensitive to trypsin but is degraded by

pancreatin and subulopeptidase. As expected, catalase had no effect on the antimicrobial activity of the 3147 bacteriocin thus eliminating the possibility that the antimicrobial activity may be due to hydrogen peroxide. In addition, none of the strains producing lacticin 3147 were capable of feπnenting sucrose. This provides additional evidence that this bacteriocin is not nisin. since the genes encoding nisin (nis A) is linked to genes responsible for sucrose cataboiism on the nisin-sucrose transposon. Tn5276 (Horn et. al. 1991). Thus, all those Nip^* (nisin-producing) strains studied to date are in addition Sue^* (Sucrose-fermenting). As expected. L. lactis NCD0496 was found to be Sue^* while neither the DPC3220 nor DPC2949 strains could utilize sucrose as a fermentable substrate. In addition, the minimum concentrations (MIC) of pure nisin required for inliibition of strains DPC3147 and NCD0496 were detemiincd. It was found diat DPC3 I47 was inhibited at 100 μg/ml of nisin while NCD0496 was not inhibited until 400 μg/ml of nisin was added Tliese differing MIC values of nisin for DPC3147 and NCD0496. together with all of the above results all strongly suggest that the DPC3 I47 bacteπocin is distinct from nisin.

Initially, certain strains were investigated for their genetic potential to produce nisin by the polymerase chain reaction (PCR) Using the published sequence of the nisin stmctural gene. Dodd et al (1990). two primers were synthesized which are complementary to sequences occurring proximal to the 3' and 5' ends of the nts A gene Tliese should amplify a PCR product of 166 base pairs from nts A -containing template DNA Indeed, an amplified product of approximately' that size was consistently amplified from genomic DNA isolated from the NCD0496 (Figure 3) strain. In contrast, no amplified product was observed when genomic DNA from DPC3147. DPC2949 and DPC3220 was used. The DNA amplified from L lactis 496 representing most of the nis A gene was then used as a gene probe to DNA isolated from lactococcal strains NCD04 6. DPC3147. DPC2949 and DPC3220. Tliese DNAs were first digested with Hind III and electrophorcscd on a 1% (w/v) agarose gel prior to transfer to a nylon membrane As shown in Figure 4. the nis A gene probe hybridised to a 3 5 kb Hind III fragment on the 496 genome In contrast, no hybridizing DNA was observed in DNA isolated from strains MG 1614 nor DPC3147 This suggests that the bacteriocin produced by DPC3147 is not a close homolocue of nisin Given the inliibition spectrum of the bacteriocin produced by L. lactis DPC 3147 and the results of the experiments discussed above demonstrating that it was not nisin. it was concluded that DPC 3147 produced a novel, broad-spectrum bacteriocin which was designated lacticin 314 Growth ofZ.. laciis DPC3147 and production of lacticin 3147 was monitored in GM 17 and TYT30 over a 24 hour period. As shown in Figure 5 higher cell numbers and bacteriocin activity are obtained in GM17. the richer medium. In both media, lacticin 3147 is produced during exponential phase and peaks during early stationary phase. Subsequently, the bacteriocin activit^y declines gradually during stationary phase. Production was also determined in a variety of different media including MRS, BHl. TYP, RSM and whole milk. Activity (mm²) was measured from the filtered supernatant of an overnight culture. Production was found to be greatest in MRS with an activity of 170 mm². The activity in RSM and whole milk was 90% and 73% of that found in MRS.

16

SUBSTTTUTE SHEET (RULE 26) Tliis indicates diat lacticin 3147-producιng starter cultures could be used to produce lacticin 3147-contaιnιng dairy products.

