INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Bacteriocin production has been described for several genera of lactic acid bacteria (LAB), including Lactobacillus, Carnobacterium, Pediococcus, Lactococcus, Enterococcus, Streptococcus, and Leuconostoc (8, 15, 22, 27, 30, 37). Studies of LAB bacteriocins have attracted increasing interest in recent years because of their potential use as biopreservatives in the food industry to eliminate spoilage and food-borne pathogenic bacteria. Bacteriocin-like peptides are also shown to be involved in gene regulation, and some of these peptides may serve dual functions by both expressing antimicrobial activity and being a signal molecule (peptide pheromone) (23, 31, 36, 42).

Most LAB bacteriocins characterized to date are small (less than 6 kDa), heat-stable, cationic, and hydrophobic and/or hydrophilic peptides and can be classified into two main groups (39): group I, consisting of modified bacteriocins (the lantibiotics), and group II, consisting of the unmodified peptide bacteriocins (the nonlantibiotics). Group II includes the pediocin-like bacteriocins, which contain as a common motif the amino acid sequence Tyr-Gly-Asn-Gly-Val-Tyr at their N terminus (10, 18, 38); the non-pediocin-like bacteriocins; and the two-peptide bacteriocins, which require the complementary action of two peptides for full antimicrobial activity. Bacteriocins are ribosomally synthesized as precursor peptides with an N-terminal leader sequence or a signal peptide (37), which is cleaved off concomitantly with export across the cytoplasmic membrane by dedicated ATP-binding cassette (ABC) transporters and their accessory proteins (20, 24) or by the general secretory pathway (10, 33, 35, 46, 48) (for a comprehensive review, see references 17 and 40). However, we have recently shown that the unmodified non-pediocin-like enterocins L50 (EntL50A and EntL50B) produced by Enterococcus faecium L50 and encoded by a 50-kb plasmid, pCIZ1 (8), unlike other bacteriocins, are synthesized without an N-terminal leader sequence or signal peptide (11). Other unusual characteristics of the EntL50 system are that the structural genes are not cotranscribed with a gene encoding an immunity protein (unpublished results) and that the two peptides constituting the bacteriocin are not synthesized as inactive precursors (11). The two peptides EntL50A and EntL50B act synergistically (11) and display a broad antimicrobial spectrum, which includes food-borne pathogens, such as Listeria monocytogenes, Staphylococcus aureus, Bacillus cereus, Clostridium perfringens, and Clostridium botulinum (8, 9, 12), and human and animal clinical pathogens, such as Streptococcus pneumoniae, Streptococcus mitis, Streptococcus oralis, Streptococcus parasanguis, and Streptococcus agalactiae (unpublished results).

In this work, we present biochemical and genetic evidence demonstrating that E. faecium L50 produces, in addition to EntL50, and under temperature regulation, the sec-dependent pediocin-like enterocin P (EntP), previously identified in E. faecium P13 (6, 10), and a new unmodified non-pediocin-like bacteriocin, termed enterocin Q (EntQ), synthesized without an N-terminal leader sequence or signal peptide. EntL50A and EntL50B are maximally synthesized in the low temperature range (16 to 25°C), whereas EntP and EntQ production is optimal at higher temperatures (37 to 47°C).

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and media. LAB used for the evaluation of antimicrobial activities are listed in Table 1. All strains were aerobically cultured in MRS broth (Difco, Detroit, Mich.) at 30°C. Bacteriocinogenic strain E. faecium L50 was grown in MRS broth at 16, 20, 25, 30, 37, 42, and 47°C.

Antimicrobial activity assays. Cell-free culture supernatants of E. faecium L50 grown at the temperatures cited above were obtained by centrifugation at 12,000 × g for 10 min at 4°C, neutralization with 1 M NaOH, and subsequent filter sterilization through 0.22 µm-pore-size filters (Millipore Corp., Bedford, Mass.). The bacteriocin activity of supernatants and fractions obtained during the purification processes was quantified in a microtiter plate assay (26). Briefly, twofold serial dilutions (50 µl) of bacteriocin extracts in MRS broth were prepared in microtiter plates. The wells were then filled to 200 µl by the addition of 150 µl of a diluted (in MRS) fresh overnight culture of the indicator microorganism. Growth inhibition was measured spectrophotometrically at 620 nm with a microtiter plate reader (Labsystems iEMS Reader MF, Labsystems, Helsinki, Finland) after 12 h of incubation at 30°C. One bacteriocin unit (BU) was defined as the reciprocal of the highest dilution of bacteriocin causing 50% growth inhibition (50% of the turbidity of the control culture without bacteriocin). The antimicrobial activity of in vitro synthesized EntL50A and EntL50B was tested against the indicator microorganisms listed in Table 1 by using the microtiter plate assay described above.

