Sec-Mediated Secretion of Bacteriocin Enterocin P by Lactococcus lactis

ABSTRACT Most lactic acid bacterium bacteriocins utilize specific leader peptides and dedicated machineries for secretion. In contrast, the enterococcal bacteriocin enterocin P (EntP) contains a typical signal peptide that directs its secretion when heterologously expressed in Lactococcus lactis. Signal peptide mutations and the SecA inhibitor azide blocked secretion. These observations demonstrate that EntP is secreted by the Sec translocase.

Ribosomally synthesized antimicrobial peptides (bacteriocins) produced by lactic acid bacteria have potential applications as biopreservatives in the food industry (32). Enterocin P (EntP) is a class IIa bacteriocin (10) produced by Enterococcus faecium P13. It inhibits the growth of some food spoilage and food-borne pathogens, such as Listeria monocytogenes, Staphylococcus aureus, and Clostridium botulinum (5). The genetic determinants responsible for EntP production are organized in one operon comprising the bacteriocin structural gene, entP, and orf2, encoding a putative immunity protein (5). The EntP precursor (preEntP) consists of 71 amino acid residues (7.5 kDa) and contains a signal peptide of 27 amino acids that is removed proteolytically upon secretion, yielding the mature EntP (4.6 kDa).
The absence or presence (and type) of an N-terminal extension determines the secretion mechanism of class II bacteriocins. So far, the enterocins L50A, L50B, and Q (6) and the lactococcal bacteriocin LsbB (12) are the only examples of class II bacteriocins that are synthesized without N-terminal extension. The secretion mechanism of these enterocins is unknown, although LsbB has been reported to be secreted by LmrB, a multidrug resistance-like ABC transporter (12). Most class II bacteriocin precursors contain a double-glycine-type signal peptide and are translocated and processed by dedicated ABC transporters (15). Finally, preEntP and some other class II bacteriocins (5,9,22,26,37,39) contain a typical Sec signal peptide (19) that is believed to direct these bacteriocins to the Sec translocase embedded in the cytoplasmic membranes (7). However, preEntP is a small, amphiphatic polypeptide that to some extent resembles procoat M13, the precursor of a small phage protein which does not utilize the Sec translocase despite containing a typical Sec-dependent signal sequence (2). Therefore, the objective of this work was to determine the secretion mechanism of preEntP.

