OUP user menu

Diversity of enterococcal bacteriocins and their grouping in a new classification scheme

Charles M. A. P. Franz, Marco J. Van Belkum, Wilhelm H. Holzapfel, Hikmate Abriouel, Antonio Gálvez
DOI: http://dx.doi.org/10.1111/j.1574-6976.2007.00064.x 293-310 First published online: 1 April 2007


Enterococci are lactic acid bacteria of importance in food, public health and medical microbiology. Many strains produce bacteriocins, some of which have been well characterized. This review describes the structural and genetic characteristics of enterocins, the bacteriocins produced by enterococci. Some of these can be grouped with typical bacteriocins produced by lactic acid bacteria according to traditional classification, whereas others are atypical and structurally distinct from the general classes of bacteriocins. These atypical enterocins recently played an important role in and prompted reclassification of the class II bacteriocins into a new scheme. In this review, a more simplified classification scheme for enterocins based on amino acid sequence homologies is proposed. Enterocins are of interest for their diversity and potential for use as food biopreservatives. The emergence of multiple antibiotic-resistant enterococci among agents of nosocomial disease and the presence of virulence factors among food isolates requires a careful safety evaluation of isolates intended for potential biotechnical use. Nevertheless, enterococcal bacteriocins produced by heterologous hosts or added as cell-free preparations may still be attractive for application in food preservation.

  • enterocin
  • bacteriocin
  • genetics
  • enterococci bacteriocin classification


Enterococci are Gram-positive, catalase-negative, coccus-shaped bacteria that have a DNA G+C content of <40 mol% and produce lactic acid as the major end product of glucose fermentation. Therefore, they meet the general characteristics of lactic acid bacteria (LAB). The enterococci were originally known as the group D streptococci but in the 1980s, based on DNA–DNA and DNA–RNA hybridization studies, Streptococcus faecalis and Streptococcus faecium were moved to the new genus Enterococcus (Schleifer & Kilpper-Bälz, 1984, 1987). Since then, major developments in the classification, number of species and clinical association of the enterococci have occurred (Devriese et al., 1993; Devriese & Pot, 1995; Stiles & Holzapfel, 1997; Franz et al., 2003; Franz & Holzapfel, 2004). Enterococci are commensals of the animal and human intestinal tract (Noble et al., 1978; Devriese & Pot, 1995). Since the 1980s, they have been recognized as emerging pathogens associated with nosocomial infections and superinfections such as endocarditis, bacteraemia, and urinary tract and other infections (Murray et al., 1990; Morrison et al., 1997; Franz & Holzapfel, 2004). Enterococci can be used as indicator organisms for faecal contamination of foods, they can cause spoilage of heat-treated meats and they are involved in ripening of many artisanal cheeses. Despite the concerns for enterococci as opportunistic pathogens, they have long been used as human and/or animal probiotics (Franz et al., 1999a).

The enterococci produce antibacterial peptides (bacteriocins), generally called enterocins. Bacteriocins are ribosomally synthesized, antimicrobial peptides with activity that is directed usually against closely related species (Klaenhammer et al., 1993; Nes et al., 1996). Interest in bacteriocins produced by some LAB has been stimulated by the fact that they are active against Gram-positive food-borne pathogens such as Listeria (L.) monocytogenes, Staphylococcus (S.) aureus, Bacillus cereus, and vegetative cells and spores of Clostridium (C.) botulinum (Cleveland et al., 2001; Chen & Hoover, 2003; Deegan et al., 2006). Because of this activity against food-borne pathogens and consumer demands for more ‘natural’ preservatives, bacteriocins have been suggested for use as ‘biopreservatives’ in foods, and their use in food or in model food studies has received much attention in recent years (Ohlsson, 1994; Holzapfel et al., 1995; McMullen & Stiles, 1996; Muriana, 1996; Schillinger et al., 1996; Stiles et al., 1996; Cleveland et al., 2001; Chen & Hoover, 2003; Deegan et al., 2006). Despite this, the class I, lantibiotic-type bacteriocin nisin is the only bacteriocin that is currently licensed for use as a food additive. Other bacteriocins are adventitiously present in foods through production by starter cultures or as a result of their presence in bacterial ferments that are added as food ingredients.

Enterocins produced by Enterococcusfaecium, Enterococcus faecalis and Enterococcus mundtii include the well-characterized enterocins A, P, CRL35, 1071A and B, mundticin, mundticin KS, bacteriocin 31, RC714, T8 and enterolysin A. These enterocins can be readily grouped in the current bacteriocin classification systems (see below) of Klaenhammer (1993) and Nes et al., (1996), but enterocins AS-48, B, EJ97, RJ11, MR10A&B, Q and L50A&B and bacteriocin 32 have unusual structural or genetic characteristics, different from known bacteriocins produced by other LAB, that prohibit grouping into these classification schemes. This review illustrates shortcomings in current bacteriocin classification of the enterocins and suggests a new, simplified classification scheme based on enterocin structure and amino acid sequence similarities. The diversity of bacteriocins produced by Enterococcus spp., the ecological significance of bacteriocin production by enterococci and the practical importance of enterocins for use in foods and as possible therapeutic agents are discussed.

Current classification of bacteriocins

Klaenhammer (1993) defined four classes of bacteriocins produced by LAB. Class I bacteriocins or ‘lantibiotics’ are small, ribosomally synthesized peptides that undergo extensive post-translational modification. They contain lanthionine and β-methyl lanthionine residues, as well as dehydrated amino acids. Class II bacteriocins are small (4–6 kDa), heat-stable, ribosomally synthesized peptides that do not undergo extensive post-translational modification, except for cleavage of a leader peptide during transport out of the cell. They are produced as a prepeptide that generally contains a leader sequence consisting of 14–30 amino acids, with a conserved processing site of two glycine residues at positions -2 and -1. Both class I and class II bacteriocins are secreted by ATP binding cassette (ABC) transporter proteins. Three subgroups of class II bacteriocins were included in Klaenhammer's classification (see Table 1). Nes et al., (1996) regrouped the class II bacteriocins, retaining class IIa and IIb but changing class IIc to include bacteriocins that contain a typical signal peptide and that are secreted by the general translocase (sec) pathway of the cell (Pugsley, 1993; Economou, 1998). Nes et al., (1996) excluded class IV bacteriocins because they have not been chemically characterized. In the classification system proposed by van Belkum & Stiles (2000), the class II bacteriocins were subdivided based on the number of cysteine residues, using the nomenclature proposed by Jack et al., (1995). In their classification, the bacteriocins that are secreted by the prepeptide translocase (sec) pathway (type IIc of Nes et al., 1996) are not classified as a separate group because of the diversity of the bacteriocins that are secreted by this pathway. However, some of the criteria to establish bacteriocin classification can be misleading when applied to enterocins. As an example, two-peptide bacteriocins can be found in several classes. Although the two peptides usually act synergistically, in several cases each of the components was shown to have individual antimicrobial activity (Garneau et al., 2002). Disulphide bridges can also be found in different bacteriocins, and they probably stabilize different bacteriocin structures (sometimes with no homologies) depending on their position in the primary sequence. In the present study, comparison of sequence similarities, together with the type of post-translational modification, have been chosen as the main criteria to arrive at a more natural classification of bacteriocins according to their relatedness. A simplified classification scheme is proposed for enterocins, including three classes: Class I enterocins (lantibiotic enterocins), Class II enterocins (small, nonlantibiotic peptides), Class III enterocins (cyclic enterocins) and Class IV enterocins (large proteins). Class II can be subdivided into three subclasses: II.1, enterocins of the pediocin family; II.2, enterocins synthesized without a leader peptide; and II.3, other linear, nonpediocin-type enterocins.