Bacteriocin purification. Purification of bacteriocins was achieved by using a modification of the procedure described by Casaus et al. (7). All the chromatographic equipment was obtained from Pharmacia-LKB (Uppsala, Sweden), and all the purification steps were performed at room temperature if not otherwise stated. The bacteriocins were purified from a 2-liter E. faecium L50 culture which was grown in MRS at 47°C until the early stationary phase (A₆₂₀ = 1.0). The cells were removed by centrifugation at 12,000 × g for 10 min at 4°C, and 40 g of Amberlite XAD-16 (Supelco) was added to the supernatant. The sample was kept at 4°C with stirring for 2 h. The matrix was washed with 100 ml of distilled water and 75 ml of 40% (vol/vol) ethanol in distilled water, and the bacteriocin activity was eluted with 200 ml of 70% (vol/vol) 2-propanol in distilled water (pH 2.0). The eluate was further subjected to cation-exchange (SP Sepharose fast flow) and hydrophobic-interaction (octyl-Sepharose CL-4B) chromatography followed by reverse-phase chromatography in a C₂ to C₁₈ column (PepRPC HR5/5) integrated in a fast-performance liquid chromatography system (6, 9). The bacteriocins were eluted from the reverse-phase column with a linear gradient of 2-propanol (Merck) in aqueous 0.1% (vol/vol) trifluoroacetic acid at a flow rate of 0.5 ml min¹. The bacteriocin activity of fractions obtained during the purification procedure was determined against Lactobacillus sakei NCFB 2714, E. faecium P13, E. faecium T136, and Pediococcus acidilactici 347 by the microtiter plate assay described above. Fractions with high and specific bacteriocin activity were mixed and rechromatographed on the same reverse-phase column to obtain chromatographically pure bacteriocins. Purified bacteriocins were stored in 60% 2-propanol containing 0.1% trifluoroacetic acid at 20°C.

Amino acid sequence analysis. The N-terminal amino acid sequences of purified bacteriocins were determined by Edman degradation with an Applied Biosystems 477A (Foster City, Calif.) automatic sequencer with an online 120A phenylthiohydantoin amino acid analyzer, as described by Cornwell et al. (14).

Mass spectrometry. Determination of the molecular mass of purified bacteriocins was performed at Novo Nordisk A/S (Gentofte, Denmark) with a PE Sciex API1 electrospray mass spectrometer.

