Heterologous Inducible Expression of Enterococcus faecalis pCF10 Aggregation Substance Asc10 in Lactococcus lactis and Streptococcus gordonii Contributes to Cell Hydrophobicity and Adhesion to Fibrin

Abstract

Aggregation substance proteins encoded by the sex pheromone plasmid family of Enterococcus faecalis have been shown previously to contribute to the formation of a stable mating complex between donor and recipient cells and have been implicated in the virulence of this increasingly important nosocomial pathogen. In an effort to characterize the protein further, prgB, the gene encoding the aggregation substance Asc10 on pCF10, was cloned in a vector containing the nisin-inducible nisA promoter and its two-component regulatory system. Expression of aggregation substance after nisin addition to cultures of E. faecalis and the heterologous bacteria Lactococcus lactis and Streptococcus gordonii was demonstrated. Electron microscopy revealed that Asc10 was presented on the cell surfaces of E. faecalis and L. lactis but not on that of S. gordonii. The protein was also found in the cell culture supernatants of all three species. Characterization of Asc10 on the cell surfaces of E. faecalis and L. lactis revealed a significant increase in cell surface hydrophobicity upon expression of the protein. Heterologous expression of Asc10 on L. lactis also allowed the recognition of its binding ligand (EBS) on the enterococcal cell surface, as indicated by increased transfer of a conjugative transposon. We also found that adhesion of Asc10-expressing bacterial cells to fibrin was elevated, consistent with a role for the protein in the pathogenesis of enterococcal endocarditis. The data demonstrate that Asc10 expressed under the control of the nisA promoter in heterologous species will be an useful tool in the detailed characterization of this important enterococcal conjugation protein and virulence factor.

The gram-positive intestinal commensal Enterococcus faecalis is host to a variety of mobile genetic elements ranging from conjugative plasmids to conjugative transposons (7). Among the conjugative elements, members of the sex pheromone plasmid family exhibit a novel mechanism of plasmid transfer. Expression of conjugation functions in donor cells is triggered by the secretion of hydrophobic hepta- or octapeptide sex pheromones produced by plasmid-free recipient cells (14). These peptides are sensed by a plasmid-containing donor cell, and the uptake of the peptide induces the expression of the surface protein aggregation substance (AS) (34). AS allows the formation of macroscopic aggregates of donor and recipient cells, with subsequent plasmid transfer reaching frequencies of 10⁻¹ to 10⁻² transconjugants per donor (12).

A multitude of sex pheromone plasmids have been described (51), and with one exception (plasmid pAM373) all encode ASs that have highly homologous DNA and protein sequences (23, 25). Most genes for AS encode proteins of approximately 137 kDa. Sequence information on the three ASs of the plasmids pAD1, pCF10, and pPD1 has become available to date (21, 22, 29). The genes encode proteins consisting of roughly 1,300 amino acids with apparent molecular masses of approximately 150 to 160 kDa, as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (PAGE). A 74- to 78-kDa fragment corresponding to the amino terminus of the protein is commonly detected in cell surface extracts (17, 24). The proteins each contain a relatively long signal peptide of 44 amino acids and possess the widespread gram-positive cell wall anchor motive LPXTG (39). Several groups used electron microscopy and immunological techniques to determine the appearance of induced cells and the distribution of the protein on the cell surface (24, 40, 49). No apparent structural features are prominent in the protein, and its predicted overall shape is globular. One intriguing feature is the presence of two RGD motifs in the proteins. This amino acid sequence is present in a number of eucaryotic proteins and is implicated in the binding of these proteins to eucaryotic cell surface molecules of the integrin family (43).

The presence of the RGD sequences suggested involvement of AS in virulence, and the work of several groups, using different model systems, supports this hypothesis (4, 31, 44). The regulation of AS expression in the wild-type plasmid context in the two best-characterized systems, pCF10 and pAD1, is rather complex. In the case of pCF10, a promoter 5 kb upstream of the structural gene prgB is involved (6), whereas in the pAD1 system a trans-acting regulatory protein is necessary for expression (37, 47). Small RNA molecules are implicated in the regulation of AS expression in both plasmids (2, 9). To simplify characterization of the AS protein by avoiding possible complicating factors of the pheromone plasmid regulatory systems, we chose to use the well-characterized nisin-inducible promoter nisA, which was recently demonstrated to be functional in E. faecalis (18).