View this table:
Table 1

Comparison of classification systems for bacteriocins of lactic acid bacteria

ClassKlaenhammer et al. (1993)Nes et al., (1996)van Belkum & Stiles (2000)Proposed classificationExample
ILantibioticsLantibioticsLantibioticsClass I. Lantibiotic enterocins Class II enterocinsCytolysin
IIaPediocin-like or ‘Listeria’-active with YGNGVXC motif near N-terminus and GG leader peptidePediocin-like or ‘Listeria’-active with YGNGVXC motif near N-terminus and GG leader peptideCystibiotics with two disulphide bridges with YGNGVXC motif near N-terminusII.1. Enterocins of the pediocin familyEnterocin A
bTwo-component peptides with GG leader peptideTwo-component peptides with GG leader peptideCystibiotics with one disulphide bridge with YGNGVXC motif near N-terminusII.2. Enterocins synthesized without a leader peptideEnterocin L50A
cThiol-activated peptides with GG leader peptideBacteriocins secreted by signal peptide pathwayCystibiotics with one disulphide bridge but without YGNGVXC motif near N-terminusII.3. Other linear nonpediocin-like enterocinsEnterocin B
dThiolbiotics with one or no cysteine residues
eTwo-component peptides with GG leader peptide
fAtypical bacteriocinsClass III. Cyclic antibacterial peptidesEnterocin AS-48
IIILarge, heat-labile proteinsLarge, heat-labile proteinsLarge, heat-labile proteinsClass IV. Large proteinsEnterolysin A
IVProtein complexes
  • * Terminology from Jack et al., (1995).

Class I. Lantibiotic enterocins

Cytolysin produced by E. faecalis is the only lantibiotic-type enterocin currently known (Booth et al., 1996). Cytolysin is a two-peptide bacteriocin and both structural subunits contain lanthionine residues (Booth et al., 1996). Cytolysin is haemolytic and it is active against eukaryotic cells (erythrocytes) and Gram-positive bacteria (Gilmore et al., 1994; Booth et al., 1996). The haemolytic phenotype is more often associated with clinical than environmental isolates of enterococci and, based on animal pathogenicity models, it is considered that haemolysin may contribute to bacterial virulence (Ike et al., 1984, 1987; Jett et al., 1992, 1994; Coque et al., 1995).

The genetic locus for cytolysin production is on a 58-kb pheromone-responsive plasmid pAD1 with five known ORFs (Gilmore et al., 1994). Expression of all five ORFs is required for production of haemolytic and bacteriocin activities (Gilmore et al., 1994). The cylLL and cylLS ORFs encode the two ribosomally synthesized peptides. Both peptides are required for activity and are presumably modified intracellularly by the product of cylM that is similar to genes associated with post-translational modification of other lantibiotics (Gilmore et al., 1994). The modified, lanthionine-containing cytolysin subunits CylLL* and CylLS* contain leader peptides that are typically found in class II bacteriocins. They are exported from the cell by the product of cylB, which is similar to ABC transporter proteins and contains a sequence in the N-terminus that resembles the proteolytic domain of ABC transporters for class II bacteriocins (Håvarstein, 1995; Booth et al., 1996). During translocation, the leader peptides are cleaved by CylB and outside of the cell the cytolysin subunits are subject to proteolytic activation mediated by CylA, a serine protease, to yield the fully active subunits CylLL” and CylLS” with a molecular mass of 3437.98 and 2031.81 Da, respectively (Gilmore et al., 1994; Booth et al., 1996) (Table 2). It has been shown that CylLL” and CylLS” form a complex that dissociates when CylLL” binds to target cells enabling CylLS” to autoinduce cytolysin expression (Coburn et al., 2004).

View this table:
Table 2

Amino acid sequence, molecular weight and proposed classification of mature enterocins

EnterocinAmino acid sequence of mature peptideMolecular weight (Da)
Class I. Lantibiotic enterocins
Class II enterocins
Class II.1. Enterocins of the pediocin family
Class II.2. Enterocins synthesized without a leader peptide
Class II.3. Other linear nonpediocin-like enterocins
Class III. Cyclic antibacterial peptides
Class IV. Large proteins
  • Blue, amino acid residues conserved in all bacteriocins of the same group.

  • Red, residues conserved to a variable degree.

  • * Theoretical molecular weight was calculated from deduced amino acid sequence with DNAstar programme.

  • RJ-11, MR10A and MR10B: leader peptide not determined.

  • Putative sequence of the mature peptide deduced from the bacteriocin structural gene.

Because cytolysin is a two-component bacteriocin consisting of two linear peptides, it is structurally different to other linear (group A) lantibiotics, such as nisins A and Z, that consist of only one linear peptide. It is also different to the smaller and globular (group B) lantibiotics, which are predominantly produced by Streptomycetes (Jung et al., 1991; de Vos, 1995; Sahl et al., 1995). Nevertheless, the fact that this bacteriocin contains lanthionine residues dictates that it should be considered as a two-component lantibiotic.

Class II enterocins

Enterocins of the pediocin family

Class II.1 bacteriocins or pediocin-like bacteriocins are a well-known group of antimicrobial peptides. All of them contain a hydrophilic cationic region with the conserved YGNGVXC ‘pediocin box’ motif and two cysteine residues joined by an S–S bridge stabilizing the formed β-sheet structure. It has been shown with phospholipid vesicles that electrostatic interactions of positively charged N-terminal residues and not the YGNGVXC-consensus motif govern binding of pediocin PA-1 (Chen et al., 1997). However, substitutions made in the YGNGVXC motif of pediocin PA-1 and carnobacteriocin B2 showed reduced antibacterial activity against target organisms (Quadri et al., 1997, Miller et al., 1998).

The C-terminal part of pediocin-like bacteriocins is more diverse in amino acid sequence, with three recognized subgroups (Fimland et al., 2005). Peptides from subgroups 1 and 2 form a C-terminal hairpin-like structure stabilized by cysteine and/or tryptophan residues that are present at or near the C-terminal end, while peptides from subgroup 3 lack such structure-stabilizing residues. However, all of them contain a conserved tryptophan residue in the middle of the C-terminal part that may position at the membrane interface and stabilize the bacteriocin structure while it interacts with the bacterial membrane.

Class II.1 enterocins cluster in two subgroups according to sequence similarities, corresponding to the subgroups 1 and 3 described by Fimland et al., (2005) (Fig. 1). The first subgroup includes enterocin A, the mundticins and enterocin CRL35. The second subgroup is represented by bacteriocin 31, bacteriocin T8, bacteriocin RC714, enterocin SE-K4 and enterocin P. Bacteriocins from subgroup 1 contain seven unique conserved amino acid residues that differentiate them from subgroup 2 bacteriocins (which in turn contain six unique conserved amino acid residues, including Ala and Thr as first and second N-terminal positions) (Fig. 2a). At genetic levels, while bacteriocins from the first subgroup include ABC-type transporter proteins involved in bacteriocin export, the second subgroup of bacteriocins are exported by the bacterial preprotein translocase and their genetic determinants include basically the bacteriocin structural gene plus an immunity protein (Fig. 2b).