Effect of pH and temperature on bacteriocin stability: The effect of pH and temperature on die stability of lacticin 3147 was investigated and results are shown in Figure 6. Three samples of bactenocin were brought to a pH of 5. 7 and 9 and aliquots of each were subsequently heated to 60°C. 70°C. 80°C. 90°C. 100°C. 1 10°C and 12 PC for 10 minutes. The activity at each pH prior to heat treatment was measured and was taken to be 100% for that pH value From the graphical representation in Figure 6. it can be seen that lacticin 3147 is heat-stable, particularly at an acid pH Indeed at pH 5. die bactenocin survives autoclaving at 121°C for 10 minutes and. at pH 9. maintains more than 50% of its activity up to a temperature of 100°C

This means that lacticin 3147 has a potential for use in bodi high-acid and low-acid canned foods Low-acid foods (pH4.5) should receive sufficient heat treatment to destroy heat resistant spores of pathogenic C. boiulinntm By adding lacticin 3147 to diese foods, it should be possible to reduce the extent of heat-processing required, thus resulting in improved flavour, increased nutritional value and an overall more economical process Such applications may be particularly beneficial for products such as canned milk puddings here heat penetration is often a problem Lacticin 3147 could also be potentially used quite successfully in high-acid foods (pH<4 5). given its acid pH optimum Even though, a substantial proportion of lacticin 3147 is lost on heating to temperatures exceeding 100°C. its efficiency as a food preservative should not be compromized by heat, as heat-treated bacteπal spores display greater sensitivity to bactenocms Hence, inliibition of such spores would require the same level of active lacticin 3147 For the reasons stated above lacticin 3147 also has a role in meat preservation

Molecular weight determination: The molecular weight of lacticin 3147 was estimated by SDS polyacn la ide gel eiectrophorcsis according to the method of Swank and Munkres ( 1 71 ) As a control, a sample of lactococcin A. which has a molecular weight of 3 kDa was also anahzed Tlie gel. which was subsequently overlaid with agar seeded with L. lactis indicator strain HP is shown in Figure 7 It can be seen that lacticin 3147 is somewhat smaller than lactococcin A and its molecular weight was estimated at 2.8 kDa Molecular weight markers ranging from 2.5 to 17 kDa were used as a standard in the other half of die gel. Genetic studies: During initial genetic work with lacticin 3147. it was observed that the bactenocin-producing property was an easily lost trait. Preliminary experiments were attempted to establish if bacteriocin production by DPC3147 was encoded on a conjugally transmissable piece of DNA. These conjugal matings involved using DPC3147 as the donor strain and die plasmid-free strain. L. laciis subsp. lactis MG1614. which is streptomycin resistant as the recipient. The selection of transconjugants from such matings was achieved using die plasmid-encoded bacteriocin immunity/resistance determinants as a selectable marker. This involved the incoφoration of lacticin 3147 into the selective media. Even when plated at a high cell density (up to 10^s-10⁹ CFU/ml). L lactis subsp. lactis MG1614 sensitive cells failed to grow on such media, indicating a very low level of spontaneous resistance occurring for this strain to the 3147 bactenocin. The incoφoration of lacticin 3147 into selective media may thus form the basis for a novel food-grade marker system for use in genetic manipulation of lactococci. In matings involving strains MG 1614 and DPC3147. putative transconjugants (bac^,mm. strep^r). were isolated at a frequency' of 10^"3 per donor cell These cells yvere subsequently' found to be lactose deficient and also had acquired the ability- to produce bacteπocin As expected, these putative transconjugants exhibited cross-immunity' to DPC3147 and also, could inhibit a yvide range of Gram positive bacteria. Evidence that tliese actually represented true transconjugants y as obtained on analysing their plasmid complements In all cases, the bac^". bac^ιrnm. strep' cells had acquired a 63 kDa plasmid. designated pMRCOl A similar sized plasmid is clearly evident m plasmid profiles of the DPC3147 strain (Figure 8) This indicates that the genetic determinants encoding lacticin 3147 arc encoded on die pMRCO l coniugative plasmid Furthermore, similar numbers of colonies yycre obtained from the mating yyhen plated on media containing streptomycin and bactenocin. and when plated on media containing streptomycin soleh Moreover, yvhen diese strep' colonies were overlaid with strain MG1614. all y ere observed to be bac^*. This would suggest that all those MG 1614 cells which had not received the plasmid during mating had been killed off by lacticin 3147 produced by DPC3147 in die mating mix This yyould negate the requirement for incoφoratmg bacteπocin to select transconjugants in such matings Mastitis:- Since lacticin 3147 yvas shoyvn to be effective in inhibiting i^' a reus ATCC25923 and S ihermophilus HA and ST1 12 (Table I ), a separate study yvas initiated to investigate its ability- to inhibit a variety of clinical isolates obtained from mastitic animals Tliese vinilcnt bacteria yycrc obtained from the Depamient of Dairy Husbandry-. Teagasc. Moorepark. Femioy. Co Cork In all six streptococci and ten staphylococci yvere tested. DPC3147 yvas most effective against these pathogenic strains, as shoyvn in Table 3 All the streptococci, except S. dvsgalaciiae strain M yycre quite sensitive to it as yvere all the staphylococci. except strain 12. A similar outcome yvas seen y^y it NCD0496. although in some cases DPC3147 is slightly more successful against the streptococci DPC3220 does not display- any- activity' against either the streptococci or the staphylococci In general the streptococcal strains yvere more sensitive to DPC3147 than the staphy lococcal strains yyhereas the reverse is true for the nisin producers NCD0496 and NCD0497 This suggests that it yvould be particularly effective to use lacticin 3147 in combination with another bactenocin. such as nisin. as a mastitis treatment. In such a situation it is much less likely that the infecting organisms would develop resistance to both bacteπocins. Clearly, the use of bactenocms in the treatment of mastitis may' mean that milk would not have to be withheld as would be the case ^yvith an antibiotic-based treatment.

