Complete Sequence of the Enterocin Q-Encoding Plasmid pCIZ2 from the Multiple Bacteriocin Producer Enterococcus faecium L50 and Genetic Characterization of Enterocin Q Production and Immunity

Abstract

The locations of the genetic determinants for enterocin L50 (EntL50A and EntL50B), enterocin Q (EntQ), and enterocin P (EntP) in the multiple bacteriocin producer Enterococcus faecium strain L50 were determined. These bacteriocin genes occur at different locations; entL50AB (encoding EntL50A and EntL50B) are on the 50-kb plasmid pCIZ1, entqA (encoding EntQ) is on the 7.4-kb plasmid pCIZ2, and entP (encoding EntP) is on the chromosome. The complete nucleotide sequence of pCIZ2 was determined to be 7,383 bp long and contains 10 putative open reading frames (ORFs) organized in three distinct regions. The first region contains three ORFs: entqA preceded by two divergently oriented genes, entqB and entqC. EntqB shows high levels of similarity to bacterial ATP-binding cassette (ABC) transporters, while EntqC displays no significant similarity to any known protein. The second region encompasses four ORFs (orf4 to orf7), and ORF4 and ORF5 display high levels of similarity to mobilization proteins from E. faecium and Enterococcus faecalis. In addition, features resembling a transfer origin region (oriT) were found in the promoter area of orf4. The third region contains three ORFs (orf8 to orf10), and ORF8 and ORF9 exhibit similarity to the replication initiator protein RepE from E. faecalis and to RepB proteins, respectively. To clarify the minimum requirement for EntQ synthesis, we subcloned and heterologously expressed a 2,371-bp fragment from pCIZ2 that encompasses only the entqA, entqB, and entqC genes in Lactobacillus sakei, and we demonstrated that this fragment is sufficient for EntQ production. Moreover, we also obtained experimental results indicating that EntqB is involved in ABC transporter-mediated EntQ secretion, while EntqC confers immunity to this bacteriocin.

Bacteriocins are a heterogeneous group of ribosomally synthesized antimicrobial peptides produced by both gram-negative and gram-positive bacteria (14, 33, 52). They often have a narrow inhibitory spectrum; i.e., they are most active against closely related bacteria likely to occur in the same ecological niche as the producer. However, many lactic acid bacteria (LAB) seem to compensate for the narrow inhibitory spectra by producing several bacteriocins (19). In many cases, bacteriocin production is a plasmid-encoded trait; however, bacteriocin genes have been also found on the chromosome and mobile genetic elements, such as transposon-like elements or phage DNA (17).

LAB bacteriocins belonging to class II are nonmodified small heat-stable peptides (19, 39, 40, 41) that are grouped into three subclasses; (i) subclass IIa comprises pediocin-like bacteriocins with a conserved N-terminal motif (YGNGVXC), (ii) subclass IIb comprises bacteriocins whose full activity is dependent on the presence of two different peptides, and (iii) subclass IIc includes other peptide bacteriocins. Class II bacteriocins are commonly synthesized as biologically inactive precursors containing an N-terminal extension (the so-called double-glycine-type leader sequence or the sec-dependent leader peptide), which is cleaved off concomitant with externalization of the active bacteriocin (30, 33, 39). Interestingly, a few bacteriocins described to date are synthesized without an N-terminal extension, including enterocin L50 (11), enterocin Q (13), enterocin EJ97 (47), aureocin A70 (42), aureocin A53 (43), and LsbB (28). Production of most subclass IIa bacteriocins relies on a well-conserved genetic organization including at least the following four genes that are often organized in one or two operon-like structures in gene clusters: (i) the structural gene encoding the prebacteriocin; (ii) a gene encoding the immunity protein, which confers producer self-protection against the toxicity of the bacteriocin; (iii) a gene encoding a dedicated ATP-binding cassette (ABC) transporter required for processing and transport of the bacteriocin; and (iv) a gene encoding an accessory protein required for proper bacteriocin externalization (30, 39, 52).

Enterococcus faecium L50, a strain isolated from a Spanish dry-fermented sausage (8, 9, 11), harbors two plasmids, pCIZ1 (ca. 50 kb) and pCIZ2 (ca. 7.4 kb), and produces three different bacteriocins; (i) enterocin P (EntP) is a pediocin-like bacteriocin (subclass IIa), (ii) enterocin L50 (EntL50) is a subclass IIb bacteriocin whose full activity is dependent on the peptides EntL50A and EntL50B, and (iii) enterocin Q (EntQ) is a subclass IIc bacteriocin. While EntL50 and EntQ are synthesized as leaderless bacteriocins and are secreted by a hitherto unknown mechanism, EntP is synthesized with a sec-dependent leader peptide (11, 13) and is secreted by the Sec translocase (31). Moreover, the production of multiple bacteriocins by E. faecium L50 is a temperature-regulated process; EntL50 is produced at 16 to 32°C but production is negligible when the growth temperature is above 37°C, whereas EntP and EntQ are synthesized at temperatures ranging from 16 to 47°C (13; R. Criado, J. Gutiérrez, M. Martín, C. Herranz, P. E. Hernández, and L. M. Cintas, submitted for publication). The production of several bacteriocins confers to E. faecium L50 a broad spectrum of antimicrobial activity, including activity against food-borne pathogens, such as Listeria monocytogenes, Clostridium perfringens, and Clostridium botulinum, and activity against human and animal clinical pathogens, such as Streptococcus pneumoniae, Streptococcus mitis, Streptococcus oralis, Streptococcus parasanguis, and Streptococcus agalactiae (11, 13).

In this work, the genetic locations of all three bacteriocins in this multiple bacteriocin producer were established by plasmid curing and Southern analysis, and both plasmid pCIZ1 and plasmid pCIZ2 were found to be bacteriocinogenic. The complete nucleotide sequence of plasmid pCIZ2, which is responsible for EntQ production and immunity, was determined. Furthermore, the minimum genetic requirements for EntQ production and immunity were established by heterologous expression experiments.

MATERIALS AND METHODS

Bacterial strains, plasmids, media, and culture conditions.