PCR analysis and DNA sequencing. The total DNA of E. faecium L50, E. faecium P13 (EntP producer) (6, 10), E. faecium AA13 and E. faecium G16 (EntP producers) (6, 25), and E. faecium E27 and E. faecium LA5 (EntL50 producers) (8) was obtained by the alkaline lysis method of Anderson and McKay (1) and was used as a DNA template for PCR reactions. Oligonucleotide primers used for PCR and DNA sequencing (Table 2) were obtained from KEBOLab (Spanga, Sweden). Different samples of total DNA from E. faecium L50 were digested with EcoRI, BamHI, EcoRI/HindIII, the blunt-end cutter EcoRV, HincII, PvuII, RsaI, ScaI, StuI, or SspI and were ligated (T4 DNA ligase; Promega Corp., Madison, Wis.) to dephosphorylated pBluescript II SK+ (Stratagene, La Jolla, Calif.) digested with EcoRI, BamHI, EcoRI/HindIII or HincII, respectively. Restriction enzymes were obtained from New England Biolabs Inc. (Beverly, Mass.) and were used in accordance with the supplier's instructions. These ligation reactions, which together represents libraries of overlapping fragments from the total DNA of E. faecium L50, were subsequently used as templates to amplify DNA segments near the bacteriocin structural gene. Two specific PCR products consisting of 750 and 1,000 bp were obtained with the SspI and EcoRI/HindIII ligation reactions when the polylinker primer SK2 was used in combination with the degenerated primers ENTQ-1 or ENTQ-2, respectively (Table 2). The sequence of these primers was derived from the amino acid sequence of EntQ obtained by Edman degradation (Fig. 1). The PCR conditions included a hot start at 97°C (2 min), annealing at 55°C (30 s), polymerization at 72°C (2 min), and denaturation at 94°C (1 min). Amplification reactions (35 cycles) were carried out with the Gene Amp PCR kit in a DNA thermal cycler according to the supplier's instructions (Perkin-Elmer Cetus, Norwalk, Conn.). PCR products were analyzed by electrophoresis on 0.8 or 2.0% (wt/vol) agarose gels (1× Tris-acetate-EDTA buffer [pH 8.0]). The gels were run at 80 V for 90 min, using the 67- to 622-bp pBR322 DNA-MspI digest ladder (New England Biolabs Inc.) as a molecular weight standard. PCR fragments to be sequenced were eluted from the agarose gels with a QIAEX II agarose gel extraction kit (Qiagen GmbH, Hilden, Germany) and were further purified by using a QIAquick PCR purification kit (Qiagen). DNA sequencing was performed by using the ABI prism dye terminator cycle sequencing ready reaction kit with AmpliTaq DNA polymerase and dye-labeled terminators (Perkin-Elmer, Applied Biosystems Division, Foster City, Calif.) and an ABI Prism 377 DNA sequencer (Perkin-Elmer). Analysis of the sequence was performed with the ABI Prism sequencing analysis (version 3, Ofc1) and the Autoassembler (version 1.4) software packages (Perkin-Elmer). Based on the sequence information obtained from these DNA fragments, new sequence-specific primers were synthesized (Table 2), and the procedure described above was repeated until the complete sequence of the 1,020-bp fragment of E. faecium L50 containing the EntQ structural gene (entQ) and the putative immunity protein gene (orf2) was determined (Fig. 2).

View larger version (15K): [in this window] [in a new window]   FIG. 1.   Partial amino acid sequence of the N-terminal parts of fraction A (enterocin P) (A) and fraction B (enterocin Q) (B) purified from a 2-liter culture of E. faecium L50 grown at 47°C. Parentheses indicate that the residue was not determined with certainty. Xaa indicates that the residue could not be identified.

View larger version (54K): [in this window] [in a new window]   FIG. 2.   Nucleotide sequence of a 1,020-bp fragment from E. faecium L50 containing the structural gene of enterocin Q (entQ) and the putative immunity protein gene (orf2). The deduced amino acid sequences of EntQ, ORF2, and ORF3 are shown below the DNA sequence. Putative ribosomal binding sites (RBS) are overlined. The putative 10 and 35 promoter sequences are overlined. The horizontal arrows indicate a potential rho-independent transcription terminator sequence.

View larger version (44K): [in this window] [in a new window]   FIG. 3.   Nucleotide sequence of a 963-bp fragment from E. faecium L50 containing the structural gene of enterocin P (entP), the putative immunity gene (entiP), and the putative orf3. The deduced amino acid sequences of EntP, EntiP, and ORF3 are shown below the DNA sequence. Putative ribosome binding sites (RBS) are overlined. The cleavage site of the prebacteriocin is indicated by a vertical arrow. The putative 10 and 35 promoter sequences and ribosome binding sites are underlined; direct repeats (DR1 and DR2) within the conserved regulatory-like boxes are overlined, and inverted repeat sequences (IRR and IRL) are shown in italics.

Computer analysis of DNA and protein sequences. Analyses of DNA and protein sequences were performed using the PC/Gene program package (version 6.8; Intelligenetics, Inc., Mountain View, Calif.) and the ExPASy WWW molecular biology server from the Swiss Institute of Bioinformatics (3).