Figure 1

Cluster analysis of enterococal bacteriocins according to sequence homology. Consensus amino acids are indicated in blocks.

Figure 2

Similarities and differences between enterocins included in class II.1. (a) Comparison of amino acid sequences of representative class II.1 enterocins. Identical residues are in bold. Other highly conserved residues are denoted by x. Particular cysteine (C) and tryptophane (W) residues found at the C-terminal region are also shown. Conserved residues common to subgroups 1 (represented by enterocin A) and 2 (represented by bacteriocin T8) as well as the consensus sequence are shown. Unique conserved residues differentiating enterocins from subgroups 1 and 2 are shown in blue and red, respectively. (b) Simplified representation of genetic determinants of class II.1 enterocins. Genes encoding for bacteriocin prepeptides (blue), immunity (green) and ABC transporters (yellow) are indicated. ORFs are not drawn to scale.

Enterocin A (EntA) can be unequivocally grouped as a class II.1 bacteriocin (Table 1). It contains the characteristic YGNGVXC-motif at the N-terminus and four cysteine residues (Table 2, Fig. 2a) that are hypothesized to form two disulphide bridges. These characteristics make it equivalent to pediocin PA-1/AcH (Marugg et al., 1992). EntA is produced by several E. faecium strains: CTC492, T136 and P21, which were isolated from Spanish fermented sausages (Aymerich et al., 1996; Casaus et al., 1997; Herranz et al., 2001); by strain BFE 900 from black olives; by strains DPC 1146, WHE 81 and EFM01 from dairy sources (Franz et al., 1999b; O'Keeffe, 1999; Ennahar & Deschamps, 2000; Ennahar et al., 2001); and by N5 from nuka (a Japanese rice-bran paste) (Losteinkit et al., 2001). EntA is active against Enterococcus, Lactobacillus and Pediococccus spp., as well as Listeria spp., including L. monocytogenes. Activity against L. monocytogenes is a characteristic feature of class IIa bacteriocins (Klaenhammer et al., 1993; Nes et al., 1996).

EntA consists of 47 amino acids with a theoretical molecular weight of 4829 Da (Table 2). EntA is produced as a prepeptide with an 18-amino-acid double-glycine-type leader peptide (Table 3) (Aymerich et al., 1996), indicating that it is secreted by the products of an ABC transporter and accessory protein genes (Nes et al., 1996). An ORF that immediately follows the EntA structural gene encodes the 103-amino-acid EntA immunity protein (O'Keeffe, 1999). It shares 40% identity with a similar ORF in the leucocin A operon that encodes the leucocin A immunity protein (van Belkum & Stiles, 1995; Aymerich et al., 1996). Five other ORFs are involved in EntA production: entF, K, R, T and D that encode an induction factor, protein histidine kinase, response regulator, ABC transporter and accessory proteins, respectively (O'Keeffe, 1999) (Fig. 2b). The genes entF, K and R are similar to those of a three-component regulatory system (Nes et al., 1996; O'Keeffe, 1999), indicating that EntA production is regulated. Although the genes for response regulator and histidine kinase proteins were not detected for E. faecium CTC492, the gene for the induction factor (EntF) was located on the chromosome and sequenced (Nilsen et al., 1998). It encodes a prepeptide with a characteristic double-glycine-type leader peptide of 16 amino acids and mature EntF consists of 25 amino acids. However, O'Keeffe et al., (1999) reported a longer, 23-amino-acid leader peptide for EntF, and suggested that the 16-amino-acid leader peptide reported by Nilsen et al., (1998) was erroneous and may be the result of a sequencing error. A chemically synthesized, mature EntF peptide induced bacteriocin production by E. faecium CTC492 and DPC1146 that no longer produced bacteriocin as a result of dilution in liquid medium (Nilsen et al., 1998; O'Keeffe, 1999), confirming its role as an induction factor. The double-glycine-type leader peptide for EntA indicates that ABC transporter gene (entT) and accessory gene (entD) products comprise a dedicated transport system for this bacteriocin (O'Keeffe, 1999).

View this table:
Table 3

Amino acid sequence and type of enterocin N-terminal extensions

EnterocinN-terminal extension sequenceType
Mundticin KSMKKLTAKEMSQVVGGdouble-glycine-type leader peptide
Enterocin CRL35MKKLTSKEMAQVVGGdouble-glycine-type leader peptide
Enterocin AMKHLKILSIKETQLIYGGdouble-glycine-type leader peptide
Enterocin BMQNVKELSTKEMKQIIGGdouble-glycine-type leader peptide
Enterocin 1071AMKQYKVLNEKEMKKPIGGdouble-glycine-type leader peptide
Enterocin 1071BMKNIKNASNIKVIEDNELKAITGGdouble-glycine-type leader peptide
Bacteriocin 31MKKKLVICGIIGIGFTALGTNVEAsignal peptide
Bacteriocin 32MKKTKLLVASLCLFSSLLAsignal peptide

Mundticin produced by E. mundtii ATO6 isolated from processed vegetables (Bennik et al., 1998) is virtually identical to Mundticin KS produced by E. mundtii NFRI 7393 (Kawamoto et al., 2002), except for the last two C-terminus amino acids, which are reversed (Table 2). It consists of 43 amino acids and contains the YGNGVXC-consensus motif, as well a single disulphide bridge at the N-terminus of the molecule. It lacks the C-terminal stabilizing disulphide bridge, although it contains two conserved tryptophane residues at the central and C-terminal region (Fig. 2a). Mundticin has a molecular weight of 4287 Da, as determined by mass spectrometry (Bennik et al., 1998). The mature bacteriocins contain 43 amino acids (Table 2) and mundticin KS is produced as a preprotein bearing a 15-amino-acid leader peptide of the double-glycine type (Kawamoto et al., 2002) (Table 3). The activity spectra of these mundticins include Enterococcus, Lactobacillus, Leuconostoc and Pediococcus spp., as well as the food-borne pathogens L. monocytogenes and C. botulinum. Differences in the activity spectra of these bacteriocins, especially against Lactobacillus species, are claimed to be due to sequence differences (Kawamoto et al., 2002). The mundticins exhibit sequence similarity to class II bacteriocins such as piscicolin 126, sakacin P, pediocin PA-1, leucocin A, mesentericin Y105, carnobacteriocins BM1 and B2, curvacin A and sakacin A (Bennik et al., 1998).

The gene cluster for mundticin KS production was cloned and shown to consist of three genes designated munA, munB and munC that encode the mundticin KS preprotein, a 674-amino-acid ABC transporter and a 98-amino-acid immunity protein, respectively (Fig. 2b). The MunB transporter has homology with the ABC transporters for enterocin A and carnobacteriocin B2, while the immunity protein has homology with the piscicolin 126 and sakacin P immunity proteins (Kawamoto et al., 2002). Interestingly, a gene for an accessory protein for bacteriocin transport was not found on the 5.3-kb sequenced DNA fragment that contained the mundticin KS gene locus. Moreover, Kawamoto et al., (2002) showed that the bacteriocin could be successfully expressed in the heterologous hosts E. faecium, Lactobacillus curvatus and Lactococcus lactis by cloning of the munA, munB and munC genes, indicating that indeed an accessory protein is not required for the secretion of mundticin KS from the Gram-positive cell (Kawamoto et al., 2002). This is a novel situation for class II bacteriocin transport systems, because an accessory protein was considered an absolute requirement for secretion of these types of bacteriocins (Nes et al., 1996).