Bacterial strains and plasmids used in this work are listed in Table 1. All LAB strains were grown in MRS broth (Oxoid Ltd., Basingstoke, United Kingdom) at 32°C, unless otherwise stated. Escherichia coli cells were grown in LB broth (Sigma-Aldrich Inc., St. Louis, Mo.) at 37°C with shaking. The following antibiotics (Sigma) were added to the media as selective agents when required: ampicillin (100 μg/ml for E. coli) and erythromycin (300 μg/ml for E. coli, 20 μg/ml for E. faecium, and 5 μg/ml for Lactococcus lactis and Lactobacillus sakei). Pure enterocin Q was chemically synthesized at the Molecular Biology Unit, University of Newcastle Upon Tyne, Newcastle Upon Tyne, United Kingdom.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant characteristics	Source or reference
Bacterial strains
Escherichia coli TOP10	Host strain, F−mcrA Δ(mrr-hsdRMS-mcrBC) φ8lacZΔM15 ΔlacW74 deoR recA1 araD139 Δ(ara-leu)7697 galU galK rpsL endA1 nupG	Invitrogen
Enterococcus faecium strains
L50	Wild-type strain (pCIZ1+, pCIZ2+); EntL50, EntP, and EntQ producer; Emr	12
LA5	E. faecium L50 cured derivative (pCIZ1+, pCIZ2−); EntL50 and EntP producer; EntQs	13
L50/30-2	E. faecium L50 cured derivative (pCIZ1+, pCIZ2−); EntL50 and EntP producer; EntQs Emr	This study
L50/14-2	E. faecium L50 cured derivative (pCIZ1−, pCIZ2−); EntP producer; EntL50s EntQs Ems	This study
L50/14-2(pMG36e)	E. faecium L50/14-2 derivative carrying pMG36e; EntP producer; EntL50s EntQs Emr	This study
L50/14-2(pRCG04-1)	E. faecium L50/14-2 derivative carrying pRCG04-1; EntP producer; EntL50s EntQr Emr	This study
L50/14-2(pRCG04-2)	E. faecium L50/14-2 derivative carrying pRCG04-2; EntP producer; EntL50s EntQr Emr	This study
P13	EntP producer; indicator microorganism; EntL50s EntPr EntQs	10
T136	EntA and EntB producer; indicator microorganism; EntL50s EntPs EntQr	7
Enterococcus faecalis JH2-2	Plasmid-free recipient strain, non-bacteriocin producer (Bac−); Rifr Fusr Erys	34
Pediococcus acidilactici 347	Pediocin PA-1 producer; indicator microorganism; EntL50s EntPr EntQr	12
Lactobacillus sakei strains
Lb790	Plasmid-free host strain, non-bacteriocin producer (Bac−); EntQr Ems	49
Lb790(pELS200)	L. sakei Lb790 derivative carrying pELS200	This study
Lb790(pRCG01)	L. sakei Lb790 derivative carrying pRCG01; entqA	This study
Lb790(pRCG02)	L. sakei Lb790 derivative carrying pRCG02; entqB entqC	This study
Lb790(pRCG03)	L. sakei Lb790 derivative carrying pRCG03; entqA entqB entqC; EntQ producer; Emr	This study
Lactococcus lactis DPC5598	Plasmid-free host strain; Ems	55
Plasmids
pCIZ1	50-kb plasmid from E. faecium L50; EntL50+ Imm+ Emr	13
pCIZ2	7.4-kb plasmid from E. faecium L50; EntQ+ Imm+	13
pGEM-3Zf(+)	3-kb cloning vector; Ampr	Promega
pELS200	7.8-kb E. coli-Lactobacillus shuttle vector; Ampr Emr	54
pMG36e	3.6-kb gene expression vector carrying P32 promoter; Emr	56
pGEM-CIZ2	pGEM-3Zf(+) derivative containing a 7.4-kb HindIII pCIZ2 fragment	This study
pRCG01	pELS200 derivative carrying entqA	This study
pRCG02	pELS200 derivative carrying entqB-entqC	This study
pRCG03	pELS200 derivative carrying enqA-entqB-entqC	This study
pRCG04-1	pMG36e derivative carrying truncated entqB (due to nucleotide substitution at coordinate 1519) and entqC	This study
pRCG04-2	pMG36e derivative carrying truncated entqB (due to nucleotide deletion at coordinate 1961) and entqC	This study

Bacteriocin and immunity assays.

The antimicrobial activity of individual colonies was screened by a stab-on-agar test as previously described by Cintas (8), using as the indicator microorganisms Pediococcus acidilactici 347 (EntL50 sensitive [EntL50^s], EntP resistant [EntP^r], and EntQ resistant [EntQ^r]) (12), E. faecium T136 (EntL50^s EntP^s EntQ^r) (7), and E. faecium P13 (EntL50^s EntP^r EntQ^s) (10) (Table 1). The antimicrobial activities of cell-free culture supernatants and chemically synthesized EntQ were detected by an agar diffusion test (ADT) and/or a microtiter plate assay (MPA) as previously described (13). The immunity tests were performed by using ADT and MPA in which the target strains were challenged with either chemically synthesized EntQ, supernatants from a culture of E. faecium P13 grown at 32°C containing EntP (10), or supernatants from a culture of E. faecium mutant strain LA5 grown at 32°C containing EntP and EntL50 (13) (Table 1). The MICs of chemically synthesized EntQ were determined by an MPA.

DNA isolation and manipulations.

DNA for Southern hybridization was isolated from E. faecium L50 by the alkaline lysis method, modified as previously described (8). Small-scale plasmid DNA isolation from E. coli was carried out using a QIAgen Spin Miniprep kit (QIAGEN GmbH, Hilden, Germany). Large-scale plasmid DNA isolation from E. coli, E. faecium, and L. lactis was carried out using a QIAgen Plasmid Midi kit (QIAGEN) after an initial treatment with lysozyme (final concentration, 5 mg/ml) at 37°C for 15 min when required. Restriction enzymes (New England Biolabs Inc., Beverly, Mass.), calf alkaline phosphatase (Promega Corporation, Madison, Wis.), T4 DNA ligase (Promega), and Platinum Taq DNA polymerase (Invitrogen Life Technologies, Carlsbad, Calif.) were used according to the suppliers' instructions. PCR products were purified by using a QIAquick PCR purification kit (QIAGEN). Most DNA manipulations were carried out using standard procedures described elsewhere (46).

Plasmid curing.

Bacteriocin-deficient variants of E. faecium wild-type strain L50 were obtained by plasmid curing using novobiocin. E. faecium L50 was inoculated into MRS broth (about 1 × 10⁵ CFU/ml) with increasing concentrations (0 to 10 μg/ml) of novobiocin (Sigma) and incubated at 32°C for 72 h. The culture that grew at the highest novobiocin concentration was serially diluted and plated onto MRS agar (1.5%, wt/vol) plates to obtain individual colonies. Randomly selected colonies were replica plated onto MRS agar plates and screened for bacteriocin production by the stab-on-agar test and ADT. The presence of bacteriocin structural genes for EntL50 (entL50A and entL50B), EntP (entP), and EntQ (entqA) was investigated by PCR using the specific EntL50S1/EntL50S2, EntPS1/EntPS2, and EntQS1/EntQS2 primer pairs (Table 2) designed on the basis of the previously published EntL50, EntQ, and EntP operon sequences (11, 13). Plasmid profiles of the wild-type strain and the derived mutants were determined and analyzed by agarose (0.8%, wt/vol) gel electrophoresis.