MATERIALS AND METHODS
Bacterial strains, plasmids, and growth conditions. The bacterial strains and plasmids used in this study are shown in Table 1. E. faecium P13 (5) and T136 (3) were used as an EntP producer strain and sensitive indicator strain, respectively. Enterococci were grown at 30°C in MRS broth (Oxoid Ltd., Basingstoke, United Kingdom). Lactococcus lactis NZ9000 was used in combination with the nisincontrolled expression system (20) for protein overexpression. L. lactis was grown at 30°C in M17 medium (Difco, Le Pont de Claix, France) supplemented with 0.5% (wt/vol) glucose and 5 g of chloramphenicol per ml when appropriate. Chemically defined medium (CDM) (21) was used to grow L. lactis for pulselabeling experiments.
Recombinant DNA techniques. Cloning and DNA manipulation were performed as described by Sambrook et al. (36). Oligonucleotide primers used for PCR are shown in Table 2. E. faecium P13 and L. lactis MG1363 (13) chromosomal DNAs were used as templates for the PCR amplification of entP (5) and usp45 (38), respectively. The PCR overlap extension method (17) was used to introduce single-amino-acid substitutions and to construct gene fusions of the signal peptides of preEntP (SP Ent ) and preUsp45 (SP Usp ) to the mature regions of preUsp45 and preEntP, respectively. PCR products were cloned in frame with a C-terminal hexahistidine tag on the pHLP5 expression plasmid (33). Sequencing of the PCR-amplified DNA fragments was performed at Westburg Genomics (Leusden, The Netherlands).
Purification of histidine-tagged EntP and Usp45. Ammonium sulfate (45% [wt/vol]) was gradually added to the culture supernatants of L. lactis NZ9000 cells producing C-terminal hexahistidine-tagged EntP (EntP-His) or Usp45 (Usp45-His). After centrifugation, the precipitated proteins were resuspended in 50 mM sodium phosphate buffer (pH 8.0) with 300 mM NaCl (buffer A). Fractions were desalted by gel filtration (G-25 PD-10 columns, Amersham Biosciences, Uppsala, Sweden) and mixed with Ni-nitrilotriacetic acid (NTA) resin (QIAGEN, Hilden, Germany) equilibrated with buffer A (pH 8.0), and suspensions were gently shaken overnight at 4°C. The resin was transferred to a Bio-spin column (Bio-Rad, Hercules, Calif.) and subsequently washed with 10 column volumes of buffer A (pH 8.0) containing 75 or 10 mM imidazole for EntP-His and Usp45-His, respectively. Proteins were eluted with buffer A (pH 7.0) containing either 1 or 0.25 M imidazole for EntP-His and Usp45-His, respectively. Eluted fractions were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting using monoclonal antibodies directed against the hexahistidine tag (Dianova, Hamburg, Germany) for chemiluminiscence detection with the Lumi-Imager F1 (Roche Diagnostics). N-terminal sequencing of EntP was performed at the University of British Columbia (Vancouver, Canada).
Pulse-labeling analysis. Overnight cultures of L. lactis cells transformed with the appropriate plasmids and grown on CDM were freshly diluted (1:40) in 5 ml of the same medium. After incubation at 30°C to an optical density at 660 nm of 0.6, cells were collected by centrifugation, washed, and resuspended in CDM without methionine. Nisin was added, and incubation at 30°C was continued for 30 min. Where stated, cells were treated with sodium azide prior to the pulse to inhibit Sec-dependent protein secretion. Pulse-labeling was performed by the addition of 16 Ci of the Pro-mix L-( 35 S) in vitro cell labeling mix per ml (Amersham Biosciences, Roosendaal, The Netherlands). During the chase, nonradioactive methionine (25 mM) and cysteine (10 mM) were added, and at various times, 500-l samples were removed and immediately centrifuged. The proteins present in the supernatants were precipitated with trichloroacetic acid (TCA; 10% [vol/vol]). Precipitates were washed with 80% (vol/vol) ice-cold acetone, dried, resuspended in SDS-PAGE sample buffer, and analyzed by SDS-PAGE and phosphorimaging. Cell pellets were resuspended in 100 l of lysis buffer (50 mM NaCl, 10 mM EDTA, 20% [wt/vol] sucrose, 10 mM Tris, pH 8.1, and 10 mg of lysozyme per ml) and incubated at 50°C for 10 min. Next, 10 l of 10% (wt/vol) SDS was added, and after boiling for 10 min, the samples were diluted 10 times with buffer A (pH 8.0) containing 0.05% (wt/vol) dodecyl-␤-Dmaltoside (Anatrace, Maumee, Ohio) and 100 l of pre-equilibrated Ni-NTA resin. After 2 h at 4°C with shaking, proteins were eluted with buffer A (pH 7.0) containing 0.05% (wt/vol) DDM and 250 mM imidazole and further analyzed as described above.
Antimicrobial activity assays. The antimicrobial activity of EntP-His was measured by the agar well-diffusion assay (4) and quantified by a microtiter plate assay (18) using E. faecium T136 (3) as the indicator microorganism.

RESULTS
Heterologous expression of entP in L. lactis. Production of EntP-His (6 kDa) in L. lactis NZ9000 was achieved when cells containing entP were induced with nisin. Coomassie-stained SDS-PAGE of the protein purified by Ni-NTA affinity chromatography showed a single band that corresponded to EntP-His ( Fig. 1), as verified by N-terminal sequence analysis of the first six amino acid residues. This result indicates that preEntP is processed by L. lactis in an identical manner as in E. faecium. The heterologously produced EntP-His displayed antimicrobial activity against E. faecium T136 (not shown).
Signal peptide swapping between preEntP and preUsp45. Hybrid genes encoding the precursor proteins composed of the SP Ent fused to the mature Usp45 (entP-usp45) and vice versa   (usp45-entP), were constructed. Expression of entP-usp45 resulted in high-level production of a protein of 50 to 75 kDa (Fig. 2) that corresponded to Usp45-His, as verified by Western blotting (results not shown). Although the native signal peptide supported more efficient secretion of Usp45-His (Fig.  2), the data show that the SP Ent can direct the secretion of a Sec-dependent protein. Nisin induction of L. lactis cells containing usp45-entP resulted in the production of EntP-His (Fig.  3A), which displayed a normal antimicrobial activity (Fig. 3B). These results show that the SP Ent and the SP Usp are functionally exchangeable and indicate that the SP Ent is a typical Secdependent signal peptide.

Secretion of EntP-His directed by mutated signal peptides.
To further investigate the characteristics of the SP Ent , point mutations were introduced into its hydrophobic core. Replacement of the isoleucine residues at positions 11 and 13 with asparagine and of the glycine at position 12 with lysine, yielded the mutated signal peptides I11N, I13N, and G12K, respectively. Analysis of the secretion of EntP-His by a microtiter plate assay showed that only 0.6, 0.5, and 2.5% of the antimicrobial activity produced by the cells expressing the wild-type precursor was obtained when secretion was directed by the mutated signal peptides I11N, G12K, and I13N, respectively.
Effect of azide on EntP secretion. Azide, a potent inhibitor of SecA, blocks Sec-dependent protein secretion in Escherichia coli and Bacillus subtilis (29,31). The effect of azide on the SP Ent -mediated secretion in L. lactis was evaluated by using pulse-labeling assays. Secretion of EntP-His or Usp45-His directed by the SP Ent was drastically reduced in the presence of azide ( Fig. 4A and B), while for SP Ent -Usp45-His, the amount of precursor remaining inside the cells increased (Fig. 4C). Similarly, secretion of preUsp45-His was blocked by azide (re-sults not shown). These data indicate that EntP secretion in L. lactis is mediated by the Sec translocase.