Generation of E. faecium T136 and E. faecium P13 cells resistant to enterocins L50A and L50B. In order to obtain large enough amounts of pure EntL50A and EntL50B (EntL50AB) to generate E. faecium T136 and E. faecium P13 cultures resistant to these bacteriocins (phenotype, EntL50AB^r), both bacteriocins were synthesized together in vitro with the Escherichia coli T7 S30 extract system for circular DNA (Promega) by using as a DNA template a recombinant plasmid containing entL50A and entL50B (pRSETB-entL50AB), as previously described (11). E. faecium T136 and E. faecium P13 (both strains are sensitive to EntL50AB; phenotype, EntL50AB^s) were inoculated at 1% in MRS broth, and following incubation at 30°C for 18 h, aliquots of cultures containing 5 × 10⁶ CFU were exposed in a microtiter plate to twofold dilution of a 10-µl sample of in vitro synthesized EntL50AB (4 × 10⁵ BU/ml). Plates were analyzed for bacterial growth after incubation at 30°C for 24 and 48 h, and turbid growth cultures from the wells containing the maximum bacteriocin concentration were exposed to in vitro synthesized EntL50AB as described above. This procedure was repeated until cultures of E. faecium T136 and E. faecium P13 did not show sensitivity to in vitro synthesized EntL50AB. EntL50AB^r cultures of E. faecium T136 and E. faecium P13 were subcultivated in MRS broth without bacteriocin and tested for the stability of this phenotype and maintenance of sensitivity to EntP (phenotype, EntP^s) and EntQ (phenotype, EntQ^s), respectively, by a microtiter plate assay.

Bacteriocin production in liquid medium at different growth temperatures. E. faecium L50 was grown in MRS broth at 16, 20, 25, 30, 37, 42, and 47°C until cultures had reached stationary phase. Cell-free culture supernatants were obtained as described above and assayed for EntL50AB, EntP, and EntQ activity by a microtiter plate assay, using as indicators P. acidilactici 347 (EntL50AB^s-EntP^r-EntQ^r) and the resistant cultures obtained from E. faecium T136 (EntL50AB^r-EntP^s-EntQ^r) and E. faecium P13 (EntL50AB^r-EntP^r-EntQ^r).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Purification of the bacteriocins. E. faecium L50 was able to produce antimicrobial activity at temperatures above 37°C, and the activity in the supernatant showed a different behavior from that of the enterocin L50 previously characterized in this strain (8, 9). The antimicrobial activity from the growth media at 37 to 47°C was lost by protease treatment but withstood heat treatments (100°C for 5 min), like most LAB bacteriocins. These observations suggest that the antimicrobial activity found in the growth broth at high temperatures is due to peptide bacteriocins.

Amino acid sequence and mass spectrometry analysis. Purified bacteriocins from the last reverse-phase chromatography step (fractions A and B, Table 3) were subjected to Edman degradation analyses in order to determine their partial amino acid sequences (Fig. 1). All the identified amino acid residues of the first 29 N-terminal steps in the degradation reaction of fraction A (Fig. 1A) were identical to those of enterocin P, including the conserved YGNGVY motif in positions 5 to 10 which are found in the pediocin-like bacteriocins (10, 18, 38). Most of the amino acid residues obtained by the N-terminal amino acid sequencing of fraction B were also unequivocally determined (Fig. 1B). This sequence, which included an N-terminal methionine, lacked the pediocin-like consensus sequence and did not match any amino acid sequences in the protein data banks. This finding suggests that fraction B contains a new bacteriocin, which was termed enterocin Q. The molecular mass of EntQ was determined to be 3,980 Da by mass spectrometry (results not shown).

Genetic organization of the enterocin Q locus. By DNA sequencing a number of PCR products from both DNA strands, the sequence of 1,020 contiguous nucleotides was obtained (Fig. 2). Analysis of the DNA sequence revealed the presence of two consecutive and diverging open reading frames (ORFs). The sequence of EntQ was found in the downstream ORF in this DNA sequence, and it was termed entQ. entQ encodes a 34-amino-acid protein with a theoretical molecular mass of 3,952 Da, corresponding closely to the mass (3,979.8 Da) obtained for the purified EntQ. The DNA sequence allowed identification of the unknown residues at positions 26 and 29 as Cys and Trp, respectively, and the determination of the last five amino acid residues of the C terminus as Glu-Lys-Ile-Ser-Cys. Furthermore, the N-terminal amino acid sequence obtained by Edman degradation gave the N-terminal amino acid sequence MNFLK, which indicates that the deduced gene product of entQ has been secreted without the removal of any N-terminal extension (Fig. 1B and 2). In the DNA sequence, an ATG start codon is preceded 9 bp upstream by a potential ribosome binding site (GAAAGGAGG). A likely 10 consensus promoter region (Pribnow box) (TATATT) is located at position 769, and a sequence (TTAATA) showing resemblance to the ⁷⁰ promoter 35 region is located at the optimal distance of 16 nucleotides upstream of the 10 region. Two inverted repeats of 25 nucleotides (24 identical) separated by 15 nucleotides were identified 4 nucleotides downstream of entQ. These sequences with dyad symmetry can form a stable stem-loop structure with an estimated G of 32.3 kcal mol¹ (135.0 kJ mol¹), which may serve as a rho-independent transcription terminator.