TABLE 2.

PCR primer pairs used in this study

Oligonucleotides		PCR product size (bp)
Designation	Sequence (5′-3′)a
EntL50S1	TCCCTACAGTCTCCCTTCC	877
EntL50S2	TCTAGCGTTAAGCCGAATG
EntPS1	TTATGCGCGTTATTGTGTG	682
EntPS2	CATAACTCAAAGTCCCGACC
EntQS1	AACAAGAAAATTGCGGCTG	678
EntQS2	AGGGCTACTTGGATAGTACAC
EntQC1XbaI	CGCGTAACTCTAGAGAAATATTGCATAGCAACCTAGTG	2,207
EntQC2XhoI	GTACTACCTCGAGATCCATTTTGCGATACCATT
EntQC1KpnI	GGTGGTGGTACCGAAATATTGCATAGCAACCTAG	2,207
EntQC2SacI	GGTGGTGAGCTCATCCATTTTGCGATACCATT
EntQC1XbaI	CGCGTAACTCTAGAGAAATATTGCATAGCAACCTAGTG	2,371
EntQC4XhoI	TAATTAATCTCGAGCTAATCGTATAAATAACGAAGGAAATA
EntQC3XbaI	CGCGTAACTCTAGATACTGATATGAGACTAAAAAGATTGG	381
EntQC4XhoI	TAATTAATCTCGAGCTAATCGTATAAATAACGAAGGAAATA

^aRestriction sites are underlined. Boldface type indicates nucleotides in the nucleotide tail added to the specific sequences to ensure proper functioning of the restriction enzymes.

Southern hybridization.

DNA from E. faecium L50 was isolated, separated by agarose (0.8%, wt/vol) gel electrophoresis, and subsequently transferred onto a positively charged nylon membrane (Hybond-N+; Amersham Biosciences GmbH, Freiburg, Germany) by vacuum blotting (VacuGene XL; Pharmacia LKB, Uppsala, Sweden) as described by Sambrook and Russell (46). Specific DNA probes for the EntL50 operon (877 bp), EntQ operon (678 bp), and EntP operon (682 bp) were prepared by PCR amplification from E. faecium L50 DNA using the specific EntL50S1/EntL50S2, EntQS1/EntQS2, and EntPS1/EntPS2 primer pairs (Table 2) and were labeled with horseradish peroxidase (ECL Direct nucleic acid labeling kit; Amersham Biosciences) according to the manufacturer's instructions. Prehybridization (42°C, 15 min), hybridization (42°C, overnight), and posthybridization washes (0.5× SSC-0.4% sodium dodecyl sulfate [SDS] at 42°C for 10 min twice; 2× SSC at 25°C for 5 min twice [1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate]) were carried out with a Minioven MKII (Hybaid, Middlesex, United Kingdom). Hybridization detection was performed as suggested in the ECL Direct nucleic acid detection system user's guide for Southern hybridization (Amersham Biosciences). The chemiluminescence signals were detected following exposure of blue-light-sensitive autoradiography films (Hyperfilm-ECL; Amersham Biosciences).

DNA sequencing of pCIZ2 and sequence analysis.

Plasmid pCIZ2 was linearized with HindIII, and the purified 7.4-kb fragment obtained was ligated into the HindIII site of the pGEM7Z vector, resulting in pGEM-CIZ2, which was used as the starting point for sequencing by primer walking. The DNA on both strands was sequenced by the dideoxy chain termination method (48), using an ABI Prism BigDye cycle sequencing Ready Reaction kit with AmpliTaq DNA polymerase and dye-labeled terminators and an ABI Prism 377 DNA sequencer (Perkin-Elmer, Applied Biosystems Division, Foster City, Calif.). Sequences were assembled using the Autoassembler software package (Perkin-Elmer) and the GCG software package of the Genetics Computer Group (University of Wisconsin, Madison, Wis.). To predict the location of putative open reading frames (ORFs) in pCIZ2, the ORF-finder software (National Center for Biotechnology Information [NCBI], U.S. National Library of Medicine) (http://www.ncbi.nlm.nih.gov/gorf/gorf.html) was used. The BLAST software (NCBI) (http://www.ncbi.nlm.nih.gov) was used to conduct similarity searches with GenBank and EMBL sequence databases. Multiple-sequence alignments of related amino acid sequences were constructed using the ClustalW software (Baylor College of Medicine Human Genome Sequencing Center) (http://searchlauncher.bcm.tmc.edu) and the Boxshade software (Swiss Node of EMBnet, Swiss Institute for Experimental Cancer Research) (http://www.ch.embnet.org). Transmembrane helices in putative proteins were predicted by using the TMHMM software (Center for Biological Sequence Analysis, Technical University of Denmark) (http://www.cbs.dtu.dk). The presence of protein motifs was examined by using the CDD software (NCBI) (http://www.ncbi.nlm.nih.gov) and InterProScan software (Swiss Institute of Bioinformatics) (http://www.expasy.org). The presence of nucleotide and amino acid conserved sequences was examined by using the Fuzznuc software and the Fuzztran software, respectively (The German National Genome Research Network) (http://ngfnblast.gbf.de). The presence or absence of coiled-coil regions in putative proteins was predicted by using the PepCoil software (The German National Genome Research Network) (http://ngfnblast.gbf.de).

Plasmid construction, DNA transformation, and mating experiments.

To investigate the biological functions of entqB and entqC, several expression clones were constructed in E. coli TOP10 or L. lactis DPC5598 (55) and subsequently transferred to L. sakei Lb790 (49) or E. faecium L50/14-2 (Table 1). Inserts were obtained by PCR amplification of entqA, entqBC, and entqABC using E. faecium L50 plasmid DNA as the template and the specific primers carrying terminal restriction sites listed in Table 2. Recombinant plasmids (Table 1) were constructed by cloning entqA and/or entqBC into the cloning vectors pMG36e (56) and pELS200 (54). The entqA gene was PCR amplified from pCIZ2 with primers EntQC3XbaI and EntQC4XhoI as a 381-bp fragment, digested with XbaI and XhoI, and ligated into the corresponding sites in dephosphorylated pELS200, resulting in pRCG01. DNA fragments containing entqB and entqC were amplified from pCIZ2 with the EntQC1XbaI/EntQC2XhoI and EntQC1KpnI/EntQC2SacI primer pairs and digested with the appropriate restriction enzymes prior to ligation into the corresponding sites in dephosphorylated pELS200 and pMG36e, resulting in pRCG02 and in pRCG04-1 and pRCG04-2, respectively. The entqBC insert in pRCG04-1 and pRCG04-2 contains a single mutation (substitution of C by T at coordinate 1,519 in pRCG04-1 and deletion of T at coordinate 1,961 in pRCG04-2); the result of this is that both plasmids encode intact EntqC but truncated variants of EntqB. These mutations were spontaneously introduced during PCRs. A 2,371-bp fragment encompassing entqABC was amplified from pCIZ2 using primers EntQC1XbaI and EntQC4XhoI, digested with the appropriate restriction enzymes, and ligated into the corresponding sites in dephosphorylated pELS200, resulting in pRCG03. The integrity of all PCR-generated fragments was confirmed by DNA sequencing.