Phage Resistance: - Tlie conjugal plasmid. pMRCOl. also conferred an increased level of phage resistance to L lactis subsp lactis MG 1614 In contrast to the MG 1614 parent, transconiugants containing die plasmid exhibited total resistance to the small isometric-headed phage 712 In addition, the burst size of prolate headed phage c2 appeared drastically reduced (as evidenced by pmpoint plaques) Further studies demonstrated that die resistance mechanism encoded by the pMRCO 1 plasmid did not effect the abihtv of phage to adsorb to the cells, nor did it appear to inhibit phage DNA replication once the infecting phage DNA yvas internalised inside the host Consequently, the life cycle of the phage yvas inhibited at some point after phage DNA replication had occurred This potent phage resistance mechanism yvas thus assumed to be an abortive infection (or Abi) mechanism In lactococci catabolism of lactose is usually plasmid-linked Therefore, the plasmid-free strain L laciis subsp lactis MG1614 cannot feπnent lactose Since MG1614 transconiugants containing the multifunctional pMRCOl are also lac deficient, this y ould suggest that the genes nccessarv for lactose catabolism are not encoded b pMRCOl plasmid MG1614 containing pMRCOl was then mated yvith a lactococcal cheese starter strain HP Putative transconiugants yycre selected from such matings based on their abihtv to ferment lactose and become resistant to lacticin 3147 These lac^' cells also noyv produced lacticin 3147 and had also become totally resistant to phage which noπnalh infect the HP strain Examination of the plasmid complements of these strains revealed that they had acquired an extra plasmid of 63 kb. the size of pMRCO 1 This demonstrates that bactenocin linked phage resistance may prove to be a very efficient method in the improvement of commercial cheese starters

These results indicate that lacticin 3147 producers ma\ actually be used as cheese starter strains raϋicr than adding a bactenocin preparation as w ith nisin Tlie pnmarv advantage of this is that by directly adding the producing strain, food additive legislation may be overcome as there is no rcstnction on the use of the strain itself C heese-Making Trials:- Problems associated yyith nisin producing starters include such undesirable cliaracteristics as an inability to produce sufficient acid for cheescmakmg. that they are usually- proteinase deficient and also that they are phage sensitive As outlined above the latter characteristics are not associated with lacticin 3147 producers since bactenocin functions and phage resistance are linked on pMRCO 1 To test the ability of such strains to act as cheese starters two separate cheese trials yvere performed In the first, three strains yvere used, one of yvhich ^yvas strain DPC3147 The other two. DPC3204 and 3256 are also kefir isolates, both of yvhich exhibit crossirnmunity with DPC3147 Consequently, these are lacticin 3147 producers as well. As illustrated in Figure 9. these strains produced acid much like die fast acid commercial strain 303 during cheesemaking Thus it can be concluded that with regard to acid production these strains make acceptable starters During cheddar cheese ripening non-starter lactic acid bacteria (NSLAB) can reach levels exceeding 10⁷cfu/g Since these can be mainly lactobacilli. it yvas important to investigate their groyvth or odierwise in cheese made yvith lacticin 3147 producers (remembering that lacticin 3147 inhibits all lactobacilli tested). As shoyvn in Figure 10. no NSLABs yvere detected in cheese made with the bactenocin-producing strains even after six months (the average duration of cheddar npening). In contrast. NSLABs had reached levels of 10⁷⁵cfu/g after approximately' 4 months in controls produced without the bactenocin. In the second cheese trial, a transconjugant of strain 303 yvhich produces lacticin 3147 yvas used as a starter yvith 303 again as the control strain. Tlie results shoyvn in Figure 1 1 again demonstrate that die 303 transconjugant perfoπned satisfactory as a starter duπng cheese manufacture In addition the NSLAB levels appearing in the cheese made in this trial (Figure 12) yvere significantly loyyer (in excess of 100-fold) than in the control cheese. Sensory analyses subsequently demonstrated diat there yvere no major differences in flavour and aroma between die tyvo cheeses