Genetic evidence for enterocin P production in E. faecium L50. The amino acid sequence of the purified bacteriocin in fraction A strongly suggested that enterocin P was produced by E. faecium L50. To confirm this hypothesis and to determine the genetic organization of the bacteriocin operon in this strain, a number of PCR products were obtained and the nucleotide sequence (both DNA strands) of 963 contiguous nucleotides was determined. The 963-bp sequence contained the 755-bp fragment previously obtained for the entP locus in E. faecium P13 (10), which demonstrates that E. faecium L50 also produces the sec-dependent pediocin-like enterocin P. Sequence analysis of the new and additional 208-bp sequence located upstream of entP revealed a putative ORF, orf3, encoding a protein composed of 65 amino acid residues with a calculated molecular mass of 7,978.4 Da. orf3 is colinearly arranged to entP and partially overlaps the direct and inverted repeats found in the promoter region of entP. The ATG start codon found in orf3 is preceded 12 nucleotides upstream by a potential ribosome binding site (AGGG). A likely 10 consensus promoter region (TATAAA) is located at position 139, and a 35 consensus promoter region (TTCACT) is located at the optimal distance of 18 nucleotides upstream of the 10 region. Searches in gene banks with the deduced gene product of orf3 showed no obvious sequence homology to any known protein.

Presence of entL50A and entL50B, entP, and entQ in other bacteriocin-producing enterococci. Several bacteriocinogenic enterococcal strains have previously been isolated from dry fermented sausages in our laboratory (6, 7, 8, 9, 10, 11). It was consequently relevant to determine if other enterococcal isolates from fermented meat contain genes encoding the enterocins described in this work. Both enterocin P- and enterocin L50-producing enterococci were selected for this study. Five strains were tested: E. faecium E27, an EntL50-producing strain containing the 50-kb pCIZ1 and the 7.2-kb pCIZ2 plasmids identical to that of E. faecium L50 (8); E. faecium LA5, which is a novobiocin-treated E. faecium L50 producing EntL50 and lacking the pCIZ2 plasmid (8); and the enterocin P-producing strains E. faecium P13 (6, 10), E. faecium AA13, and E. faecium G16 (6, 25). Total DNA from these strains was prepared, and pairs of primers specific for structural genes of EntL50, EntP, and EntQ were used to detect the appropriate bacteriocin genes by PCR analysis. The presence of EntL50A and EntL50B, EntP, and EntQ structural genes in the enterococcal strains would result in PCR fragments of 556, 132, and 653 bp, respectively, as obtained from DNA of the control strain E. faecium L50 (Fig. 4). DNA from E. faecium E27 generated three amplified bands of these sizes, suggesting that it contains entL50, entP, and entQ. However, PCR analysis with DNA from E. faecium LA5 resulted in the amplification of entL50 and entP but not of entQ, which suggests that this strain retains the ability to encode EntL50 and EntP but may have lost the structural gene of EntQ, presumably because of the loss of the 7.2-kb pCIZ2 (results not shown). The three EntP producer strains (E. faecium P13, E. faecium AA13, and E. faecium G16) showed an amplification band only for the enterocin P structural gene and not for entL50 and entQ.

View larger version (87K): [in this window] [in a new window]   FIG. 4.   Agarose gel electrophoresis of PCR fragments generated from total DNA of E. faecium L50 (1), E. faecium E27 (2), E. faecium LA5 (3), E. faecium P13 (4), E. faecium AA13 (5), and E. faecium G16 (6) with specific oligonucleotide primers for enterocins L50A and L50B (A), enterocin P (B), and enterocin Q (C) structural genes. M, molecular weight marker.