Enterocin CRL35 produced by E. mundtii CRL35 (previously reported as E. faecium) isolated from an Argentinean artisanal cheese also contains the YGNGV motif near the N-terminus (Farías et al., 1996). Enterocin CRL35 is a 43-amino-acid peptide with molecular mass of 4900 Da. The peptide sequence was recently determined (Saavedra et al., 2004), being identical to mundticin KS (Table 2). The leader peptide of enterocin CRL35 differs by two amino acid residues from that of mundticin KS (Table 3). Nucleotide sequence analysis of the enterocin CRL35 gene cluster revealed an identical gene organization with the mundticin KS cluster (Saavedra et al., 2004). The ABC transporter and immunity protein for enterocin CRL35 showed 96% and 98% identity, respectively, with MunB and MunC. Enterocin CRL35 shows both antilisterial and antiviral activity (Wachsman et al., 1999).

Bacteriocin 31. This enterocin is produced by E. faecalis YI717, which was isolated from a clinical sample (Tomita et al., 1996). The genetic determinant is located on a pheromone-responsive plasmid. Bacteriocin 31 is secreted by the bacterial preprotein translocase and contains a 24-amino-acid signal peptide. The mature bacteriocin consists of 43 amino acids (Table 2) and it is active against Enterococcus spp. and L. monocytogenes. Genetic analysis showed that the structural gene for bacteriocin 31 was located in an operon containing three ORFs (Fig. 2b). The ORF immediately downstream of the structural gene encodes the 94-amino-acid immunity protein. The function of the third ORF located downstream of the immunity protein gene is not known, but it has homology with ORF-X2, the presumptive immunity gene for carnobacteriocin BM1 that is located immediately downstream of the structural carnobacteriocin BM1 gene (Quadri et al., 1994; Tomita et al., 1996).

Bacteriocin 31 contains an N-terminal YGNGLXC sequence that differs from the conserved YGNGVXC-motif of other class IIa or IIb bacteriocins of the classification according to Nes et al., (1996). Nevertheless, bacteriocin 31 shows sequence similarity to other class II bacteriocins such as sakacins A and P, curvacin A, carnobacteriocin BM1 and leucocin A (Tomita et al., 1996). The consequence of the deviation from the YGNGVXC-consensus motif is not clear. Bacteriocin 31 clusters in a subgroup with the highly homologous bacteriocin RC714, bacteriocin T8 (which is identical to bacteriocin 43 and to hiracin JM79), and also with enterocin SE-K4 and enterocin P (Fig. 1).

Bacteriocin RC714 was isolated from a vancomycin-resistant E. faecium strain (Del Campo et al., 2001). The peptide sequence contains 42 amino acid residues and shows 88% identity with bacteriocin 31 (Table 2). Although the genetic determinants have not been described, bacteriocin activity cotransferred together with vancomycin resistance to E. faecalis JH2-2. Bacteriocin T8/bacteriocin 43. The bacteriocin T8 is produced by E. faecium T8 isolated from vaginal secretion of a child infected with HIV (de Kwaadsteniet et al., 2006). At the amino acid level, it is 98% homologous to bacteriocin RC714 (being two amino acid residues longer) and 88% homologous to bacteriocin 31. The bacteriocin is produced as a 74-amino-acid prepeptide bearing a 30-amino-acid signal peptide that shares homology at its C-terminal end with the signal peptide of bacteriocin 31 and also shares the homologous KVDA sequence with enterocin P at the processing site (Table 3). The 44-amino-acid mature bacteriocin has a calculated molecular mass of 5.1 kDa (de Kwaadsteniet, 2006). Bacteriocin T8 has a conserved YGNGLXC where the conserved valine of the typical YGNGVXC box sequence is replaced by a leucine, similar to the case of bacteriocin 31 (see above). Furthermore, it has two cysteine residues in the N-terminal half of the molecule and two tryptophane residues near its C-terminal end (Fig. 2a). The bacteriocin T8 structural gene is located on a 7-kb plasmid and is followed downstream by an ORF that could encode a 95-amino-acid peptide (Fig. 2b), and that shows 50% homology to the immunity protein of bacteriocin 31 (de Kwaadsteniet, 2006). Further bacteriocin transport or regulatory genes were not determined (de Kwaadsteniet, 2006). Bacteriocin T8 was active on uropathogenic strains of E. faecalis, a Lactobacillus sp., two Enterococcus sp., and a Propionibacterium sp.

The vancomycin-resistant clinical isolate E. faecium VRE82 produces an antimicrobial peptide that has been named bacteriocin 43 (Todokoro et al., 2006). The structural gene for bacteriocin 43 (bacA) encodes for a precursor of 74 amino acid residues, and is followed by an ORF (bacB) encoding an immunity protein of 95 amino acids. A blast search carried out on the NCBI protein database (http://www.ncbi.nlm.nih.gov/BLAST/) reveals that the deduced products of bacA and bacB are 100% identical to the deduced structural and immunity proteins, respectively, of bacteriocin T8 described by de Kwaadsteniet et al., (2006) and hiracin JM79 produced by Enterococcus hirae DCH5 isolated from mallard ducks (Anas platyrhynchos) (J. Sánchez et al., unpublished data: NCBI protein database, accession numbers ABG47453 and ABG47454). The genetic determinants of bacteriocin 43 are located on the mobilizable plasmid pDT1 (6.2 kb). A 737-bp DNA fragment containing bacA and bacB conferred bacteriocin production and immunity to enterococcal transformants, suggesting that no auxiliary genes are needed for production of this bacteriocin. The genetic determinants bacA and bacB were also shown to be present in pDT1-like plasmids of 21 vancomycin-resistant E. faecium clinical isolates. Todokoro et al., (2006) reported that bacteriocin 43 is active against E. faecalis, E. faecium, E. hirae, E. durans and L. monocytogenes.

Enterocin SE-K4 is produced by an E. faecalis strain isolated from silage (Eguchi et al., 2001). It was purified to homogeneity and found to be a 43-amino-acid bacteriocin with an experimentally determined mass of 5356.2 Da (Eguchi et al., 2001). It shows a high degree of homology to bacteriocins 31 and T8 (Table 2). The 60-kb plasmid pEK4S is involved in the production of and immunity to enterocin SE-K4 (Doi et al., 2002). This bacteriocin shows antimicrobial activity against Gram-positive bacteria, E. faecium, E. faecalis, Bacillus subtilis, C. beijerinckii and L. monocytogenes.