NSLAB yvhich are found in ripening cheeses may contπbute to the flavour of the cheese It yvould be possible to use lacticin 3147 or a commercial starter strain producing it to control the entire microbial population of a cheese, dius allowing the flavour of a cheese to be designed Lacticin 3147 yvas found to be particularly' active against Clostridia tyrobutyricum and C sporogenes It is yyell established that the outgroyvth of milk-contaminating anaerobic spore-formers such as C tyrobutyricum and C. bttiyrtcum is primarily' responsible for the problem of late-gas bioyving in some cheeses i.e. the formation of hydrogen gases and carbon dioxide during ripening resulting in the development of large holes. Tliese bacteria convert lactic acid into buty ric acid giving rise to off-flavours and aromas Thus, lacticin 3147 also has a use in such products given its potency against clostridia! strains.

Tlie introduction of this genetic material into lactococcal industrial strains introduces complications regarding the availability of food-grade selection markers, and the possibility of the loss of industrially important plasmid-encoded functions such as lactose fermentation, proteoh tic activity. bacteπophage resistance and citrate utilization. In this regard lacticin 3147 production and immunity may be incoφorated as a desirable phenotype of some industrial starter cultures used in many commercial applications The results described here demonstrate that incoφoration of this bactenocin as a selectable marker into the media itself can form the basis for a novel selectable system Tlie gene(s) encoding immunity to lacticin 3147 can be linked to desirable traits on plasmids and transconjugants containing diese plasmids yvould then selectively- groyv on bactenocin-producing media. It has been observed that die level of spontaneous resistance to lacticin 3147 by some commercial cheese making strains is very low and additionally, plasmid maintenance by pMRCOl -containing strains would be assured in fermentations smce cells yvhich lose the plasmid yvould be killed by the lacticin 3147 produced, bodi properties yvhich are required for an effective selectable system. This is a significant advantage as most selectable markers to date are not of food-grade quality-. Indeed, most are actually antibiotics yvhich cannot be used because of the danger of the occurrence of antibiotic resistant strains of clinical importance.

Strain DPC2949: The biological activity exhibited by __. lactis DPC2949 yvas quite similar to that of the previously characterized L. lactis CNRZ481 in that common sensitive strains include lactococci. Clostndium tyrobutyricum. Leuconostoc and some, but not all lactobacilli In addition. Bacillus substilis, Enterococcus faecalis. Listeria innocua and L. monocytogenes yvere not inhibited by either bactenocin producer. However, cross-sensitivity studies indicated that the*.' yvere actually different since both inhibited each odier. Unexpectedly. DPC2949 gave slight inliibition of the tyvo Gram negative strains Escherichia coll and Pseudomonas aeroginosa Strain DPC2949 thus produces an intermediate spectrum bacteriocin yvhich has applications such as the prevention of late gas-bloyving in Cheddar cheese. It could also be applied to the control of non-starter lactic acid bacteπa (NSLABs) and consequently, to assess dieir effect on Cheddar cheese quality, hich still remains to be established