Production of EntL50A and EntL50B, EntP, and EntQ in E. faecium L50 at different temperatures. In order to evaluate the production of EntL50, EntP, and EntQ by E. faecium L50 at various growth temperatures, the strain was grown in MRS broth at 16, 20, 25, 30, 37, 42, and 47°C. Based on the use of indicator strains that are able to differentiate among the three bacteriocins, the activities were determined in cell-free supernatants obtained from growth media at different temperatures, as seen in Fig. 5. EntL50 is maximally accumulated in the medium at the growth temperatures 16, 20, and 25°C (480 BU/ml). At 30°C, a twofold decrease in EntL50 activity was observed, and at 37°C or above, no EntL50 activity was detected. On the other hand, EntP and EntQ activities were detected at all temperatures, and the amount of activity gradually increased with the growth in temperature. EntQ activity reached a maximum (160 BU/ml) at 37°C and remained constant at 42 and 47°C, which represents a two- to fourfold increase over those detected at lower temperatures. Maximum EntP activity was detected at 47°C (320 BU/ml), which represents a 1.3- to 4-fold increase relative to those obtained at 16 to 42°C.

View larger version (12K): [in this window] [in a new window]   FIG. 5.   Production of enterocins L50A and L50B (EntL50) (), EntP (), and EntQ () by E. faecium L50 grown in MRS broth at 16, 20, 25, 30, 37, 42, and 47°C. The antimicrobial activity of cell-free culture supernatants was assayed by a microtiter plate assay against bacteriocin-specific indicator microorganisms.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The biochemical and genetic data presented in this report show that E. faecium L50, a naturally occurring strain isolated from a Spanish dry-fermented sausage (8, 9, 11, 12), produces three different bacteriocins at various amounts depending on the temperature of growth. Several LAB have been reported to produce more than one bacteriocin (4, 7, 13, 16, 29, 37, 41, 43, 47) but it has not been reported that growth temperature may influence the production of various bacteriocins differently. The most studied multiple bacteriocin producer, Lactobacillus plantarum C11, has been reported to contain five consecutive operons within a 16-kb locus responsible for the production of two two-peptide bacteriocins, plantaricins EF and JK, one one-peptide bacteriocin-like peptide, plantaricin N, and the induction factor/bacteriocin, plantaricin A (16). Recently, the six bacteriocin-like peptides have been both partly purified from the cell-free culture supernatant of L. plantarum C11 and chemically synthesized (2). All the peptides except plantaricin N possess some antagonistic activity. Strong synergistic activity between the two peptides constituting both two-peptide bacteriocins (plantaricins EF and JK) was shown to take place (2). E. faecium L50 has now been shown to produce three bacteriocins (four peptides) that have been purified to homogeneity and biochemically and genetically characterized (9-11); also the present work). However, the three bacteriocins produced by E. faecium L50 are encoded by three operons which, unlike the plantaricin system, are located in separate loci and are apparently neither genetically linked nor controlled by a common regulatory system. EntL50 and EntQ are apparently encoded by plasmids pCIZ1 and pCIZ2, respectively. EntP is shown to be frequently produced by several enterococcal strains in the absence of the other two bacteriocins (10, 25, and the present work).

Bacteriocin purification followed by amino acid and DNA sequencing identified a new leaderless bacteriocin termed enterocin Q. EntQ did not have significant amino acid sequence homology to any other characterized bacteriocin but, like EntL50 (11), is synthesized without any leader since no N-terminal extension was found in the translated entQ gene. The absence of a signal leader peptide in these bacteriocins suggests that they can be externalized by some ABC transporters, as has been demonstrated for other bacterial proteins secreted without an N-terminal leader sequence or signal peptide (34, 45). EntQ shares the fundamental physicochemical properties of a typical LAB bacteriocin (size, molecular weight, isoelectric point, and hydrophobicity). However, some characteristics of EntQ clearly distinguish it from EntL50. EntQ is smaller than EntL50A and EntL50B peptides (34 compared with 43 and 44 amino acid residues, respectively) and possesses less pronounced cationic and hydrophobic properties (Table 4). EntQ is apparently a one-peptide bacteriocin since no additional bacteriocin-like encoding gene is found in the vicinity of its structural gene as found with entL50 (11) or other two-peptide bacteriocins (37), and furthermore, only one peptide was found in the purified peak of EntQ activity.