Nisin belongs to the lantibiotic class of peptide antibiotics produced by certain strains of L. lactis (33). In the nisin biosynthetic cluster, the promoters preceding nisA, the nisin peptide structural gene, and nisF, involved in immunity to nisin, are inducible by mature nisin. Nisin serves as a peptide pheromone sensed by a classic two-component regulatory system, with NisK as a histidine kinase and NisR as a response regulator (19, 30, 32). In L. lactis, expression of the nisA promoter increases linearly with the amount of nisin present; however, in other bacterial species, this linearity is seen only in the nanomolar range of nisin concentrations (18).

Here we describe the expression of the pCF10 AS under the control of the nisA promoter in E. faecalis as well as in the heterologous hosts Lactococcus lactis and Streptococcus gordonii. The expression of AS increases the cell surface hydrophobicity remarkably and allows increased adherence to fibrin, thus supporting the role of the protein in virulence. The functionality for adhesion of the expressed protein to E. faecalis cells is demonstrated by the increased transfer of a conjugative transposon into AS-expressing hosts.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

E. faecalis and S. gordonii were grown without shaking at 37°C in Todd-Hewitt broth (Difco, Detroit, Mich.). L. lactis was grown without shaking at 30°C in M17 medium (Difco) supplemented with 0.5% glucose. The host strain for molecular cloning was Escherichia coli DH5α grown at 37°C in Luria-Bertani medium or, if erythromycin was used as a selective marker, in brain heart infusion broth (Difco). Agar plates contained 1.5% agar. The following antibiotic concentrations were used: for E. coli, kanamycin at 35 μg/ml and erythromycin at 100 μg/ml; for E. faecalis, erythromycin at 10 μg/ml and streptomycin at 1 mg/ml; and for L. lactis and S. gordonii, erythromycin at 10 μg/ml.

PCR amplification of the prgB gene.

The prgB wild-type fragment containing plasmid pINY1801 (5) was utilized as the template for cloning of prgB. The gene was amplified in two segments. The 5′ end was amplified with the primer pair 5′ TGATCCATGG-------------AGGAGGATGATACATG 3′, containing an NcoI site (broken underline) and a base change at position 13 (boldfaced) from A→G to change the ribosome binding site (RBS; double underlined) to GGAGG and thereby improve it, and 5′ TTTTCACATATGAGCCATTAATGTATTTGAACGC 3′. The start codon for the prgB gene is single underlined. The primers used for amplification of the remainder of the gene were 5′ GCAACTTCCAGAATTCC 3′ and 5′ AATTACTCGAG-------------CCAATTTTTCCCCTCC 3′, with the latter containing an XhoI site (broken underline). PCR was performed with a Hybaid thermocycler, using Vent polymerase (New England Biolabs, Beverly, Mass.). The amplified products were purified by the use of a DNA clean-up kit (Promega Corporation, Madison, Wis.) and employed in further manipulations.

Molecular cloning procedures.

Restriction enzymes (Promega) and T4 DNA ligase (Gibco BRL, Gaithersburg, Md.) were used in accordance with the manufacturers' recommendations. Plasmid DNA was isolated from E. coli by the use of a plasmid minikit (Qiagen Inc., Chatsworth, Calif.). Electrotransformation was performed with a Gene Pulser apparatus (Bio-Rad Corp., Richmond, Calif.) as previously described for E. coli (20), E. faecalis (13), and L. lactis (27). S. gordonii was transformed as described previously for Streptococcus mutans (45).

Nisin induction.

A preparation containing 2.5% nisin (Sigma Chemical Co., St. Louis, Mo.) was used. A stock solution of 10 mg/ml (effective nisin concentration, 250 μg/ml) in H₂O was prepared and stored at −20°C. Appropriate dilutions were made for induction of the bacterial strains.