E. coli TOP10 was used as an intermediate host for amplification of plasmids pRCG01, pRCG02 and pRCG03, whereas L. lactis DPC5598 was used for amplification of pRCG04-1 and pRCG04-2. Recombinant plasmids were reisolated and then transferred to L. sakei Lb790 or E. faecium L50/14-2 by electroporation as previously described (2, 32).

The transferability of bacteriocin production, as well as erythromycin resistance (Ery^r) determinants, was tested by conjugation using a filter method (18), with plasmid-free Enterococcus faecalis JH2-2 as the recipient strain (rifampin resistant [Rif^r], fusidic acid resistant [Fus^r], erythromycin susceptible [Ery^s], and nonbacteriocin producer [Bac⁻]) and E. faecium L50 as the donor strain (Rif^s Fus^s Ery^r Bac⁺).

NCI-ELISA.

The presence of EntQ in cell-free supernatants and intracellular extracts (5) obtained from growing (log-phase) or overnight cultures of L. sakei Lb790 and the recombinant strains L. sakei Lb790(pRCG01), L. sakei Lb790(pRCG02), and L. sakei Lb790(pRCG03) was determined and quantified by a noncompetitive indirect enzyme-linked immunosorbent assay (NCI-ELISA) essentially as previously described (29), using rabbit polyclonal antibodies with specificity for EntQ (anti-EntQ-KLH) (Criado et al., submitted for publication). Briefly, wells of flat-bottom polysterene microtiter plates (Maxisorp; Nunc, Roskilde, Denmark) were coated overnight (4°C) with supernatants or intracellular extracts. After coating, wells were blocked and then washed. Next, diluted anti-EntQ-KLH serum was added to each well, unbound antibodies were removed by washing, and goat anti-rabbit immunoglobulin G-peroxidase conjugate (Cappel Laboratories, West Chester, PA) was added. The amount of bound peroxidase was determined using ABTS [2,2′-azinobis(3-ethylbenzthiazoline-6-sulfonic acid)]) (Sigma) as the substrate by measuring the absorbance at 405 nm of the wells with an iEMS reader (Labsystems, Helsinki, Finland). The plates included control wells coated with (i) coating buffer (0.1 M sodium carbonate-bicarbonate buffer, pH 9.6)-MRS broth to determine the background level of the plate and (ii) six twofold dilutions of samples containing known concentrations of pure EntQ (determined from the A₂₈₀ using the molar extinction coefficient) in coating buffer-MRS broth to determine a standard curve.

Bacteriocin purification and mass spectrometry.

The EntQ heterologously produced by L. sakei Lb790(pRCG03) was purified from a 1-liter culture grown in MRS broth at 32°C for 14 h, as previously described by Cintas et al. (13). The fraction corresponding to the peak containing the purified bacteriocin was subjected to mass spectrum analysis by using a matrix-assisted laser desorption ionization—time of flight Voyager-DE STR mass spectrometer (PerSeptive Biosystems, Foster City, Calif.).

Tricine-SDS-PAGE analysis, Western blotting, and overlay assay.

The purity of EntQ at the final reversed-phase fast protein liquid chromatography purification step was also analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis on 16% Tricine gels (Tricine-SDS-PAGE) (51) after silver staining. Western blotting using anti-EntQ-KLH antibodies was performed essentially as previously described (29). To determine the antimicrobial activity of the purified heterologously produced EntQ, an overlay assay (6) was performed using E. faecium P13 (about 1 × 10⁵ CFU/ml) as the indicator microorganism.

Nucleotide sequence accession number.

The nucleotide sequence reported in this paper has been deposited in the GenBank database under accession no. DQ832184.

RESULTS

Localization of the bacteriocin structural genes for EntL50, EntQ, and EntP.

E. faecium L50 harbors two plasmids, pCIZ1 (about 50 kb) and pCIZ2 (about 7.4 kb) (11, 13). To examine whether the genetic determinants responsible for bacteriocin production are plasmid encoded in E. faecium L50, plasmid curing was performed using novobiocin (3 μg/ml) as the curing agent. Two plasmid profiles were obtained from the plasmid-cured mutants; the type I profile had only the larger plasmid, pCIZ1, while the type II profile had no plasmid, as shown for one isolate (designated 30-2) with the type I profile and one isolate (designated 14-2) with the type II profile in Fig. 1B and Table 3. Isolate 30-2 still produces EntL50 and EntP but is not able to produce EntQ (Fig. 1C), indicating that one or more of the genetic determinants required for EntQ production are located on pCIZ2. Likewise, only EntP activity and not EntL50 or EntQ activity were detected in isolate 14-2, indicating that EntP determinants are chromosomally located. Together, the results also imply that the genetic determinants for EntL50 are located on pCIZ1. These results were confirmed by Southern analysis using DNA probes specific for the different bacteriocin genes. Thus, we found that the entP-specific probe hybridized to the chromosome, while the probes specific to entL50AB and entqA hybridized to pCIZ1 and pCIZ2, respectively (Fig. 1A).

FIG. 1.

(A) Southern hybridization analysis of total DNA from E. faecium L50: agarose gel stained with ethidium bromide showing plasmid profile of E. faecium L50 (W) and autoradiograms of membrane-bound DNA that hybridized with labeled specific-PCR probes for entL50AB (L50), entP (P), and entqA (Q). (B) Plasmid profiles of wild-type strain E. faecium L50 (W) and the derived mutants E. faecium L50/30-2 (30-2) and E. faecium L50/14-2 (14-2). Chr., chromosomal DNA; pCIZ1, 50-kb plasmid; pCIZ2, 7.4-kb plasmid; M, Supercoiled DNA ladder (Gibco-BRL). (C) Bacteriocin activity in supernatants from cultures of wild-type strain E. faecium L50 (W) and the derived mutants E. faecium L50/14-2 (14-2) and E. faecium L50/30-2 (30-2) grown in MRS broth at 32 and 42°C as determined by an ADT using P. acidilactici 347 (EntL50^s EntP^r EntQ^r), E. faecium P13 (EntL50^s EntP^r EntQ^s), and E. faecium T136 (EntL50^s EntP^s EntQ^r) as indicator microorganisms. The bacteriocins responsible for the antimicrobial activity are indicated below the inhibition halos. (D) Bacteriocin immunity and sensitivity of wild-type strain E. faecium L50 (W) and the derived mutants E. faecium L50/14-2 (14-2) and E. faecium L50/30-2 (30-2) to supernatants from cultures of the mutant E. faecium LA5 containing EntL50 and EntP (L50+P) and E. faecium P13 containing EntP (P), both grown in MRS broth at 32°C, and to chemically synthesized EntQ (0.5 mg/ml) (Q).