Cheese-Making Trials - As stated previously . cheese made yvith Iactιcm3147-producιng starters had significantly' less non-starter lactic acidbactena (NSLABs) in them yvhen compared to corresponding cheeses made yvith a commercial starter. Since then yve have assayed the bacteriocin in these cheeses. In Cheese trial I the presence of lacticin 3147 in take test cheese was confirmed (usmg L. lactis AM2 as die indicator strain) at a level of approximately 1.280 AU/ml which remained constant in the cheese over the 28-yveek ripening period (Fig 13 A) In contrast, no anti¬ microbial activity- yvas detected in the control cheese Similarly-, the amount of bacteriocin detected in cheese in trial 2 yvas approximately 1.280 AU/ml yvhich also remained constant over the ripening period (Fig 13B) Thus the level of bacteriocin observed in the cheese corresponds to the number of NSLABs yvhich occur in the cheese during ripening

Genetic studies - To evaluate the potential usefulness of the lacticin 3147-encodιnu plasmid (pMRCOl) for starter strain improvement it yvas transferred into a variety of lactococcal strains including those currently- used for Cheddar cheese and lactic butter manufacture These transfers yvere performed in a Food Grade manner via conjugations (bacterial matings) after Λ Inch the ncy^yh modified strains yvere selected based on their immunity- to the bactenocin Such matings first necessitated the construction of a Food Grade donor strain for the plasmid yvhich is sensitive lo antibiotics. Depending on the starter recipient used one of three possible results for each of the matings yvas recorded. In the first, starters yvere isolated yvhich had improved phage resistance properties and produced (and yvere resistant to) bacteriocin (Table 4). Genetic analyses confirmed diat they had received die pMRCOl plasmid which could efficiently be mated back out of die strain These strains retained their commercially important characteristics for example ability to produce acid (for cheese starters) and/or diacetyi (for lactic butter production. Table 2). In addition, the plasmid appeared stable in these strains and was maintained over a number of successive subcultures. One such strain yvas subsequently used for pilot-scale Cheddar cheese manufacture Another result of these conjugations yvas die identification of strains hich produced the bacteπocin but had not improved phage resistance. Genetic analysis of one such strain demonstrated that such a phenotype yvas associated yvidi plasmid co-integration in this strain. Lastly, a number of strains yvere found to be recalcitrant to the pMRCO 1 plasmid. One possible explanation for this observation is that die plasmid maybe incompatible widi a resident plasmid of the strain (e.g. pMRCO 1 yvas found to be incompatible yvidi die lactococcal plasmid pNP40 in this study) The overall significance of die results of these matings can be summarized as folloyvs 1 ) Tlie genetic determιnant(s) encoding immunity to lacticin 3147 is very- useful as a food grade selectable marker for starter strain improvement The lactococcal strains tested in this study proved very sensitive to the bacteriocin incoφorated solid media unless they had received the bactenocin genes. Indeed, die bacteπocin is as convenient to use as conventional antibiotics for such studies 2) A number of new starter strains have noyv been generated using the bactenocin Many of these have also been improved with regard to their phage resistance Importantly- all diese neyv strains now produce the bactenocin and may be used as starters for products yy here the bactenocin might impart a desirable effect. Examples include reduction of NSLAB numbers in cheese products or elimination of undesirable bacteria from some fermented meat and fish products Mastitis:- As a result of the growing concern over the use of antibiotics for the treatment and prevention of disease in animals the potential of using lacticin 3147 in the prevention of bovine mastitis y\as investigated. To achieve this the bactenocin yvas incoφoratcd into teat seals preparations. Tliese seals are manufactured by Cross Vctphaπn (Broomhill Road. Tallaght. Dublin 24. Ireland) and act as a physical barrier in the cow against infection The bactenocin provides an additional anti-microbial barrier over the physical one provided by the seal itself Incoφoration of the bactenocin into die teat seals first required the preparation of the bactenocin in a highly- concentrated and semi-purified foπn. Tlie anti-microbial effectiveness of lacticin 3147 in the environment of the seal required the addition of either Ty een 20 or 80 in concentrations of 2% (Figure 14A). Where the detergent yvas not present negligible bactenocin activity yvas recorded in the treated seals Having successfully prepared effective bacteπocin-containing teat seals a number of animal trials yvere dien performed Tliese involved exposure of the modified seals to the animal for different time periods after yvhich die seals yvere removed Overall such studies have demonstrated that die animals tolerated die bacteriocin-containing seals as evidenced by low- associated somatic cell counts. In addition, seals extracted from the animal yvere later shoyvn to retain bacteriocin activity. Since lacticin 3147 was found to inhibit a range of mastitic streptococci its effectiveness on Streptococcus dysgalactiae M. a strain previously isolated from an infected animal, yvas studied. Addition of die bacteriocin (10.240 AU/ml) to stationary- phase cells of this strain resulted in a 99.99 % kill in just 2 hours (Fig. 14B). Other aspects of this study focused on the frequency at which a mastitic strain can develop resistance to lacticin 3147 since this could limit its practical usefulness as an effective antimicrobial under certain conditions However, only 0 003% bacterial cells of the M strain developed increased tolerance to the bactenocin after prolonged exposure