Expression of AS and Western blot analysis.

The strains were inoculated in fresh medium at a 1:10 dilution from overnight cultures and grown for 4 h under appropriate conditions. Nisin was added as described above. Cells were sedimented, and a lysozyme cell surface extraction was performed as described previously (22). For analysis of AS in culture supernatant, cells were sedimented and the supernatant was filtered through a 22-μm-pore-size filter (Millipore, Bedford, Mass.). The filtrate was precipitated with 4 volumes of ethanol overnight at 4°C, resuspended in H₂O, and used for PAGE. Protein concentrations were determined by the bicinchoninic acid method (Pierce, Rockford, Ill.). The extracts were separated by sodium dodecyl sulfate–7.5% PAGE and transferred to a BA 85 nitrocellulose membrane (Schleicher and Schuell, Keene, N.H.). Western blot analysis was performed with an antibody against pAD1 AS (kindly provided by A. Muscholl-Silberhorn, Universität Regensburg, Regensburg, Germany) in accordance with an enhanced chemiluminescence protocol (Pierce).

Electron microscopy.

Cells were grown and induced as described above. A 1-ml volume of the cell culture was pelleted, washed with phosphate-buffered saline (PBS), resuspended in PBS containing 5% goat serum, and incubated with a 20-μg/ml solution of a monoclonal antibody against AS for 2 h. The cells were washed three times with PBS–5% goat serum, incubated for 1 h with goat anti-mouse immunoglobulin G conjugated to 12-nm-diameter colloidal gold particles (Jackson ImmunoResearch Laboratories, West Grove, Pa.) diluted 1:50, and subsequently washed with PBS three times. The cells were concentrated by centrifugation at low speed and then placed on poly-l-lysine (Sigma)-covered glass supports (5 by 10 mm). The cells were allowed to adhere for 30 min and then were washed and fixed in 0.1 M cacodylate buffer containing 3% glutaraldehyde and 7.5% sucrose. The bacteria were examined by backscatter electron imaging, using an AUTRATA modified YAG detector with a Hitachi S-900 field emission scanning microscope at 5 keV as described previously (40).

Transposon transfer.

E. faecalis INY1010 harboring a single copy of the conjugative transposon Tn925 was used as the donor strain. The donor and recipient strains were inoculated into fresh medium at a 1:10 dilution from overnight cultures, grown for 1 h under the appropriate conditions, and induced with nisin when applicable. The recipient cultures were then inoculated with INY1010 at a 1:10 (donor/recipient) ratio. The cultures were incubated for 8 h under the culture conditions most favorable for the recipient strain. To dissolve cell aggregates, 100 μl of 0.5 M EDTA was added before serial dilutions in 0.9% NaCl were made to determine donor and transconjugant numbers.

Hydrophobicity and fibrin adherence assays.

Cell surface hydrophobicity assays were performed, as previously described (42), with hexadecane (0.25 ml) as the hydrocarbon providing the hydrophobic phase. Hydrophobicity was expressed as the percentage of cells adhering to the hexadecane phase. For fibrin adherence assays, the strains were inoculated into fresh medium at a ratio of 1:10 from overnight cultures, induced with nisin or cCF10 when indicated, and allowed to grow for 2 h under the appropriate conditions. The fibrin adherence assay was performed as described previously (1) with the following modifications. Cells were allowed to adhere to the plates for 30 min at room temperature. After being washed, the fibrin plates were incubated at room temperature for 30 min with a 0.25% trypsin–EDTA solution (Gibco BRL). Dilutions in 0.9% NaCl were prepared for determination of levels of adherent bacteria.

DNA sequencing.

Correct PCR amplification was verified by automated sequencing of the AS construct in plasmid pMSP7516 (Microchemical Facility, University of Minnesota).

RESULTS

Cloning of prgB behind the nisA promoter.

The prgB gene was PCR amplified in two fragments with a primer containing an improved RBS and an NcoI site at the 5′ end and cloned as an NcoI-EcoRI fragment into the plasmid pET28b, creating pMSP7513. The remaining part of prgB was amplified from the EcoRI site to the end of the gene with an XhoI restriction site, creating pMSP7514. Sequencing was performed to verify correct amplification.