TABLE 3.

Presence of EntL50 (entL50A and entL50B), EntP (entP), and EntQ (entqA) structural genes, phenotypes, and plasmid profiles for wild-type strain E. faecium L50 and the derived mutants E. faecium L50/30-2 and E. faecium L50/14-2

Strain	Structural genes			Phenotypes	Plasmid profile
	entL50ABa	entP	entqA
E. faecium L50	+	+	+	EntL50+ Imm+/EntP+ Imm+/EntQ+ Imm+	pCIZ1+/pCIZ2+
E. faecium L50/30-2	+	+	−	EntL50+ Imm+/EntP+ Imm+/EntQ− Imm−	pCIZ1+/pCIZ2−
E. faecium L50/14-2	−	+	−	EntL50− Imm−/EntP+ Imm+/EntQ− Imm−	pCIZ1−/pCIZ2−

^aentL50AB corresponds to a PCR fragment containing entL50A and entL50B amplified using primers EntL50S1 and EntL50S2.

To examine whether the genes conferring immunity are genetically linked with their cognate bacteriocin structural genes, as commonly found in other bacteriocin systems (17, 40), we examined their immunity potentials by exposing the isolates to various bacteriocins on agar plates. As expected, isolate 30-2, which did not contain the EntQ-encoding plasmid (pCIZ2), was sensitive only to EntQ and not to EntP or EntL50, whereas isolate 14-2, which did not contain either plasmid, was sensitive to both EntL50 and EntQ but not to EntP, which is chromosomally encoded (Fig. 1D).

DNA sequence of pCIZ2.

The complete nucleotide sequence of pCIZ2 (Fig. 2) was determined by primer walking. It is 7,383 bp long, and its G+C content is 32.3%, which is lower than the value reported for the E. faecium genomic DNA (38.3 to 39%) (38). Computer analysis of the pCIZ2 sequence revealed the presence of 10 putative ORFs (orf1 to orf10) (Table 4), each preceded by a putative ribosome binding site (Fig. 2). All of these ORFs except orf1 have the same direction on the plasmid (Fig. 2). The ORFs on pCIZ2 seem to be organized into three distinct functional regions (Fig. 3).

FIG.2.

Nucleotide sequence of the 7,383-bp plasmid pCIZ2. The deduced amino acid sequences are shown below the DNA sequence. Putative ribosome binding sites (RBS) are indicated by boldface type. The putative −35 and −10 promoter sequences are underlined, and start codons are indicated by boldface italics. Stop codons are indicated by asterisks at the ends of protein sequences. The horizontal arrows under the nucleotide sequence indicate inverted repeats. Direct repeats are enclosed in boxes and overlined. Primers used in this study are indicated by dashed underlining. Transmembrane segments are underlined in the amino acid sequence. Nucleotide and amino acid motifs and conserved sequences are enclosed in boxes. The vertical solid and open arrows indicate the nucleotides substituted and deleted in the truncated entqB encoded by pRCG04-1 and pRCG04-2, respectively. The 1,020-bp sequence from coordinates 40 to 1060 was previously published by Cintas et al. (13).

TABLE 4.

ORFs in 7,383-bp plasmid pCIZ2 from E. faecium L50 and the closest relatives of the deduced proteins

ORF	Position in nucleotide sequence		Gene		Protein
	5′	3′	Designation	G+C content (%)	Length (amino acids)	Molecular mass (kDa)	pI	Closest relative (length; e value, level of amino acid identity/level of amino acid similarity; microorganism)	Accession no.
ORF1	259	155	entqA	33.33	34	3.95	9.39	Structural gene of enterocin Q	A4318955
ORF2	340	2058	entqB	25.48	572	64.66	9.51	Ej97B, ABC transporter (583 amino acids; e−123, 41%/65%; E. faecalis)	CAD35294
								LmrB, MDR transporter (567 amino acids; e−119, 41%/63%; L. lactis)	NP_861550
ORF3	2069	2272	entqC	20.59	67	7.69	9.63	No significant similarity found
ORF4	2686	3012	orf4	44.95	108	12.11	9.89	Mobilization protein (129 amino acids; 5e−46, 78%/78%; E. faecium)	EAN08864
								MobC, mobilization protein (127 amino acids; 3e−16, 61%/77%; E. faecalis)	NP_863267
ORF5	2994	3908	orf5	39.23	304	35.77	9.36	Relaxase/mobilization nuclease domain (304 amino acids; e−167, 94%/97%; E. faecium)	EAN08860
								MobA, mobilization protein (346 amino acids; 3e−64, 44%/58%; E. faecalis)	NP_863268
ORF6	3700	4146	orf6	40.94	148	17.70	5.98	Hypothetical protein (538 amino acids; 1e−4, 34%/55%; Mycoplasma synoviae)	YP_278611
								ATP synthase F0, subunit b (171 amino acids; 0.001, 26%/49%; Helicobacter pylori)	NP_223781
ORF7	4148	4837	orf7	33.33	229	27.09	8.74	Hypothetical protein (229 amino acids; 2e−120, 94%/98%; E. faecium)	EAN08861
ORF8	5690	6427	orf8	26.00	245	29.17	9.46	RepE, replication protein (240 amino acids; 2e−44, 40%/62%; E. faecalis)	NP_863355
ORF9	6455	6991	orf9	33.89	178	21.10	5.40	Protein with unknown function (179 amino acids; 1e−91, 96%/98%; E. faecium)	EAN08855
								RepB, replication protein (168 amino acids; 1e−11, 28%/52%; L. casei)	NP_542222
ORF10	7007	7177	orf10	30.41	56	6.33	9.16	No significant similarity found

FIG. 3.

Physical map showing the genetic organization of the 7,383-bp plasmid pCIZ2. ORFs are represented by arrows indicating the direction of transcription. Inverted repeats are indicated by stem-loop symbols. Direct repeats are indicated by boxes.