Enterocin P is produced by E. faecium P13, AA13, G16 and L50 that were isolated from Spanish fermented sausages (Cintas et al., 1997, 2000; Herranz et al., 1999). It has a broad antimicrobial spectrum that includes activity against Enterococcus, Lactobacillus and Pediococcus spp., as well as the food-borne pathogens B. cereus, C. botulinum, C. perfringens, L. monocytogenes and S. aureus (Cintas et al., 1997). Mature enterocin P consists of 44 amino acids and based on the nucleotide sequence it has a theoretical molecular weight of 4493 Da (Table 2). Enterocin P is produced as a prepeptide of 71 amino acids with a 27-amino-acid signal peptide (Table 3). When aligned with the signal peptide of bacteriocin 31, the enterocin P signal peptide was shown to share sequence identity of 12 amino acids (Cintas et al., 1997). Secretion of enterocin P, therefore, occurs by the bacterial preprotein translocase (Cintas et al., 1997; Herranz & Driessen, 2005). A 755-bp chromosomal DNA fragment containing the enterocin P structural gene was cloned and sequenced. This fragment contains two ORFs arranged in an operon (Fig. 2b). The ORF immediately downstream of the EntP structural gene encodes an 88-amino acid, putative immunity protein (Cintas et al., 1997). Enterocin P has sequence similarity with other class II bacteriocins containing the YGNGVXC-motif (Cintas et al., 1997). It shares several conserved amino acid residues with bacteriocin 31, including a tryptophan residue (Trp35) at its C-terminal part (Fig. 1).

II.2. Enterocins synthesized without a leader peptide

This group includes several bacteriocins that lack the YGNGVXC motif and are synthesized as leaderless peptides requiring dedicated export systems. A main subgroup includes the two-peptide bacteriocins enterocin L50 and the highly related peptides MR10A and MR10B (Fig. 1). Other leaderless bacteriocins consisting of individual peptides, such as enterocin RJ-11, enterocin Q and enterocin EJ97 are also included in this group. All of them share conserved Lys10 and Lys21 residues (Lys8 and Lys19 in enterocin RJ-11). Leaderless enterocins are thought to be exported through dedicated transport systems involving ABC transporters, similarly to other leaderless bacteriocins such as aureocin A70 from S. aureus (Netz et al., 2001) and bacteriocin LsbB from Lactococcus lactis BGM-1 (Gajiç et al., 2003). Bacterial ABC transporters involved in export are often composed of an ATP-binding protein and one or more accessory proteins (Fath & Kolter, 1993; Håvarstein et al., 1995). ABC transporters could also contribute to bacteriocin immunity by pumping bacteriocin molecules out of the cells (Diaz et al., 2003; Sánchez-Hidalgo et al., 2003). The presence of genes encoding for ABC transport proteins in leaderless enterocins is indicated in Fig. 3.

Figure 3

Simplified representation of genetic determinants of leaderless enterocins. Genes encoding for bacteriocin peptides (blue) and ABC transporters (yellow) are indicated. The ABC transporter system homologous to the enterocin AS-48 as48EFGH system is marked by dashed lines. ORFs are not drawn to scale.

Enterocin L50 is a plasmid-encoded bacteriocin that is produced by E. faecium L50 and 6T1a (Cintas et al., 1998; Floriano et al., 1998), and also by E. faecium F58 isolated from jben, a Moroccan traditional cheese (Achemchem et al., 2005). Cintas et al., (1998) showed by in vitro transcription and translation that enterocin L50 consists of two peptides, L50A and L50B, which individually have antimicrobial activity and exhibit synergy when combined. The genetic loci for enterocin L50 are identical in strains L50 and 6T1a, and the 44-amino-acid peptide (EntL50A and EntI) and 43-amino-acid peptide (EntL50B and ORF2) have the same theoretical molecular mass of 5190 and 5178 Da (Table 2), respectively (Cintas et al., 1998; Floriano et al., 1998). The A and B peptides have 31 amino acid residues in common and they are 72% identity. EntL50 exhibits a wide antimicrobial spectrum and it is active against strains of Enterococcus, Lactobacillus, Lactococcus lactis, Pediococcus pentosaceus as well as the food-borne pathogens L. monocytogenes and B. cereus (Cintas et al., 1998; Floriano et al., 1998).

EntL50A and B, similar to enterocins Q and EJ97, are produced as leaderless peptides. Although enterocin L50 may at first glance be considered a class IIb bacteriocin in the Klaenhammer (1993) and Nes et al., (1996) classification systems, it has more in common with a group of staphylococcal peptide toxins which include δ-lysin, SLUSH A to C and AGS1 to 3, produced by S. aureus, S. lugdunensis and S. haemolyticus, respectively (Cintas et al., 1998). Indeed, when amino acid sequences of enterocins L50A and B are aligned they exhibit low but significant sequence homology to these peptide toxins (Cintas et al., 1998). Based on these observations Cintas et al., (1998) suggested that EntL50 and the staphylococcal peptides constitute a separate, hitherto unrecognized, family of peptide toxins. The structural genes were not cotranscribed with an immunity gene (Cintas et al., 1998). According to recent reports, the genetic determinants for EntL50A and B are located on the 50-kb plasmid pC17 in E. faecium L50 (Criado et al., 2006), while in strain 6T1a they are located on the 21-kb plasmid pEF1, in which the bacteriocin cluster has been suggested to act as a functional plasmid stabilization module (Ruiz-Barba et al., 2006). The bacteriocin region of pEF1 contains 13 ORFs (Fig. 3), including the two bacteriocin structural genes and a set of four genes (as-48EFGH) that are highly homologous to the second ABC transport system described for the cyclic bacteriocin AS-48 (Diaz et al., 2003).

Enterocin RJ-11 is an enterocin with a calculated molecular weight of 5049 Da produced by E. faecalis RJ-11 that was isolated from rice bran. It is also highly homologous to EntL50A (Yamamoto et al., 2003). However, as genetic data are lacking at present it is not known whether the RJ-11 structural gene encodes for any leader peptide.

Recently, two antimicrobial peptides (enterocins MR10A and MR10B) with molecular masses of 5201.58 and 5207.7 Da, respectively, were purified from an E. faecalis strain isolated from the uropigial gland of the hoopoe (Upupa epops) (Martin-Platero et al., 2006). These enterocins are highly homologous to enterocins L50A and L50B. The corresponding sequences are almost identical, except that there is a conservative change (Glu 38 to Asp) in MR10A and two changes in residues (Thr 9 to Ala, and Leu 15 to Phe) in MR10B. The two peptides show a broad antimicrobial spectrum and also act synergistically. It is not known whether these bacteriocins are also synthesized without a leader peptide, but their genetic determinants seem to be located on the bacterial chromosome (Martin-Platero et al., 2006).

Enterocin Q is produced by E. faecium L50 in addition to enterocins P and L50 (Cintas et al., 2000). Enterocin Q is relatively small (34 amino acids) and it has a theoretical molecular mass of 3952 Da (Table 2) (Cintas et al., 2000). Similar to EntL50, it is not produced as a prepeptide. Enterocin Q does not contain the YGNGVXC consensus sequence, but it does contain two cysteine residues (Cys26 and Cys34) and it is suggested that a disulphide bridge is formed between these cysteine residues (Cintas et al., 2000). Contrary to other bacteriocins such as enterocin B and carnobacteriocin A, the disulphide bridge does not span the N- and C-sections of the molecule or just the N-terminal. Instead, it is located in the C-terminal half of the molecule. The structural gene of enterocin Q (entqA) is located on the 7.4-kb plasmid pCIZ2 of E. faecium L50, and is preceded by two divergently orientated genes, entqB and entqC (Criado et al., 2006) (Fig. 3). EntqB is highly similar to ABC transporter proteins and might be involved in export of enterocin Q from the cell, similar to other bacterial proteins that do not contain an N-terminal leader or signal peptide. EntqC encodes a putative 67-amino-acid residue peptide with two probable transmembrane segments, but it has no significant homology to known proteins.