Inhibition of Oral Streptococci - A number of oral streptococci have also been tested for sensitivity to lacticin 3147 to investigate the use of the bactenocin in such applications as mouth yvashes. dental products etc The strains tested included four caπogenic streptococci isolated from infected patients All four strains tested proved sensitive to die bactenocin ( 12 5% as sensitive as L lactis HP) Importantly, fermented dairy products manufactured with lacticin 3147-producιng starters may therefore have additional anti-caπogenic characteristics

23

SUBSTTTUTE SHEET (RULE 26) Table 1. Inhibition Spectrum of Laclococcus lactis DPC3147. L. lactis NCD0496 and !, lactis DPC3220. [No Zone (NZ); 0 to 1 mm (-); 1 to 5 mm (+): 5 to 15 mm (++). 15 mm over (+++)]

^' Temperature of Incubation: '21°C; 37°C; ³30°C; ⁴42°C. * Anaerobic conditions DPC3147 produces lacticin 3147; NCD0496 produces nisin; DPC3220 produces a bacteriocin with a narrow spectrum of inliibition. Table 2. Strains and levels of inocula used in cheεsemaking noculation rate (v/v%) 1

0.7 0.7 0.7 1 1

Table 3. Sensitivity of mastitic strains of Streptococci and Staphylococci to Laclococcus lactis DPC3147. L. lactis NCD0496 and L. laciis DPC3220 [No Zone (NZ). 0 to 1 mm (-), 1 to 5 mm (+); 5 to 15 mm (++). 15 mm over (+++)] Strain

Streptococcus agalacttae (B) Streptococcus agalacttae (H) Streptococcus agalacttae (?) Streptococcus dysagalactiae (M) Streptococcus faecalis (I) Streptococcus bens (L) Staphylococci subspecies 2

11

12

13

89

10 1

22

32(a)

36

DPC3147 produces lacticin 3147; NCD0496 produces nisin; DPC3220 produces a bacteriocin with a narrow spectrum of inliibition. Temperature of Incubation = 37°C.

26

SUBSTITUTE SHE.ET (RULE 26) Table -■_ Conjugation results for mating pMRCOl into a variety of Laclococcus lactis starter recipients.

and all were lacticin 3147 producing.

* Mating of transconjugant with L. lactis MG1614.

Table 5 Diacetyi and acetolactate production by lactis DPC220 and _ lactis DPC220 containing pMRCOl (220 TcA).

Citrate ulilisation = 100 % for all the test cultures. 10% RSM + 0.5% tryptone at 30°C. References

Anderson. D.G.. and L.L. McKay. 1983. Simple and rapid method for isolating large plasmid

DNA from lactic streptococci. Appl. Environ. Microbiol. 46:549-552.

Bhunia. A.K.. M C. Johnson, and R. Ray. 1987. Direct detection of an antimicrobial peptide of Pediococcus aadilactici on sodium dodecyl sulfate - polvacrvlamide gel electrophoresis J Indus.

Microbiol 2 319-322.

De Vuyst. L.. and E.J. Vandamme. (ed) 1994 Bactenocms of Lactic Acid Bacteria.

Microbiology. Genetics and Applications. Blackie Academic and Professional.

Dodd. H.M . N. Horn, and M J. Gasson 1990 Analysis of die genetic determinants for production of the peptide antibiotic, nisin J Gen. Microbiol 136.555-566

Gratia. A 1925. C.R. Seanc Soc Biol 93 1040 in Mayr-Harting e t al 1972.