Region I is comprised of three ORFs (orf1 to orf3). orf1 corresponds to the enterocin Q structural gene entqA (previously reported as entQ) (13), which is preceded by two putative promoters (Fig. 2). Downstream of entqA are two inverted repeats (IR1L and IR1R) that could serve as a rho-independent transcriptional terminator. Together, these findings suggest that the structural bacteriocin gene can be transcribed as a monocistronic unit (Fig. 2).

orf2 (designated entqB), which is divergently located 79 bp upstream of entqA, is preceded by a putative consensus promoter (Fig. 2). entqB encodes a 572-amino-acid protein with high levels of sequence homology to many prokaryotic ABC transporters, including LmrB and AurT, which are involved in export of the leaderless bacteriocins LsbB and aureocin A70, respectively (28, 42) (Table 4). Sequence analyses of EntqB revealed the presence of the following features typical of ABC transporters: (i) the Walker A and B motifs; (ii) a highly conserved C motif (linker peptide) preceding the Walker B motif; and (iii) the so-called “switch region,” located after the C motif (15, 21, 30, 50) (Fig. 4A). Moreover, analysis of the primary sequence of this polypeptide revealed extensive hydrophobic stretches (residues 18 to 40, 53 to 75, 124 to 146, 153 to 175, 244 to 266, and 273 to 295) that are likely to form six putative transmembrane segments.

FIG. 4.

(A) Alignment of the predicted amino acid sequence of EntqB encoded by pCIZ2 (EntqB-pCIZ2) with the sequences of several bacterial ABC transporters, including those from Bacillus halodurans (ABC-Bh) (accession no. NP_241579), Lactobacillus johnsonii (ABC-Lj) (accession no. NP_965013), Oceanobacillus iheyensis (ABC-Oi) (accession no. NP_691347), S. aureus (AurT [accession no. AAK73550] and ABC-A-Sa [accession no. CAA62898]), Staphylococcus epidermidis (PepT [accession no. CAA90021] and ABC-Se [accession no. YP_187899]), and E. faecalis (Ej97B) (accession no. CAD35294), and with the sequences of ABC-type multidrug transporters (MDR) from Bacillus thuringiensis (MDR-Bt) (accession no. YP_036121), L. lactis (LmrB [accession no. AAC14278] and MDR-Ll [accession no. NP_266867]), and Lactobacillus brevis (HorA) (accession no. BAD80897). (B) Alignment of the predicted amino acid sequence of ORF5 encoded by pCIZ2 with the predicted amino acid sequences of several plasmid relaxases and mobilization proteins, including those from E. faecium (RLX-Ef [pRUM; accession no. NP_863170] and VirD2-Ef [accession no. ZP_00287779]), E. faecalis (MobA-Efs [pEF1071; accession no. AAN76342], RLX-Efs [pAD1; accession no. AAF72355], and orf-Efs [accession no. NP_815959]), Pediococcus pentosaceus (Mob-Pp [pMD136; accession no. AAD39619]), S. aureus (RLX-Sa [pC221; accession no. CAA26106]), and S. epidermidis (RLX-Se [accession no. NP_863255]). Identical residues and conservative substitutions are indicated by black and gray backgrounds, respectively. The conserved motifs are indicated above the sequences. Asterisks and dots in the consensus sequence indicate highly conserved and 100% conserved residues, respectively. Solid and open arrows indicate the last amino acids of the truncated EntqB proteins encoded by pRCG04-1 and pRCG04-2, respectively.

orf3 (designated entqC), located immediately downstream of entqB (Fig. 2), encodes a putative 67-amino-acid peptide containing two possible transmembrane segments. However, EntqC shows no significant sequence similarity to any known protein in data banks.

Region II is comprised of four ORFs (orf4 to orf7). orf4 is preceded by a putative promoter (Fig. 2) and encodes a putative 108-amino-acid protein which most closely resembles a mobilization protein from E. faecium and the mobilization protein MobC from E. faecalis (3, 4) (Table 4). Features typical of a transfer origin (oriT) (24, 35, 53, 58) were found in the region containing the putative promoter of orf4; these features include (Fig. 2) (i) two inverted repeats (IR3 and IR4) and (ii) a sequence (GAGCTTGC) with homology to the core sequence of the cis-acting nic site of the IncP/Mob_P family of gram-positive plasmids (RYGCTTGC) (53).

orf5, which overlaps the end of orf4, encodes a 304-amino-acid protein with high levels of sequence similarity to a relaxase/mobilization nuclease domain from E. faecium and to the mobilization protein MobA from E. faecalis (4) (Table 4). Typical features for this protein family include the presence of three highly conserved motifs (motifs I, II, and III) at the N-terminal relaxase domain (24, 44) (Fig. 4B). The presence of two mob genes and a possible oriT sequence suggested that pCIZ2 could be mobilizable. However, numerous mating experiment attempts, carried out with the plasmidless recipient strain E. faecalis JH2-2 (Rif^r Fus^r Ery^s) (34) and the donor strain E. faecium L50 (Rif^s Fus^s Ery^r), were unsuccessful.

orf6, which overlaps the end of orf5, encodes a 148-amino-acid protein which shows similarity to a hypothetical protein from Mycoplasma synoviae and to subunit b of ATP synthase F0 from Helicobacter pylori (1) (Table 4).

orf7, which is immediately downstream of orf6 (Fig. 2), encodes a putative 229-amino-acid protein exhibiting a high level of similarity to a hypothetical protein from E. faecium (Table 4).

Finally, region III is comprised of three ORFs (orf8 to orf10). orf8 is preceded by a putative promoter (Fig. 2) and is predicted to encode a 245-amino-acid protein with similarity to the replication initiator protein RepE from E. faecalis (23) (Table 4). The region containing the putative promoter of orf8 includes two sets of direct repeats (12 bp and 22 bp), which is a feature typical of a putative plasmid replication origin (ori) (20).

orf9, which is located 26 bp downstream of orf8, is predicted to encode a 178-amino-acid protein with a putative DNA-binding domain, exhibiting sequence similarity to a protein having an unknown function from E. faecium and to RepB proteins in data banks (Table 4).

orf10, found 13 bp downstream of orf9, encodes a putative 65-amino-acid peptide showing no significant sequence similarity to any known protein in data banks.

EntqBC is involved in secretion and immunity.

