More than One Tandem Repeat Domain of the Extracellular Adherence Protein of Staphylococcus aureus Is Required for Aggregation, Adherence, and Host Cell Invasion but Not for Leukocyte Activation

Abstract

The extracellular adherence protein (Eap) is a multifunctional Staphylococcus aureus protein and broad-spectrum adhesin for several host matrix and plasma proteins. We investigated the interactions of full-length Eap and five recombinant tandem repeat domains with host proteins by use of surface plasmon resonance (BIAcore) and ligand overlay assays. In addition, agglutination and host cell interaction, namely, adherence, invasion, and stimulation of proliferation, were determined. With plasmon resonance, the interaction of full-length Eap isoforms (from strains Newman and Wood 46) with fibrinogen, fibronectin, vitronectin, and thrombospondin-1 was found to be specific but with different affinities for the ligands tested. In the ligand overlay assay, the interactions of five single tandem repeat domains (D1 to D5) of Eap-7 (from strain CI-7) with fibronectin, fibrinogen, vitronectin, thrombospondin-1, and collagen I differed substantially. Most prominently, D3 bound most strongly to fibronectin and fibrinogen. Full-length Eap, but none of the single tandem repeat domains, agglutinated S. aureus and enhanced adherence to and invasion of host cells by S. aureus. Constructs D3-4 and D1-3 (in cis) increased adherence and invasiveness compared to what was seen for single Eap tandem repeat domains. By contrast, single Eap tandem repeat domains and full-length Eap similarly modulated the proliferation of peripheral blood mononuclear cells (PBMCs): low concentrations stimulated, whereas high concentrations inhibited, proliferation. Taken together, the data indicate that Eap tandem repeat domains appear to have distinct characteristics for the binding of soluble ligands, despite a high degree of sequence similarity. In addition, more than one Eap tandem repeat domain is required for S. aureus agglutination, adherence, and cellular invasion but not for the stimulation of PBMC proliferation.

Staphylococcus aureus continues to be a major human pathogen responsible for superficial skin infections as well as for serious invasive infections, such as endocarditis, osteomyelitis, and sepsis (24). An important step in the initiation of invasive staphylococcal disease is adherence to host tissues and plasma proteins. S. aureus may interact with adhesive surface sites consisting of exposed or immobilized extracellular matrix (ECM) proteins such as fibronectin (Fn), fibrinogen (Fg), vitronectin (Vn), thrombospondin-1 (Tsp-1), collagen (Cn), bone sialoprotein, elastin, and several other proteins (30). Evidence from in vitro and ex vivo studies has suggested a role for these interactions in clinical disease (13, 23, 27, 29, 33, 34).

S. aureus adhesins are either anchored to the cell wall via an LPXTG motif or bind to the surface after their secretion by noncovalent interactions. S. aureus adhesins such as Fn-binding proteins (FnBPs), Cn adhesin (Cna), and clumping factor A (ClfA) have been well characterized on the biochemical and molecular levels (1, 4, 5, 26, 31). A number of secreted adhesins (secreted expanded-repertoire adhesive molecules, or SERAMs), such as the extracellular adherence protein (Eap) (15-17), the extracellular matrix-binding protein (Emp), and the extracellular fibrinogen-binding protein (Efb), play important roles in the establishment of disease. Eap shows a broad binding spectrum, and at least seven plasma proteins have been found to bind to Eap (28).

Not only does the specific bacterial interaction with these adhesive proteins allow for adhesion and colonization of tissues, but interaction with Fn is also pivotal for the invasion of nonprofessional phagocytes such as epithelial or endothelial cells by S. aureus (36-38). The importance of Eap in the adherence of S. aureus to eukaryotic cells has been demonstrated by a decreased adherence of an eap mutant to both fibroblasts and epithelial cells. The addition of exogenous Eap increases the adherence of both the wild type and an eap-negative mutant to fibroblasts. Anti-Eap antibodies significantly decrease adherence to and invasion of epithelial cells and fibroblasts (9, 17, 21). This scenario is important if FnBPs, which are major S. aureus invasins, are absent, as in strain Newman, or produced only at a low level.

Binding of Eap to intercellular adhesion molecule 1 (ICAM-1) inhibits leukocyte adhesion to endothelial cells and as a result prevents leukocyte extravasation (3). Eap also inhibits neutrophil recruitment during peritonitis, suggesting that Eap may function as an anti-inflammatory agent (3, 8). Therefore, the effects of the blockage of ICAM-1 interactions with leukocytes and the inhibition of neutrophil recruitment and the resulting dampening of the immune response may be important factors that determine the outcome of S. aureus infection. Hence, Eap is a critical factor in S. aureus adhesion and for the development of infection. The adherence of S. aureus to matrix supramolecular structures via Eap can be supported by inflammatory reactions (12). The binding of Eap to extracellular matrix ligands is promiscuous at the molecular level but not indiscriminate with respect to supramolecular structures containing the same macromolecules (12).