The amplified prgB gene was initially intended for cloning into the nisA promoter vector pNZ8048 (Table 1), with the nisKR sensor/regulator component being supplied by plasmid pNZ9531 (18) in trans. Because constructs containing pNZ8048 were unstable in our hands, we constructed pMSP7515, containing the nisA promoter and the cloning sites of pNZ8048 as a BglII-XhoI fragment in pET28b (Fig. 1A). The prgB gene was cloned into pMSP7515 to yield plasmid pMSP7516. For use in a gram-positive host, the nisA-prgB cassette was cloned as a BglII-XhoI fragment into the shuttle vector pDL289 (3). However, this construct resulted in constitutive expression of prgB (data not shown). Therefore, we constructed pMSP3535, containing the nisin response regulator pair as well as the nisA promoter. The orientation of coding reading frames on the plasmid (nisRK, ermAM, and rep) was chosen so that no readthrough should occur. The construction of this vector and the use of the nisin system for controlled expression of other E. faecalis proteins are described by Bryan et al. (E. M. Bryan, T. Bae, M. Kleerebezem, and G. M. Dunny, submitted for publication, 1999).

TABLE 1.

Bacterial strains and plasmids used in this study

Bacterial species or plasmid	Strain or description	Reference or source
Bacteria
E. faecalis	OG1SSp	12
	OG1SSp(pMSP7517)	This study
	OG1SSp(pCF10)	15
	INY1010	48
L. lactis	NZ9800	33
	NZ9800(pMSP7517)	This study
S. gordonii	DL1 (Challis)	41
	DL1(pMSP7517)	This study
E. coli	DH5α	Bethesda Research Laboratories
Plasmids
pET28b		Novagen
pNZ8048	nisA promoter-containing plasmid	M. Kleerebezem
pMSP3535	nisA promoter- and nisRK cassette-containing shuttle vector; Ermr	Bryan et al.a
pMSP7514	prgB gene (Asc10) cloned via NcoI-XhoI into pET28b	This study
pMSP7515	BglII-XhoI fragment of pNZ8048 (nisA promoter and multicloning site)	This study
pMSP7516	prgB gene cloned via NcoI-XhoI into pMSP7515	This study
pMSP7517	BglII-XhoI fragment from pMSP7516 (nisA prgB) cloned into pMSP3535	This study

^aE. M. Bryan, T. Bae, M. Kleerebezem, and G. M. Dunny, submitted for publication, 1999. 

FIG. 1.

(A) Construction of plasmid pMSP7517. The prgB gene, encoding AS, was PCR amplified and cloned into pET28b, yielding pMSP7514. The nisA promoter of plasmid pNZ8048 was cloned in pET28b, resulting in pMSP7515, and prgB was then cloned behind it, creating pMSP7516. The nisA promoter-prgB construct was cloned as a BglII-XhoI fragment into plasmid pMSP3535, resulting in the final pMSP7517 construct. (B) Sequence of pMSP7517 from the nisA promoter −10 region to the start codon of prgB. The −10 region and the RBS are boxed; the boldfaced G indicates the base change in the RBS (see Materials and Methods). The double underline designates the prgB start codon. The arrow marks the transcription start site (10).

The complete nisA promoter-prgB fragment from pMSP7516 was cloned into pMSP3535 by BglII-XhoI restriction to generate pMSP7517. That plasmid was transformed into E. faecalis OG1SSp, L. lactis NZ9800, and S. gordonii DL1 (Challis).

Expression of AS in E. faecalis and heterologous hosts.

The expression of AS under the control of the nisA promoter in E. faecalis, L. lactis, and S. gordonii was investigated by Western blot analysis of cell surface extracts with anti-AS antibodies. We compared levels of nisin-induced expression to those induced by sex pheromone cCF10 in E. faecalis. To determine the optimal nisin concentration for induction, the three species were grown in media containing various concentrations of nisin in the nanogram-per-milliliter range. E. faecalis (Fig. 2A) showed good expression of AS upon nisin induction, with a maximum effective nisin concentration of 25 ng/ml. AS expression from this promoter was not detected in the absence of nisin. The induced cells demonstrated a very distinct ability to form aggregates, confirming that AS is sufficient for formation of tight junctions between the cells.