The mass of purified enterocin Q was determined to be 3,980 Da, while the theoretical mass determined from the translated gene was calculated to be 3,952 Da. This discrepancy can be explained by spontaneous modification of the two methionines and the two cysteines found in the peptide. It has been reported that methionines in bacteriocins are often oxidized and cysteines can form cysteine bridges in the peptide (6, 19, 21, 28). By assuming that both methionine residues have become oxidized (an addition of 32 Da to the theoretical mass) and that the two cysteines have formed a disulfide bridge (a reduction of 2 Da), the molecular mass of EntQ can be calculated from the translated gene product to be 3,982 Da, which is within the error of determination of mass of the purified peptide.

One of the most striking features concerning bacteriocin production in E. faecium L50 is the temperature regulation of production. While EntP and EntQ are produced between 16 and 47°C but maximally accumulated at 47 and 37 to 47°C, respectively, EntL50 activity was maximally accumulated at 16 to 25°C and was not detected at 37°C or above. Increased enterocin P and enterocin Q production can be explained partly by increased growth but only up to about 40°C, since above that temperature reduced growth was observed. The reduced EntL50 production at 37°C or above is more difficult to explain. The temperature-dependent production of active bacteriocins is a new feature not observed in other systems. Regulation of bacteriocin production has previously been observed by a quorum-sensing mechanism through a peptide pheromone combined with a two-component regulatory system (36). This regulation has been shown to be affected by environmental changes such as temperature, ionic strength, and media (5, 16). There is no evidence that a quorum-sensing regulation is involved in the production of these enterocins. Heat inactivation is not very likely because bacteriocins are stable to heat, and selective temperature-dependent proteolytic inactivation is not plausible because no inactivation of EntL50 is observed in crude extracts under various temperatures. The production of individual bacteriocins can be temperature dependent, and the production has a temperature optimum. The present work shows that when several bacteriocins are produced by the same bacterium the production of the individual bacteriocins has independent temperature optima, which indicates that some specific regulation is involved in their production.

In most nonlantibiotics described to date, the bacteriocin immunity-encoding gene is located immediately downstream of and in the same operon as the bacteriocin structural gene (30, 37). Both EntL50 and EntQ are exceptions to this rule. The bacteriocin structural gene of both bacteriocins is followed by a putative rho-independent transcription terminator. While no immunity-encoding gene has been identified in the vicinity (3,560-bp pCIZ1 fragment containing the entL50 locus) of the structural bacteriocin genes (entL50A and entL50B), immediately upstream of but in the opposite orientation of entQ a gene encoding a putative immunity protein composed of 221 amino acid residues is found (Fig. 2, orf2). A homology search in protein data banks revealed significant identity (33%) between ORF2 and the 197-amino-acid residue immunity protein of bacteriocin 513 (Baci513) from L. lactis subsp. lactis BGMN1-5 (results not shown). The presence of bacteriocin immunity genes in this atypical orientation has been described so far only for enterocin B produced by E. faecium BFE 900 (21).

Computer-assisted amino-acid-sequence and hydrophobic profile analyses revealed four putative transmembrane domains both in ORF2 and Baci513 and some sequence homology to the transmembrane domain of several ABC transporters which are frequently found also in other bacteriocin immunity systems. Transmembrane helices are located in both immunity factors in similar regions, comprising amino acids 20 to 38, 57 to 79, 127 to 149, and 153 to 175 in ORF2 and amino acids 19 to 37, 57 to 79, 130 to 152, and 155 to 177 in Baci513 (results not shown). The possibility that the immunity protein of EntQ also protects E. faecium L50 against the antimicrobial activity of EntL50 and that both bacteriocin loci are genetically linked can be excluded since E. faecium LA5, a mutant lacking the pCIZ2 encoding EntQ and its immunity protein, retains its ability to produce the former bacteriocins and therefore the immunity (results not shown).

Multiple bacteriocin production by E. faecium L50 may represent an adaptive advantage for this strain since bacteriocins are produced depending on the growth temperature, which can increase its biological fitness and the control of the competing microflora under different environmental conditions. The absence of cross-resistance to EntL50, EntP, and EntQ in the indicator strains, as well as the different target specificity of these bacteriocins, suggests that they may possess different mechanisms of action, and therefore their combined use as biopreservatives may contribute to reduce the frequency at which resistant bacterial populations develop. Elucidation of their mechanisms of action and optimization of bacteriocin production will be essential to evaluate their potential use in the food industry as natural antimicrobial compounds to control spoilage and food-borne pathogenic bacteria.