Enterocin EJ97 is produced by E. faecalis EJ97 that was isolated from municipal waste water. EntEJ97 has a broad range of activity against Gram-positive bacteria including other enterococci, several species of Bacillus, Listeria as well as strains of S. aureus (Gálvez et al., 1998). Mass spectrometry indicated that the purified bacteriocin has a mass of 5340 Da. The genes for EJ97 production were detected on a 60-kb conjugative, pheromone-response plasmid pEJ97, and the structural gene encodes a 44-amino-acid bacteriocin (Table 2) that lacks an N-terminal extension (Sánchez-Hidalgo et al., 2003). EntEJ97 shows some homology to other bacteriocins in this group, including three conserved lysine residues (Lys6, Lys10, Lys21), a conserved Ile27, as well as a conserved tryptophan residue close to its C-terminal end (Trp39). The structure – function relationship of these conserved residues is unknown. Genes encoding putative transport proteins such as an ABC-transporter and putative accessory proteins were found in close proximity of the Ent EJ97 structural gene (Fig. 3) and may be involved in transport and possibly immunity (Sánchez-Hidalgo, 2003).

II.3. Other linear, nonpediocin-type enterocins

This group includes unrelated linear bacteriocins synthesized with a leader peptide. In some cases, their antimicrobial activity may be reinforced by the presence of a second peptide (such as enterocin 1071), although single-peptide bacteriocins are also included.

Enterocin 1071 produced by E. faecalis BFE1071 and FAIR-E 309 (Balla et al., 2000; Franz et al., 2002) consists of two peptides, enterocin 1071A and 1071B. These bacteriocins are active against a broad spectrum of Gram-positive bacteria including strains of Clostridium, Enterococcus, Lactobacillus, Propionibacterium, Streptococcus, Micrococcus and Listeria spp. The genetic determinants for this two-peptide bacteriocin were located on plasmid DNA and DNA sequencing revealed two ORFs encoding the 39- and 34-amino-acid A and B peptides, respectively (Table 2) (Balla et al., 2000; Franz et al., 2002). These bacteriocins are encoded as prepeptides, each bearing a double-glycine-type leader peptide of 18 amino acids (Ent1071A) or 24 amino acids (Ent 1071B) (Franz et al., 2002) (Table 3). Enterocins 1071A and B share sequence similarity only with the α and β subunits of the two-peptide bacteriocin lactococcin G (Franz et al., 2002). Downstream of ent1071B, a gene eni1071 encoding a protein of 110 amino acids with an isoelectric point of 9.278 was found. It was cloned and expressed in a heterologous host and shown to confer immunity to enterocin 1071 (Franz et al., 2002). An ABC transporter gene that has similarity to other bacteriocin ABC transporter genes and an accessory protein gene were located upstream of, and on the opposite DNA strand relative to the enterocin structural genes (Franz et al., 2002; Balla & Dicks, 2005). Furthermore, by insertional inactivation of the bacteriocin structural genes, Balla & Dicks (2005) showed that Ent1071A and Ent1071B could act independently against target bacteria.

Enterocin B is an additional bacteriocin that is produced by several strains of enterococci that produce enterocin A, for example by E. faecium CTC492, T136, WHE 81 and BFE 900 (Aymerich et al., 1996; Casaus et al., 1997; Franz et al., 1999b; Ennahar et al., 2001). EntB does not contain the YGNGVXC-consensus motif at the N-terminus of the peptide and it shares sequence similarity only with carnobacteriocin A (Table 2) (Worobo et al., 1994; Casaus et al., 1997; Franz et al., 1999b). An unusual feature of EntB and carnobacteriocin A is that their immunity genes do not occur immediately downstream of, and in the same operon as, the structural gene, which is the case for most other class II bacteriocins. Instead, their immunity genes are located on the opposite DNA strand, immediately downstream of the structural genes for CbnA and the EntB gene (Franz et al., 1999b, 2000). Both EntB and CbnA contain two cysteine residues which could form a disulphide bridge in the mature peptide (Worobo et al., 1994; Franz et al., 1999b), but unlike other class II bacteriocins such as enterocin P, bacteriocin 31, leucocin A, carnobacteriocin B2 and sakacin P, the disulphide bridge is not contained in the N-terminal half of the molecule, but rather it spans positions 23 and 52 of the bacteriocin molecule (van Belkum & Stiles, 2000). Enterocin B consists of 53 amino acid residues with a molecular weight of 5463 Da (Table 2). Genetic analysis showed that the prepeptide contains an 18-amino-acid leader peptide of the double-glycine type (Table 3) (Casaus et al., 1997; Franz et al., 1999b). This indicates that EntB is secreted by a dedicated-type secretion mechanism; however, a 12.0-kb DNA fragment cloned from the chromosome of E. faecium BFE 900 that contained the structural gene for EntB did not contain ABC transporter or accessory protein genes (Franz et al., 1999b). It is possible that EntB is secreted by the dedicated transport proteins of EntA or some other transport system available in the cell (Franz et al., 1999b). EntB is active against Enterococcus and Lactobacillus spp. and the food-borne pathogens L. monocytogenes, S. aureus and C. perfringens (Franz et al., 1996; Casaus et al., 1997).

Bacteriocin 32. This bacteriocin was detected in the vancomycin-resistant strain Enterococcus faecium VRE200 (Inoue et al., 2006). The Bac32 genetic locus consists of a bacteriocin gene (bacA) and an immunity gene (bacB). The deduced bacA product is 89 amino acids in length, with a putative signal peptide of 19 amino acids (Inoue et al., 2006). Neither of the bacteriocin and immunity gene products showed significant homology to any known bacteriocins or immunity proteins. Bacteriocin 32 was active against E. faecium, E. hirae and E. durans but showed no activity against L. monocytogenes.

Class III enterocins: cyclic antibacterial peptides

Kemperman et al., (2003) suggested that a class V should be added to the classification systems of Klaenhammer (1993) and Nes et al., (1996) to include bacteriocins that are ribosomally synthesized, post-translationally unmodified and head-to-tail ligated, cyclic antibacterial peptides. Separating cyclic peptide bacteriocins from other class II bacteriocins was also suggested by van Belkum & Stiles (2000). Several cyclic antibacterial peptides have been described to date, including enterocin AS-48 from E. faecalis (Martínez-Bueno et al., 1994), gassericin A from Lactobacillus gasseri LA39 and the similar reutericin 6 from Lactobacillus reuteri LA6 (Kawai et al., 2004), butyrivibriocin AR10 from the ruminal anaerobe Butyrivibrio fibrisolvens AR10 (Kalmokoff et al., 2003), circularin A from C. beijerinckii ATCC 25752 (Kemperman et al., 2003) and subtilosin A from Bacillus subtilis (Kawulka et al., 2004). We propose to include cyclic antibacterial peptides produced by enterococci as Class III enterocins within the enteroccal bacteriocin classification scheme.