FIG. 2.

Expression of AS from pMSP7517. Western blot analysis was performed with a polyclonal antibody specific to AS. The protein concentrations in all lanes are identical. (A) E. faecalis OG1SSp(pMSP7517). (B) L. lactis (pMSP7517). (C) S. gordonii(pMSP7517). Lanes 1, OG1SSp(pCF10) induced with cCF10 (10 ng/ml); lanes 2, no nisin; lanes 3, nisin at 1 (A and B) or 0.5 (C) ng/ml; lanes 4, nisin at 5 (A and B) or 1 (C) ng/ml; lanes 5, nisin at 10 (A and B) or 2.5 (C) ng/ml; lanes 6, nisin at 25 (A and B) or 5 (C) ng/ml; lanes 7, nisin at 50 (A and B) or 10 (C) ng/ml; lanes 8, nisin at 100 (A and B) or 25 (C) ng/ml; lanes 9, nisin at 1,000 (A and B) or 50 (C) ng/ml.

L. lactis, not surprisingly, showed good inducible AS expression (Fig. 2B). In comparison to E. faecalis, larger amounts of protein were produced. In addition, there was no drop-off in promoter activity above a nisin concentration of 25 ng/ml. Although the Western analysis of S. gordonii showed expression of AS, in comparison to the other two species the amounts were smaller, with maximum production at a nisin concentration of 5 ng/ml (Fig. 2C). We also transformed pMSP7517 into Bacillus subtilis; however, AS expression was only barely detectable and appeared to be constitutive (data not shown). Also noteworthy is the difference in the AS banding patterns of the various strains. In E. faecalis, the 78-kDa form of AS was detectable at a nisin concentration of 5 ng/ml (Fig. 2A, lane 4), whereas in L. lactis this form was observed only at a nisin concentration of at least 25 ng/ml (Fig. 2B, lane 6). There appeared to be less degradation of the mature form of AS in L. lactis than in E. faecalis, for which the laddering was more abundant. S. gordonii showed very little degradation, which could be partly due to the lower level of AS expressed in this species.

Electron microscopy of nisin-induced cells.

Scanning electron microscopy was performed to determine if AS was correctly displayed on the cell surface. A monoclonal anti-Asc10 antibody was used for labeling, along with a 12-nm-diameter gold particle-labeled secondary antibody. Labeling was successful with E. faecalis and L. lactis expressing AS, as shown in Fig. 3B and C, respectively. Interestingly, nisin-induced S. gordonii(pMSP7517) carried no gold label (data not shown), suggesting that AS was not correctly expressed on the cell surface of this species. The distributions of AS on the cell surfaces of E. faecalis(pMSP7517) and L. lactis were not significantly different from that on wild-type E. faecalis(pCF10) cells induced by cCF10 (Fig. 3A). Uninduced wild-type or pMSP7517-carrying cells did not show gold labeling [Fig. 3D, OG1SSp(pMSP7517)].

FIG. 3.

Electron microscopy of AS-expressing cells. A primary mouse monoclonal antibody against AS was used, followed by a 12-nm-diameter gold particle-labeled secondary antibody (see Materials and Methods). (A) E. faecalis(pCF10), cCF10 induced. (B) E. faecalis(pMSP7517) induced with nisin at 25 ng/ml. (C) L. lactis(pMSP7517) induced with nisin at 25 ng/ml. (D) E. faecalis(pMSP7517), no nisin induction. Scale bars, 0.5 μm.

AS in the culture supernatant.