Gross. E . and J L Morell. 1971 Tlie Structure of nisin J Am Cliem Soc 93 4634-4635

Hoffman. C S.. and F Winston 1987 A ten minute DNA preparation from yeast efficiently releases autonomous plasmids for transformations of Eschenchia cυh Gene 57 267-272 Horn. N . S. Syyindeil. H. Dodd. and M Gasson. 1991. Nisin biosvndiesis genes arc encoded by a novel conjugative transposon Mol Gen. Genet. 228.129-135

Maniatis. T.. E.F Fπtsch. and J. Sambrook. 1989 Molecular Cloning A laboratory manual

Cold Spring Harbor Laboratory-. Cold Spring Harbor. N Y

Mattick. A T.R.. and A Hirsch. 1947 Further observations on an inhibitory substance (nisin) from lactic streptococci. Lancet 2:5-7

Morris. S L . R C Walsh. J N Hansen 1984 Identification and characterisation of some bacterial membrane sulfhydryl groups which arc targets of bacteπostatic and antibiotic action J

Biol Chem 259 13590-13594

Ogden K . and R.S Tubb. 1985. Inhibition of beer-spoilage Lactic Acid Bacteria by nisin J lust Brew 91 390-392

Parente. E.. and C. Hill. 1992. A comparison of factors affecting the production of tyvo bactenocms from lactic acid bacteria. J. Appl. Bacteriol 73 290-298.

Piard. J.C . P.M Muπana. M.J. Desmazeaud. and T.R. Klaenhammer. 1991. Purification and partial characterization of lacticin 481. a lanthiomne-containing bacteriocin produced by Lactococcus lactis subsp. lactis CNRZ481. Appln. Environ. Microbiol. 58:279-284

Rea. M C. and T M. Cogan. 1994. Buttermilk plants: the Irish version of Kefir. Tlie Irish

Scientist. 2.7.

Ruhr. E.. and H.G. Sahl. 1985. Mode of action of the peptide antibiotic nisin and influence on the membrane potential of whole cells and on cytoplasmic and artificial membrane vesicles. Antimicrob. Agents Chemother. 27:841-845. Steen. M T . Y J Chung and J N Hansen 1991 CharactenzaUon of the nisin gene as part of a polycistronic operon in die chromosome of Lactococcus lacϋs ATCC 11454 Appln Environ Microbiol 57 1181-1188

Stevens. K A . B W Sheldon. N A Klapes, and T R Klaenhammer 1992 Effect of treatment conditions on nisin inactivation of Gram negative bacteria J Food Prot 55 763-766 Syvank. R T . and K D Munkres 1971 Molecular weight analysis of ohgopeptides by electrophoresis in polyacrylamide gels yvith sodium dodecyl sulphate Anal Biochem 39 462-477 Maπer & Boulet. J Dairy Sci . 1958. 41. 1683 Pπll & Hammer. Iowa State Coll J Sci . 1938. 12, 385 Jordan & Cogan. Irish J Agric Food Res. 1995. 34. 39

Claims

1. A bacteriocin designated lacticin 3147 characterised by a molecular yveight of approximately 2 8 kDa. inhibiting activity against lactococci. lactobacilli. enterococci. bacilli, leuconostocs. pediococci. clostridia. straphylococci and streptococci. sensitivity to the proteases trypsin. alpha-chymotrypsin. proteinase K and pronase E but not pepsin, heat-stability. activity at acid pH. and the capability of inhibiting nisin-producmg bacterial strains

2 L lactis DPC3147 strain as deposited at the National Collection of Industrial and Marine Bacteria. Aberdeen. Scotland, on 1 1th April 1995 under the Accession No NCIMB 40716 and strains substantially similar thereto also encoding lacticin 3147 production 3 A 63 kDa plasmid encoding a bactenocin as defined in Claim 1