Enterocin AS-48 is produced by a clinical isolate of E. faecalis S-48 (Gálvezet al. , 1986). Production of identical bacteriocins, enterocin EFS2, enterocin 4 and bacteriocin 21 were subsequently reported for other strains of E. faecalis (Joosten et al., 1996; Maisnier-Patin et al., 1996; Tomita et al., 1997), as well as E. faecium 7C5 (Folli et al., 2003). The structural genes for enterocin AS-48 and bacteriocin 21 are located on the pheromone-responsive plasmids pMB2 and pPD1, respectively. They have been cloned and sequenced (Martínez-Bueno et al., 1994; Tomita et al., 1997). From the genetic analysis of the structural gene and the amino acid analysis of two peptides obtained after digestion of enterocin AS-48 with Glu-C endoproteinase, Martínez-Bueno et al., (1994) showed that the as-48 gene encodes a 105-amino-acid prepeptide consisting of a 35-amino-acid signal peptide (Table 3) and a 70-amino-acid mature peptide (Table 2). The signal peptide contains a typical, positively charged N-terminus followed by a hydrophobic domain and a polar region at the processing site; however, the histidine residue found at position -1 is atypical for a signal peptide cleavage site (Martínez-Bueno et al., 1994). It is possible that more than one cleavage step is involved in processing of the leader region.

These studies indicate that the peptide forms a 70-amino-acid cyclic molecule that resulted from a head-to-tail linkage of the N-terminal methionine (Met1) to the C-terminal tryptophan (Trp70) (Martínez-Bueno et al., 1994). Tomita et al., (1997) showed that production of bacteriocin 21 depends on the presence of nine ORFs (bacA to bacI), while Martínez-Bueno et al., (1998) and Diaz et al., (2003) reported that 10 ORFs (as-48A, B, C, C1, D, D1, E, F, G and H) are required for complete expression of AS-48 production and immunity. In both cases, all ORFs are orientated in the same direction. The presence of the structural gene for AS-48 and a gene encoding the C-terminal domain of an ABC transporter protein are common to the genetic loci (Tomita et al., 1997; Martínez-Bueno, 1998). Using homology comparisons, Martínez-Bueno et al., (1998) determined that as-48D encodes the functional, C-terminus domain of an ABC transporter, and they suggested that as-48C1 encodes the membrane spanning domain for this transporter protein. Using transposon mutagenesis with Tn5 in conjunction with deletion mutational analysis, they also determined that as-48D1 encodes the 56-amino-acid immunity protein for AS-48; however, this gene confers partial immunity and full immunity is only obtained when as-48B, C1, D, D1, E, F, G and H are expressed (Diaz et al., 2003). It was suggested that as-48B and C are involved with post-translational modification, signal peptide removal and head-to-tail cyclization (Martínez-Bueno et al., 1998). The genes as-48D, E, F and G encode a multicomponent ABC transporter that also has a function in immunity to AS-48 (Diaz et al., 2003). Thus, two ABC systems appear to be involved in AS-48 transport and immunity; whereas As-48C1D is devoted to cleavage and export of newly synthesized AS-48 as well as providing low-level immunity to AS-48, AS-48 EFGH is mainly related to self-protection, i.e. high level resistance to AS-48 (Diaz et al., 2003).

More recently, an AS-48 variant (AS-48RJ) has been characterized from an E. faecium strain isolated from home-made goat cheese. This strain contains a gene cluster identical to strain S-48, but in contrast to AS-48 it is located on the bacterial chromosome (Abriouel et al., 2005). AS-48RJ is identical to AS-48 except for Glu20 that is replaced by Val20 (Abriouel et al., 2005) (Table 2, Fig. 1).

Enterocin AS-48 has a three-dimensional structure that shows structural, but not sequence homology to NK-lysin, an antibacterial and tumorolytic polypeptide of T-cells and natural killer (NK) cells isolated from small intestine of a pig (González et al., 2000). These molecules consist of five α-helices that contain a so-called ‘saposin fold’ (González et al., 2000). Recently, a circular bacteriocin, circularin A, with homology to AS-48 at the primary sequence level was isolated from C. beijerinckii ATCC 25752 (Kemperman et al., 2003). Circularin A has 60% similarity (30% identity) with AS-48 and is produced by head-to-tail linkage (between Val4 and Tyr72).

Class IV enterocins

Enterolysin A is a bacteriocin produced by E. faecalis LMG 2333 and DPC5280 (Hickey et al., 2003, Nilsen et al., 2003). It is a large (calculated molecular weight of 34,501 Da), heat-labile bacteriocin and therefore fits the general characteristics of class IV bacteriocins (Table 2). Enterolysin A showed activity towards Lactobacillus, Lactococcus, Pediococcus, Enterococcus, Bacillus, Listeria and Staphylococcus spp. The genetic determinant for enterolysin A was cloned and sequenced (Hickey et al., 2003; Nilsen et al., 2003). The bacteriocin is encoded as a preprotein consisting of a 343-amino-acid polypeptide bearing a 27-amino-acid signal peptide (Table 3). Amino acid sequence comparisons indicated homology to several proteins in the public databases. The N-terminal region contained a domain found in the M37 family of metallopeptidases, while the C-terminal part showed similarity to different muralytic bacteriophage proteins (Hickey et al., 2003; Nilsen et al., 2003). Thus, the N-terminus of enterolysin A showed homology to lysostaphin, LytM and zoocin A produced by S. simulans biovar. staphylolyticus, S. aureus and Streptococcus zooepidemicus, respectively (Hickey et al., 2003; Nilsen et al., 2003). However, enterolysin A has a broader activity spectrum than lysostaphin, LytM and zoocin A (Hickey et al., 2003; Nilsen et al., 2003). Nilsen et al., (2003) also showed that enterolysin A, as indicated by the similarity to muralytic enzymes, possesses cell-wall hydrolysing activity. Sequence analysis of the bacteriocin showed that the protein consists of two separate domains, an N-terminal catalytic (endopeptidase) domain and a C-terminal substrate recognition domain, that are separated by a threonine- and proline-rich linker region (Hickey et al., 2003; Nilsen et al., 2003). A gene was found immediately downstream of the enterolysin A gene that could encode a putative transcription regulator (Hickey et al., 2003).

Ecological significance and practical importance

The enterocins are notable for their diversity and for the exceptional distribution among isolates from different sources. This diversity may be a reflection of the robust and ubiquitous nature of enterococci, as well as their remarkable ability to disseminate and receive genetic material. Enterolysin A and enterocin L50 have sequence similarity to staphylococcal lysostaphin and staphylococcal peptide toxins, respectively. The bacteriocin AS-48 has sequence similarity with circularin A produced by C. beijerinckii. This suggests that they may have resulted from genetic exchange between these genera. Exchange of genetic material between staphylococci and enterococci has been reported, as is shown for an E. faecalis strain producing a β-lactamase identical to that produced by S. aureus (Murray & Mederski-Samoraj, 1983; Murray et al., 1986). Well-described gene transfer mechanisms in enterococci include conjugative and nonconjugative plasmids, as well as conjugative transposons (Clewell, 1990). Of particular relevance to transfer of bacteriocin genes are the conjugative, pheromone-responsive plasmids. These generally have a narrow host range and transfer at high frequency (Clewell, 1990). Genetic determinants for cytolysin, bacteriocin 31, enterocin EJ97 and enterocin AS-48 are encoded on such plasmids. This gene transfer mechanism may promote fast and efficient transfer of these bacteriocin genes between enterococci.