The surprising finding that S. gordonii did not display AS on the cell surface led us to investigate whether AS is secreted into the culture medium differently in various species. To investigate this possibility, the culture supernatants of all species were ethanol precipitated and subjected to PAGE and Western blot analyses. Surprisingly, E. faecalis, L. lactis, and S. gordonii all showed AS in their culture supernatants (Fig. 4), suggesting that not all of the produced protein is efficiently anchored in the cell wall and that considerable amounts are released into the growth medium. Although this result could explain the lack of labeling for AS on the cell surface of S. gordonii, the presence of similar amounts of AS in the supernatants of the other species argues against there being a differential loss of AS into the medium. The supernatant of an E. faecalis(pCF10) cell culture was also investigated and was found to contain AS in comparable amounts (data not shown).

FIG. 4.

AS in the culture supernatant. A Western blot analysis of culture supernatant of pCF10- or pMSP7517-carrying cells was performed. Lanes: 1, surface extract of E. faecalis(pCF10), cCF10 induced; 2 to 7, concentrated supernatants [lane 2, E. faecalis(pMSP7517), nisin induced; lane 3, E. faecalis(pMSP7517), uninduced; lane 4, L. lactis(pMSP7517), nisin induced; lane 5, L. lactis(pMSP7517), uninduced; lane 6, S. gordonii(pMSP7517), nisin induced; lane 7, S. gordonii(pMSP7517), uninduced].

Phenotypes of cells expressing AS.

An increase in cell hydrophobicity due to expression of surface proteins is implicated in many adhesion processes in gram-positive bacteria, especially among oral streptococci (28). Therefore, we investigated the influence of expression of AS on the cell surfaces of E. faecalis and L. lactis. A hydrophobicity assay performed with hexadecane showed a strong increase in cell surface hydrophobicity in E. faecalis(pMSP7517) if the cells were induced by nisin (Table 2). A significant increase in hydrophobicity was seen in nisin-induced, AS-expressing L. lactis(pMSP7517). Pheromone cCF10-induced L. lactis(pCF10) cells also showed an increase in hydrophobicity.

TABLE 2.

Cell hydrophobicity and fibrin adhesion of E. faecalis(pMSP7517) and L. lactis(pMSP7517) strains in comparison to wild-type OG1SSp(pCF10)

Bacterium and strain	Mean hydrophobicity ± SD (%)	Mean level of adhesion to fibrin ± SD (% of input cells)
E. faecalis
OG1SSp(pMSP7517)
Uninduced	48.6 ± 11.3	4.4 ± 2.4
Nisin induced (25 ng/ml)	91.9 ± 2.7	13.7 ± 7.1
OG1SSp(pCF10)
Uninduced	56.6 ± 12.7	6.5 ± 1.4
cCF10 induced (10 ng/ml)	97.3 ± 1.7	6.9 ± 0.58
L. lactis NZ9800(pMSP7517)
Uninduced	41.9 ± 3.4	4.6 ± 3.2
Nisin induced (25 ng/ml)	66.7 ± 4.5	10.9 ± 1.0

AS was shown previously to contribute to the severity of enterococcal infective endocarditis (4, 44). Adhesion to fibrin, a component of endocardial vegetations, could be one of the mechanisms by which AS increases virulence in this model system. A fibrin adhesion assay (Table 2) demonstrated that AS-expressing cells of E. faecalis adhere to the fibrin matrix two to three times better than non-AS-expressing cells. The assay with E. faecalis cells is somewhat problematic due to the high level of aggregate formation. Binding of the cells to each other instead of directly to the fibrin matrix may lead to an overestimation of the actual number of cells interacting with fibrin. The expression of AS in L. lactis, which shows no aggregation, allowed us to assess the actual contribution of AS to adherence of AS-expressing cells to fibrin without interference from cell self-aggregation. Fibrin adhesion of nisin-induced L. lactis(pMSP7517) cells is also increased around twofold in comparison to that of uninduced cells, demonstrating the ability of AS to bind fibrin.