Genes involved in transport and immunity are often closely associated with bacteriocin structural genes (17, 41, 52). We are therefore interested in determining whether entqBC are involved in such a relationship. To examine their possible roles in transport, different DNA fragments from entqABC were cloned into the expression vector pELS200. The resulting recombinant plasmids, including pRCG01 containing only entqA, pRCG02 containing entqBC but not entqA, and pRCG03 containing entqABC (Table 1), in addition to an empty plasmid, were transferred into the heterologous host L. sakei Lb790, which is a naturally EntQ-resistant strain. Bacteriocin production in culture supernatants from the resulting Lb790 clones was assessed by an ADT. As shown in Fig. 5A, only clone Lb790(pRCG03), expressing all three genes (entqABC), was able to produce antimicrobial activity in supernatant against growth of E. faecium P13, while the other two clones, Lb790(pRCG01) and Lb790(pRCG02), were not able to do this. The presence of EntQ in the supernatant of Lb790(pRCG03) was confirmed, and the EntQ quantified by an NCI-ELISA using anti-EntQ-KLH-specific antibodies (data not shown). The maximum heterologous production of EntQ (390 ng/ml) by Lb790(pRCG03) was obtained after 14 h of growth at 32°C in MRS broth; this corresponds to 22% of the maximum amount of EntQ produced by cultures of E. faecium L50 grown under similar conditions (Criado et al., submitted for publication). While no EntQ was found in the culture supernatant of Lb790(pRCG01), the bacteriocin was detected in the intracellular extract from an overnight culture by an NCI-ELISA (data not shown), demonstrating that the region encompassing entqBC is involved in ABC transporter-mediated secretion of EntQ.

FIG. 5.

(A) Bacteriocin activity in supernatants from cultures of L. sakei Lb790(pELS200) (HpELS), E. faecium L50 (W), L. sakei Lb790(pRCG01) (Hp01), L. sakei Lb790(pRCG02) (Hp02), and L. sakei Lb790(pRCG03) (Hp03) grown in MRS broth at 32°C and MRS broth containing erythromycin (5 μg/ml) (MRS-E) as determined by an ADT using E. faecium P13 (P13) (EntL50^s EntP^r EntQ^s) as the indicator microorganism. (B) Bacteriocin immunity or sensitivity of E. faecium L50/14-2(pRCG04-1) (Hp04-1), E. faecium L50/14-2(pRCG04-2) (Hp04-2), and E. faecium L50/14-2(pMG36e) (HpMG) to supernatants from cultures of the mutant E. faecium LA5 containing EntL50 and EntP (L50+P) and E. faecium P13 containing EntP (P), both grown in MRS broth at 32°C, and to chemically synthesized EntQ (0.5 mg/ml) (Q).

During cloning of the fragment containing entqBC into lactococcal plasmid pMG36e, we obtained two plasmid clones, each of which had a mutation in entqB. The mutations caused truncation of EntqB, while EntqC was still intact in these clones. The truncated EntqB (393 amino acids) encoded by the pRCG04-1 clone lacked most of the C-terminal half, including the C motif (involved in ATP hydrolysis and mediating conformational changes induced by ATP hydrolysis between the ATPase and the transmembrane domains), the Walker B motif (involved in nucleotide binding and ATPase activity), and the switch region (involved in signal transduction to the ABC domain by sensing conformational changes in the membrane-spanning domains upon substrate binding) (15, 50). The second clone (pRCG04-2) with a truncated version of EntqB (540 amino acids) lacked the last 32 amino acids that encompass the switch region (Fig. 4A). Therefore, we assumed that the functionality of entqB, which encodes an ABC transporter and is involved in secretion of EntQ, was destroyed in the two clones, thereby providing the possibility of examining whether the remaining entqC gene in these two clones can confer immunity. These two plasmid clones were therefore transformed into the EntQ-sensitive plasmidless mutant E. faecium L50/14-2 (Table 1), and the resulting transformants were examined for immunity to EntQ. The ADT immunity analyses were performed with the transformants E. faecium L50/14-2(pRCG04-1) and E. faecium L50/14-2(pRCG04-2), and the results conclusively showed that both strains were immune to EntQ, while the control clone E. faecium L50/14-2(pMG36e) was inhibited by this bacteriocin under the same conditions (Fig. 5B). Interestingly, the acquired immunity was specific to EntQ as these two clones were sensitive to EntL50 (Fig. 5B). Besides, by using an MPA, it was shown that the EntQ immunity level of E. faecium L50/14-2(pRCG04-1) and E. faecium L50/14-2(pRCG04-2) was the same as that exhibited by the wild-type strain E. faecium L50; i.e., their growth was not affected by chemically synthesized EntQ at a concentration of 125 μg/ml. In comparison, the growth of the control clones E. faecium L50/14-2(pMG36e) and E. faecium L50/14-2 was inhibited by a much lower concentration of EntQ (2.0 μg/ml). Taken together, all these results established that EntqC is involved in EntQ-dedicated immunity, conferring full protection against the toxicity of this bacteriocin.

Purification and characterization of the heterologously produced EntQ.

The EntQ heterologously produced by L. sakei Lb790(pRCG03) was purified to homogeneity. The last reversed-phase fast protein liquid chromatography resulted in a single well-separated absorbance peak that coincided with the antimicrobial activity peak (results not shown). The purity and molecular mass of EntQ were confirmed by matrix-assisted laser desorption ionization—time of flight mass spectrometry, which revealed a single peptide with the expected molecular mass (3,952 Da) (13). The purified EntQ produced a band of the expected size similar to the band produced by chemically synthesized EntQ when it was analyzed on a silver-stained Tricine-SDS-PAGE gel (Fig. 6A). The purified bacteriocin and bacteriocin aggregates, likely due to hydrophobic interactions, were detected by Western blotting using the anti-EntQ-KLH antibodies specific to EntQ (Fig. 6B), and they were shown to be biologically active by an overlay assay (Fig. 6C). Based on these results, we concluded that the heterologously produced EntQ is identical to the EntQ produced by the wild-type strain E. faecium L50 (13).

FIG. 6.

(A) Tricine-SDS-PAGE of EntQ heterologously produced by L. sakei Lb790(pRCG03) after silver staining. (B) Western blotting using specific anti-EntQ-KLH antibodies. (C) Antimicrobial activity after overlay with the indicator strain E. faecium P13. Lane 1, 1 μg of pure chemically synthesized EntQ; lane 2, pure EntQ heterologously produced by L. sakei Lb790(pRCG03). The positions of ultra-low-range molecular mass markers (Sigma) are indicated on the left.