Efficient gene transfer mechanisms may explain the wide variety of enterocins observed, the production of multiple bacteriocins by single strains and the multiple isolation of the same enterocins by different work groups as occurred for enterocin L50 and enterocin AS-48. Using a seminested PCR method, Joosten et al., (1997) showed that AS-48 may be widespread among strains of E. faecalis and E. faecium. PCR amplification was used to demonstrate the presence of enterocin L50A and B and enterocin A genes among enterococci isolated from pig faeces (Du Toit, 2000). Similarly, production of enterocin P was reported for a range of E. faecium-like strains isolated from Spanish dry fermented sausages (Herranz et al., 1999, 2001). De Vuyst et al., (2003) showed that at least one of the bacteriocin genes for EntA, EntB, EntL50A&B, Bac31 or cytolysin occurred in 64 of 122 bacteriocinogenic enterococci isolated from diverse ecological niches. Multiple bacteriocin production by enterococci also occurs frequently. Among the strains investigated by De Vuyst et al., (2003), 39.1% showed a positive PCR signal for at least one enterocin, 48.4% were positive for two different enterocins, 10.9% were positive for three different enterocins and one strain (1.6%) was positive for four different enterocins. Abriouel et al., (2006) also showed that the structural gene for EntP could be determined on plasmids of various food enterococci strains. These studies show that enterocin production is common among enterococci isolates and that the bacteriocin genes are often located on transmissible genetic determinants.

As is shown by cytolysin producing E. faecalis strains in the eukaryotic environment, bacteriocins may contribute a competitive advantage to the enterococci. The source of enterococci in human infection has been cited as either the patient's own microbial community from the gastrointestinal or urinary tracts, or person-to-person transmission by hospital workers (Franz et al., 1999a). Pheromone-responsive plasmids that carry bacteriocin phenotypes may also harbour antibiotic resistance genes (Clewell, 1990). Transferable antibiotic resistance and bacteriocin phenotypes may enable the enterococci to establish themselves and colonize the gastrointestinal tract of patients despite pressure from antibiotics and the protective effect of the patient's natural microbial community and allow colonization. However, other factors such as adherence, invasiveness and other virulence factors may also play a role in the establishment of infection (Jett et al., 1994; Franz et al., 1999a).

Production of bacteriocins may also play a role for enterococci to colonize foods, particularly dairy or meat product fermentations, and to become a major component of the microbial communities associated with these products. However, the success of enterococci in becoming the predominant population in fermented foods such as meats or cheese cannot be explained on the basis of bacteriocin production alone; other factors such as their tolerance to heat, drying and salt are well documented (Franz et al., 1999a). Production of bacteriocins, in combination with tolerance of a wide range of adverse chemical and physical conditions, may explain why these bacteria are so robust in nature and why they occur in such a wide variety of ecological niches.

Many enterocins have a broad activity spectrum, with many of them inhibiting enterococci and other LAB, as well as other Gram-positive bacteria, including Listeria, Staphylococcus and Clostridium species. The enterocins are generally active against L. monocytogenes and may also be active against C. tyrobutyricum, a microorganism associated with the ‘late blowing defect’ of cheese. Enterococci form a major component of the predominant microbial community of certain Mediterranean cheeses and they are important for ripening and flavour development (Trovatelli & Schiesser, 1987; Tsakalidou et al., 1993; Centeno et al., 1996; Franz et al., 1999a). The use of bacteriocin-producing enterococci as starter cultures in cheese production may prevent spoilage by C. tyrobutyricum or safeguard cheese from growth of L. monocytogenes (Torri Tarelli, 1994; Giraffa et al., 1995; Giraffa et al., 1997). Challenge studies using model dairy systems have demonstrated the potential for use of bacteriocinogenic enterococci to prevent Listeria growth (Sulzer et al., 1992; Rodriguez et al., 1997). Furthermore, bacteriocinogenic enterococci predominate at the end of the fermentation of Spanish dry fermented sausages (Herranz et al., 1999). These sausages are usually produced without added starter cultures (Herranz et al., 1999). Moreover, one strain of E. faecium isolated from these sausages produced three bacteriocins, enterocins P, L50 and Q (Cintas et al., 2000). This is a considerable arsenal of bacteriocins, which may confer a competitive advantage to this strain in the sausage fermentation. In addition to a competitive advantage of the bacteria in the ecosystem, the bacteriocinogenic enterococci may also serve as an anti-Listeria factor. Indeed, enterococci producing bacteriocins or purified enterocins have been used as protective cultures in the production of various foods in model studies (Aymerich et al., 2000; Callewaert et al., 2000). The use of bacteriocinogenic enterococci in (model) food fermentations is reviewed in Giraffa (2003) and Hugas et al., (2003).


As the above discussions show, our current understanding of enterococcal bacteriocins allows us to place these peptides in a new, simplified classification scheme, which is based mostly on structural differences and amino acid sequences. This enterocin classification scheme may serve as a model also for the classification of bacteriocins produced by other LAB. What is quite unique to the enterococci is the impressive array of different bacteriocins produced by members of this bacterial group. This can be explained by their robust nature, versatility to survive in a wide range of ecological niches and, ultimately, their superior genetic exchange mechanisms. As these bacteriocinogenic substances generally are inhibitory towards many food-borne pathogenic bacteria, the idea to make biotechnological use of either the producing strain or a purified bacteriocin in food preservation has received considerable interest (Giraffa et al. 2003; Hugas et al., 2003; Franz & Holzapfel 2004).

A drawback for the use of enterococci as starter cultures is that they do not have GRAS (generally recognized as safe) status (Giraffa et al., 1997), and they have increasingly been associated with nosocomial human infection (Franz et al., 1999a). Therefore, studies regarding the virulence potential of bacteriocinogenic Enterococcus starter cultures are required to guarantee the safety of such strains, especially because food strains have been shown to carry virulence determinants (Eaton & Gasson, 2001; Franz et al., 2001). Our studies have shown that the incidence of virulence traits among food isolates is strain specific (Franz et al., 2001). As a result, it is imperative that any bacteriocinogenic Enterococcus for use as a starter culture or as probiotic should be screened for potential virulence factors, as well as antibiotic resistance genes.

Recently, the European Food Safety Authority (EFSA) has taken responsibility to launch the European initiative towards a ‘Qualified Presumption of Safety’ (QPS) concept which, similar to the GRAS system in the USA, is aimed to allow strains with established history and safety status to enter the market without extensive testing requirements (EFSA, 2004). The presence of transmissible antibiotic resistance markers in the evaluation of strains is an important safety criterion. Because enterococci are known to be associated with human disease and often carry transferable antibiotic resistances and potential virulence factors, the enterococci may be regarded as having a higher risk potential than, for example, lactobacilli (Bernardeau et al., 2006) and may have to undergo more rigorous testing in order to guarantee that only safe strains enter the food chain. By contrast, enterococcal bacteriocins produced by heterologous hosts or added as cell-free partially purified preparations may still be attractive for application in foods, as exemplified by recent advances on enterocin P production by the yeast Pichia pastoris (Gutiérrez et al., 2005) and on the application of enterocin AS-48 concentrates in several food systems (Ananou et al., 2005a, b, c; Cobo Molinos et al., 2005; Grande et al., 2005, 2006a, b; Lucas et al., 2006).


  • Editor: Ramón Díaz Orejas


View Abstract