The results obtained by electron microscopy as well as data obtained in the cell hydrophobicity and fibrin binding assays suggest that AS is functionally expressed on the cell surface if expressed from pMSP7517. To investigate whether the expressed AS is also functional in adherence to E. faecalis binding substance, an assay making use of the conjugative transposon Tn925 in the E. faecalis donor strain INY1010 was employed. An 8-h coincubation with nisin-induced or uninduced recipient cells carrying pMSP7517 resulted in transconjugants with a copy of the transposed Tn925 appearing in the recipient cells. Results for mating of INY1010 with E. faecalis or L. lactis carrying pMSP7517 are depicted in Table 3. Transfer of Tn925 was increased in both species by about 100-fold if nisin was added to induce AS expression, providing a tight cell contact between the mating partners. A mating between INY1010 and S. gordonii was also performed. However, the number of E. faecalis donor cells was reduced by 100-fold after 8 h in comparison to the numbers obtained in the matings with the other species, suggesting that an inhibitory product (potentially a bacteriocin) was produced by S. gordonii. When matings were performed at a donor-to-recipient ratio of 10:1, transconjugants were recovered only with nisin induction. Their numbers were very low (two per milliliter of culture medium), apparently approaching the limit of sensitivity for this assay. It is also noteworthy that visible cell aggregates with the heterologous host could be observed only in induced INY1010-L. lactis(pMSP7517) cocultures.

TABLE 3.

Tn925 transfer from an enterococcal donor into AS-expressing cells

Mating pair		Mean level of Tn925 transfer ± SD
Donor	Recipient	Uninduced	Nisin induced (25 ng/ml)a
INY1010	OG1SSp(pMSP7517)	4 · 10−6 ± 4.1 · 10−6	4.5 · 10−4 ± 4.1 · 10−4
INY1010	NZ9800(pMSP7517)	5.5 · 10−8 ± 5.7 · 10−8	2.5 · 10−6 ± 1.6 · 10−6

^aCell aggregates were macroscopically visible. 

DISCUSSION

AS of the enterococcal sex pheromone plasmids is a large surface protein induced by peptide pheromones produced by plasmid-free recipient cells. The presence of two RGD sequences in the protein, as well as data from several previous studies, suggests the involvement of the protein in virulence of the nosocomial pathogen E. faecalis. The AS structural gene size (∼3.8 kb) and complex regulation provide sizable obstacles to a detailed characterization of the protein. So far, only deletion analysis of fairly large stretches of the protein has been accomplished (38). A further restriction for the analysis of AS, especially in terms of adherence to eucaryotic extracellular matrix proteins or cells, is the inherent function of the protein, the formation of tight aggregates between E. faecalis cells, thus potentially providing false-positive results in adherence assays.

An inducible system with an AS, readily manipulable in a defined background, is therefore highly desirable for use in our efforts to understand the structure and function of AS. We successfully adapted the nisin-inducible promoter system for this purpose. The AS gene was cloned behind the nisA promoter and placed in the shuttle vector pMSP3535. The resulting construct, pMSP7517, was introduced into E. faecalis, L. lactis, and S. gordonii. Induction with nisin resulted in the expected formation of E. faecalis cell aggregates, whereas the other species did not show any aggregation. This result confirms that AS is the only factor on the surface of the donor cells that is needed for aggregate formation, and it confirms the specificity of AS for the E. faecalis cell surface for recognition.

Immunoblot analysis of cells expressing AS showed that the previously described linearity of the nisA promoter (11) is only in effect in L. lactis over the tested range. AS expression in E. faecalis reaches a maximum at a nisin concentration of around 25 ng/ml and falls off at higher concentrations, whereas in S. gordonii the maximum is reached at 5 ng/ml and there is no significant increase or decrease at higher concentrations. It is also notable that the minimum concentration of nisin necessary for detectable AS expression ranges from 0.5 ng in S. gordonii and 1 ng in L. lactis to around 5 ng in E. faecalis. A slight amount of uninduced AS expression can also be observed in S. gordonii. The size of the produced protein is indistinguishable from that of AS isolated from cCF10-induced enterococcal cells carrying pCF10.