DISCUSSION

Genetic determinants responsible for multiple bacteriocin production have been described for a number of systems. In the case of the plantaricin (16) and sakacin (37) systems, the genes involved are clustered in a locus-like unit and exposed to a common regulatory system. In this work, we demonstrated that the genetic determinants involved in the multiple bacteriocin production by E. faecium L50 are organized differently and occur at different genetic locations, as follows: entL50AB, entqA, and entP are located on the 50-kb plasmid pCIZ1, the 7.4-kb plasmid pCIZ2, and the chromosome, respectively (Fig. 1A). Likewise, Carnobacterium piscicola LV17 produces three carnobacteriocins, carnobacteriocins A, BM1, and B2, whose structural genes are located on the chromosome and 72-kb and 61-kb plasmids, respectively (45, 57). The genetic organization found in the enterocin system might allow production of the individual bacteriocins to be subjected to different regulatory regimens, each of which is dependent on specific environmental signals and growth conditions. In fact, the level of production of EntL50 is high at 16 to 25°C but negligible when the growth temperature is above 37°C, whereas EntP and EntQ are synthesized in a wider temperature window, from 16 to 47°C (13; Criado et al., submitted for publication).

Nucleotide sequence analysis of pCIZ2 from E. faecium L50 revealed the presence of the EntQ structural gene (entqA) and two divergently oriented genes, entqB and entqC, which were shown to be involved in ATP-mediated bacteriocin transport and immunity, respectively. These notions are based on the results of expression of these genes in L. sakei Lb790 and E. faecium L50/14-2. L. sakei Lb790(pRCG01) expressing only entqA was found to be unable to secrete the bacteriocin into the growth medium. Only when entqA was coexpressed with entqBC in L. sakei Lb790(pRCG03) was the bacteriocin secreted into the medium (Fig. 5A). Similarly, the cloned entqC (together with truncated and apparently nonfunctional entqB mutants) was found to confer immunity to E. faecium L50/14-2(pRCG04-1/2) (Fig. 5B).

The role of ABC transporters in processing and secretion of leader-containing class II bacteriocins is well documented (14, 33, 36, 39, 41, 52). Furthermore, it has been reported that secretion of the leaderless bacteriocin LsbB from L. lactis BGM-1 (28) and secretion of aureocin A70 from Staphylococcus aureus A70 (42) are mediated by the ABC-type multidrug resistance transporter LmrB and the ABC transporter AurT, respectively; also, it has been suggested that secretion of the leaderless enterocin EJ97 from E. faecalis EJ97 is mediated by the ABC transporter Ej97B (47). Interestingly, EntqB exhibits extensive homology to these ABC-type transporters (Table 4 and Fig. 4A). There are two important features that distinguish the transporters directing leaderless bacteriocin secretion from their counterparts directing leader-containing bacteriocin secretion. First, the former proteins, including EntqB, lack the typical N-terminal proteolytic domain that is found in and required by the latter proteins to process and direct transport of leader-containing bacteriocins (30). Second, in the transport of leader-containing bacteriocins, a so-called accessory protein functions together with the cognate ABC transporter to mediate bacteriocin secretion, and its gene is commonly located just downstream of the ABC transporter gene (17, 41); such a gene is not required for transporters directing leaderless bacteriocin secretion. Based on these findings, EntqB is therefore believed to be involved in transport of the leaderless bacteriocin EntQ. This notion is further supported by the coexpression of the bacteriocin structural gene (entqA) with entqB, which allows the bacterium to secrete active EntQ into the growth medium (Fig. 5A).

The immunity gene of most nonlantibiotics is located immediately downstream of, and in the same operon as, the bacteriocin structural gene (17, 41, 52). However, there are some examples of atypical locations, such as the immunity genes of bacteriocins LsbA and LsbB, enterocin B, and carnobacteriocin A, which are located next to the structural genes but have the opposite orientation (25, 26, 28). With regard to this, entqA is neither followed nor preceded by a typical immunity gene. However, entqC, which is located immediately downstream of entqB, encodes a putative 64-amino-acid protein (EntqC) with some physical properties typical of immunity proteins (14, 16, 22, 52), such as (i) a high pI value (pI 9.63), (ii) a high hydrophobic residue content (50%), and (iii) the presence of two predicted transmembrane segments, previously reported to be necessary for the integration of immunity proteins into the membrane of the bacteriocin producers (27). It is noteworthy that the genetic content of the EntQ gene cluster found in pCIZ2 is similar to that of the mundticin KS gene cluster in Enterococcus mundtii NFRI 7393 (36) and that of the enterocin EJ97 gene cluster in E. faecalis EJ97 (47). In these systems, the bacteriocin structural gene (munA or ej97A) is followed by an ABC transporter gene (munB or ej97B) and a gene (munC or ej97C) encoding a small protein (71 to 98 amino acids). While EJ97C is a hypothetical protein having an unknown function, MunC exhibits homology to other bacteriocin immunity proteins and functions as the mundticin KS immunity protein. In the present study, a 2,207-bp fragment encompassing an intact entqC (and a truncated entqB) was shown to confer full immunity to EntQ, thus suggesting strongly that EntqC plays a role in host self-protection against EntQ. However, one might argue that the remaining part of the truncated forms of EntqB in pRCG04-1 and pRCG04-2 might have an immunity function. It is noteworthy that these EntqB variants (in particular, the one encoded by pRCG04-1, which is only 393 amino acids long) are truncated from the C terminus, where important conserved motifs required for functional ABC transporters (15, 50) are missing (Fig. 4). Nevertheless, a possible role of EntqB in immunity can be clarified only by further investigation.

The genetic determinants involved in multiple bacteriocin production frequently occur in clusters; each bacteriocin has its own dedicated immunity protein, while the bacteriocins share the transport system (39, 41, 52). In this sense, it seems that secretion of the leaderless bacteriocins EntQ and EntL50 does not require a common ABC transporter, which was deduced from the following observations: (i) the loss of pCIZ2 in the mutant E. faecium L50/30-2 did not affect the production of EntL50 (Fig. 1C, left panel) and (ii) the presence of the recombinant plasmid pRCG03 in L. sakei Lb790 was sufficient for EntQ production and secretion (Fig. 5A). Likewise, it also seems that EntQ and EntL50 do not have a common immunity protein, since (i) the loss of pCIZ2 in the mutant E. faecium L50/30-2 did not alter the immunity to EntL50 (Fig. 1D, upper and lower panels) and (ii) the presence of recombinant plasmid pRCG04-1 or pRCG04-2 in the host E. faecium L50/14-2 conferred immunity to EntQ but not immunity to EntL50 (Fig. 5B, upper and middle panels).

The biochemical, genetic, and immunological data presented in this paper demonstrate that the 2,371-bp fragment from the 7.4-kb pCIZ2 plasmid encompassing entqA, entqB, and entqC contains the genetic determinants required for EntQ production and immunity. Although the presence of the ABC transporter EntqB alone is probably sufficient for exporting the leaderless EntQ, the mechanism by which EntqB acts in this process and what feature(s) on the leaderless bacteriocin peptide serves as a signal for export are still unknown. Unraveling this process is therefore a challenging task that has great scientific interest.