DISCUSSION
In this report, we have investigated the mechanism of secretion of the heterologously produced EntP-His in L. lactis. EntP belongs to the class II bacteriocins that are synthesized as precursors with an N-terminal extension that shows the typical tripartite structure of Sec-dependent signal peptides (19) (Table 3). The assumption of a Sec-dependent secretion of bacteriocin 31 and acidocin B is supported by their production by heterologous hosts in which only the bacteriocin structural and immunity genes have been introduced (22,37). For divergicin A, additional evidence is provided by the ability of its signal peptide to direct the secretion of Sec-dependent (39) and Secindependent (1,27,28) proteins. In the case of EntP, the presence of a seemingly typical signal sequence (5) and the lack of dedicated transport genes in the entP operon (16) suggested that this bacteriocin is secreted via the Sec pathway. However, there are some unusual characteristics in the preEntP amino acid sequence which prompted us to investigate its secretion mechanism. First, there is a positively charged residue at the ϩ3 position of the mature EntP. Although this feature is not uncommon in class II bacteriocins secreted by dedicated ABC transporters, the presence of pos- itively charged amino acids in the N-terminal region of the mature proteins negatively affects Sec-dependent secretion in E. coli (25) and L. lactis (23,24). Second, a lysine residue is present at position Ϫ4 relative to the cleavage site in preEntP. This attribute is not observed in other putative Sec-dependent class II bacteriocins and would render the signal peptide inactive for Sec-dependent secretion in E. coli. Finally, asparagine, a residue rarely present in prokaryotic signal peptides (14), is located in the hydrophobic core of SP Ent . Since no dedicated transporter has been defined for EntP, the signal sequence should support its secretion via the Sec translocase when expressed in a heterologous host. Indeed, L. lactis could secrete EntP-His either with the native or with the Usp45 signal peptide. On the other hand, secretion of a typical Sec substrate as Usp45 (38) is supported by the SP Ent . Finally, secretion is dramatically reduced by signal sequence mutations that interfere with the secretion of Sec-dependent substrates in E. coli (14) and is blocked by the SecA inhibitor azide. Taken together, these results lend strong support for the hypothesis that EntP is secreted into the medium by the Sec translocase.
In an attempt to detect special features of the SP Ent that could be required for the bacteriocin secretion, the ability of the SP Usp to direct EntP-His secretion was evaluated. Previous work using SP Usp for heterologous protein expression showed that the combination of this signal peptide with a negatively charged N terminus in the mature protein increased the secretion efficiency (8,24,34). Combination with the cationic N terminus of EntP appears at least as efficient as the native SP Ent . Apparently, the positive charge in the mature amino terminus does not interfere with a proper functioning of the SP Usp in L. lactis. Several bacteriocins have been heterologously produced in L. lactis, using an ABC-dedicated transport system (reviewed in reference 35). However, a limited number (28,30,39) have been produced through secretion by the Sec system. This study shows that L. lactis is an excellent host for the production of Sec-dependent bacteriocins. Since L. lactis is not sensitive to EntP (5), the bacteriocin could be expressed without the immunity protein Orf2. This makes the food-grade microorganism L. lactis an ideal host for the heterologous production of EntP, thereby avoiding concerns about the safety of enterococci in food (11). FIG. 4. Effect of sodium azide on the SP Ent -mediated secretion. L. lactis NZ9000 cells expressing SP Ent -EntP-His (A) or SP Ent -Usp45-His (B and C) were pulse-labeled with a mixture of radioactive methionine and cysteine. Where indicated, cells were treated with 3 mM sodium azide 1 min prior to the pulse. Samples were taken at the indicated times (in minutes) after the chase, and the supernatants (A and B) and the cell pellets (C) were processed as described in the text. a Shown are the amino acid sequence of the signal peptides and the first five residues after the processing site, which is indicated by an arrow.

ACKNOWLEDGMENTS
Luis M. Cintas and Pablo E. Hernández are acknowledged for kindly supplying E. faecium strains P13 and T136. We are grateful to Girbe Buist, Jan Jongbloed, Nico Nouwen, and Aldert Zomer for helpful advice with pulse-labeling experiments and to Nathalie Campo and Piotr Mazurkiewicz for stimulating discussions.
C. Herranz is supported by a Marie Curie Individual Fellowship (contract HPMF-CT-2002-01660) from the IHP Programme of the European Commission.