The correct display of the expressed AS was investigated by electron microscopy. S. gordonii surprisingly did not show any labeling, suggesting that AS is not expressed on the surface of that bacterium. E. faecalis and L. lactis were equally well labeled when AS was induced. These results led us to question whether AS is secreted and is not anchored correctly in the S. gordonii cell wall. However, AS was present in the culture supernatants of all three species upon nisin induction. This result showed that although AS was produced and obviously directed to the cell surface in S. gordonii, this apparently led to complete secretion in this species, presumably due to insufficient anchoring of AS in the cell wall. Secretion of the protein does not lead invariably to a loss of surface labeling with the AS antibody, as demonstrated by the presence of the protein in supernatants and on the cell surfaces of E. faecalis and L. lactis. The substantial amounts of AS present in the supernatants of these two species suggest that a limit to assembly into the cell wall may exist, with the excess protein being secreted.

Hydrophobicity can contribute to adherence and attachment of bacteria (28). Upon induction of E. faecalis(pCF10) with cCF10, the hydrophobicity of the cells increased considerably. AS is not the only surface protein induced by cCF10; Sec10, the surface exclusion protein, is also upregulated (29) and may contribute to the increased hydrophobicity. The expression of AS on cells carrying plasmid pMSP7517 alone shows that the protein contributes to the increase in the surface hydrophobicity of E. faecalis and L. lactis. This high-level hydrophobicity of the cells is, however, apparently not involved in aggregate formation, since L. lactis expressing AS does not show any aggregation unless E. faecalis is present in coculture.

Enterococci are commonly isolated in cases of infective endocarditis (36), and sex pheromone plasmids are present at an increased frequency in endocarditis-associated isolates (8). Adherence of bacteria to fibrin (1) can contribute to virulence in endocarditis. Our adherence assays demonstrate binding of AS-expressing bacteria to fibrin and show the advantage of expressing AS in a heterologous host. Aggregate formation by E. faecalis cells can interfere with the outcome of the assay. This effect is evidenced by the higher degree of variation in numbers of adhesive cells seen with E. faecalis(pMSP7517). The expression of AS from pMSP7517 in L. lactis, however, allows a clear demonstration of the fibrin-binding capability of AS, increasing the adherence by a factor of two. Similar values were found recently for adherence of S. gordonii CshA fibrils to fibronectin (35). AS-mediated binding of E. faecalis cells to fibrin could therefore help this bacterium establish itself on an endocardial vegetation. Interestingly, the number of adherent OG1SSp(pCF10) cells did not significantly change with induction by cCF10. The number of adherent uninduced OG1SSp(pCF10) cells was larger than the number of adherent OG1SSp(pMSP7517) cells not expressing AS. This effect could be due to the surface exclusion protein Sec10, which is constitutively expressed but upregulated upon cCF10 induction (16). The Sec10 protein has so far been neglected in studies of virulence. This protein may form a coiled-coil structure (29, 50), and moderate similarity to a Streptococcus pyogenes fibrinogen-binding protein (22% identity, 36% similar residues [47]) suggests that it may also contribute to enterococcal virulence.

Further confirmation that AS is functionally expressed on E. faecalis and L. lactis cells carrying pMSP7517 came from the transfer of the conjugative transposon Tn925 into cells that had not or had been induced with nisin, respectively. Transfer of Tn925 into E. faecalis(pMSP7517) as well as L. lactis (pMSP7517) increased around 100-fold upon nisin induction. This clearly demonstrates that AS is functionally expressed and displayed in both its homologous host, E. faecalis, and the heterologous bacterium L. lactis. The correct folding of the protein for the recognition of its cognate binding substance suggests that the overall structures of AS in E. faecalis and L. lactis are identical. Bactericidal activity of S. gordonii toward the E. faecalis Tn925 donor strain did not allow us to determine the function of AS in this species.

The expression of gram-positive cell surface adhesin in heterologous hosts has become a valuable tool, especially for surface proteins of oral streptococci, with E. faecalis and L. lactis serving as host species (26, 35). Our results confirm that AS is the sole factor required for aggregation and that it is specific for enterococcal binding substance. In addition, AS was successfully expressed on the L. lactis cell surface, increasing the cell surface hydrophobicity and enhancing adherence to fibrin. Both of these phenotypes strongly suggest a contribution of AS proteins to the virulence of E. faecalis.