4 The plasmid pMRCOl which encodes the bactenocin designated lacticin 3147. lacticin 3147 immunity gene(s) and phage resistance genes as deposited at die National Collection of Industnal and Manne Bacteria. Aberdeen. Scotland, on 1 1th Apπl 1995 under the Accession No NCIMB 40716 and plasmids substantially similar thereto also encoding bactenocin 3147 production 5 Isolated gene(s) for the production of lacticin 3147 as deposited in the plasmid pMRCO l as deposited in the National Collection of Industrial and Marine Bacteria. Aberdeen. Scotland on 1 1th April 1995 under the Accession Number NCIMB 40716 and genes substantiall similar thereto also encoding lacticin 3147 production

6 Isolated gene(s) encoding immunity to lacticin 3147 as deposited in the plasmid pMRCOl as deposited in the National Collection of Industrial and Marine Bacteria. Aberdeen. Scotland on 1 Ith

April 1995 under die Accession Number NCIMB 40716. and genes substantially- similar thereto also encoding lacticin 3147 immunity

7 Phage resistance gene(s). encoded by the plasmid pMRCOl. as deposited at the National Collection of Industnal and Marine Bacteria. Aberdeen. Scotland, on 1 Ith April 1995 under the Accession No. NCIMB 40716. and genes substantially similar thereto also encoding phage resistance gene(s), particularly gene(s) encoding total resistance to the small isometric-headed phage 712 and burst size limitation for the prolate-headed phage c2.

8 A host cell comprising a plasmid as claimed in claim 3. or 4 or genes as claimed in claims 5.6 or 7. 9 Tlie use of lacticin 3147 as claimed in claim 1 or a lacticin 3147 producing host as claimed in claim 2 or 8 selected from uses in die treatment of mastitis in cattle, to prevent clostridial spoilage in cheese, as a food preservative in pasteurised cheeses and cheese spreads, as a shelf-life extender in milk and milk products, in the production of alcoholic beverages, particularly in the breyving industry, in vegetable fermentations, by incoφoration into canned foods and in meat preservation, in incoφoration into oral healthcare products such as toothpaste and mouthyvashes. in cosmetic treatments for acne, and against Gram negative bacteπa yvhich have been treated yvith chelating agents.

10. Use of a plasmid as claimed in claim 3 or 4 or a gene(s) as claimed in claim 5 encoding die lacticin 3147 bactenocin to confer bacteπocin producing properties on a host such as a bacterium, particularly a cheese-starter culture. 1 1. A method of producing lacticin 3147 comprising cuituring a host cell as claimed in claim 2 or 8 containing lacticin 3147-encodiπg gene(s) and isolating lacticin 3147 from die culture. 12. A method of conferring lacticin 3147-producιng properties on a host such as a bacterium, comprising introducing and expressing in the host a plasmid as claimed in claim 3 or 4 or gene(s) as claimed in claim 5 encoding lacticin 3147. 13. A food-grade genetic marker system comprising genes for immunity to lacticin 3147 as claimed in claim 6 as encoded by plasmid pMRCOl. characterised in that the genetic determinants which encode lacticin 3147 immunity arc introduced into a bacterial strain together yvith any' desirable gene(s) yvhich have been linked to d em.

14. A mediod of conferring phage resistance on a host cell and particularl a cheese starter culture, comprising introducing and expressing therein a plasmid as claimed in claim 4 or genc(s) as claimed in claim 7 encoding phage resistance

15. L. lactis strain DPC2949 as deposited at the National Collection of Industrial and Marine Bacteria. Aberdeen. Scotland on 1 1 tin April 1995 under the Accession No NCIMB 40715. and strains substantially- similar thereto also encoding a bactenocin 16. The isolated bacteriocin-encoding gcne(s) ofL. lactis strain DPC2949 as claimed in claim 15. 17. The bactenocin produced by L. lactis strain DPC2949 as claimed in claim 15

18 A host cell comprising genes as claimed in claim 16

19 A mediod of preventing late gas-bloyving in Cheddar cheese production comprising either the addition of L. lactis DPC2949 as claimed in claim 18 or die bacteriocin as claimed in claim 16 or the host cell as claimed in claim 18 to the initial cheese starter culture

20. A method of controlling non-starter lactic and bacteria in Cheddar cheese production comprising either the addition of L. lactis DPC2949 as claimed in claim 18 or the bacteriocin as claimed in claim 16 or the host cell as claimed in claim 18 to the initial cheese starter culture.