It has been suggested that the combined anti-inflammatory and antiangiogenic properties of Eap not only render this bacterial protein into an important virulence factor during S. aureus infection but also open new perspectives for therapeutic applications in pathological neovascularization (40). Eap has been shown to block metastasis formation in vivo, and Eap-derived agents may represent an attractive novel treatment for the prevention of breast cancer bone metastasis (35). Furthermore, Eap has been indicated as an attractive treatment for autoimmune neuroinflammatory disorders such as multiple sclerosis (41).

PCR analysis detected the eap gene in all 597 of the S. aureus isolates tested but not in S. epidermidis isolates or other gram-positive cocci (n = 216) (15). Based on PCR amplification, three different analogous genes of 1.8, 2.0, and 2.4 kb have been identified, consisting of five, six, and seven tandem repeat domains of 93 to 110 amino acids (16). However, on the protein level, only two analogues, of 65 and 72 kDa, have been identified, and these consist of five and six tandem repeat domains, respectively. Eap analogues consisting of six tandem repeat domains from strain FDA574 (20) and strain Wood 46 have also been designated as Map and p70, respectively (19). Recently, three further homologues have been described in strain Mu50, comprising four Eap tandem repeat domains (50 kDa) and one Eap tandem repeat domain only. Based on this homologue, the three-dimensional crystal structures of three different Eap domains have been resolved. Eap domains showed homology with the C-terminal domains of bacterial superantigen staphylococcal enterotoxin C. Examination of the crystal structure of the superantigen staphylococcal enterotoxin C bound to the T-cell-receptor β-chain suggests a potential ligand binding site within Eap (6). However, Eap does not block major histocompatibility complex-T-cell-receptor interactions and is not a superantigen. Instead, it has nonspecific cross-linking activity that is dependent upon having at least two of its six 110-amino-acid tandem repeat domains (25). The structure of Eap in solution has been revealed recently and has shown that Eap adopts an elongated conformation in aqueous solution (11).

The purpose of this study was to investigate the binding of Eap and individual tandem repeat domains to Fn, Fg, Vn, Tsp-1, and Cn type I (Cn I) by use of surface plasmon resonance (SPR) and ligand overlay assays. A further aim was to compare full-length Eap and individual tandem repeat domains with regard to known functions, such as the agglutination of staphylococci and interactions with host cells.

MATERIALS AND METHODS

Bacterial strains and media.

S. aureus strains include Newman (kindly provided by T. Foster, Dublin, Ireland), clinical isolate 7 (CI-7) (16), and Wood 46 (ATCC 10832). Escherichia coli TG1 was used to express recombinant Eap (rEap) and tandem repeat domains of Eap. For the cultivation of staphylococci, tryptic soy broth or agar (Difco, Detroit, MI), brain heart infusion (BHI) broth or agar (Merck, Darmstadt, Germany), Mueller-Hinton broth or agar (Mast, Merseyside, United Kingdom), and LB broth or agar (Difco) were used, as appropriate. For the cultivation of E. coli, LB broth or agar was used.

Solubilization of staphylococcal cell surface proteins.

To prepare cell surface proteins, staphylococci were grown in 5 ml BHI broth at 37°C for 18 h and then centrifuged at 10,000 × g for 2 min. The pellet was resuspended in extraction buffer (125 mM Tris-HCl, pH 7.0, plus 2% sodium dodecyl sulfate [SDS; Sigma-Aldrich Chemie GmbH, Deisenhofen, Germany]), heated at 95°C for 3 min, and then centrifuged at 10,000 × g for 3 min. The supernatant was passed through a Nap-10 column (Amersham Pharmacia Biotech Europe GmbH, Freiburg, Germany) containing Sephadex G-25 to remove SDS. The eluate was stored at −20°C.

SDS-PAGE and ligand overlay analysis.

To 20 μl of cell surface extract, 5 μl of 5× sample buffer (60 mM Tris-HCl, pH 6.8, 25% glycerol, 2% SDS, 14.4 mM 2-mercaptoethanol, and 0.1% bromophenol blue [Merck]) was added, and the mixture was heated at 95°C for 3 min and then separated in an SDS-polyacrylamide gel electrophoresis (PAGE) minigel. For Western ligand blot analysis, proteins separated by SDS-PAGE were electrophoretically transferred (Trans-blot SD; Bio-Rad, Munich, Germany) onto a nitrocellulose membrane (Schleicher & Schüll, Dassel, Germany), and then the membrane was blocked with 3% bovine serum albumin (BSA) (fraction V; Sigma). For probing blocked blots, either Fn (Chemicon, Temecula, CA), Fg (Calbiochem, San Diego, CA), Cn I (product 7774; Sigma), or Vn purified by the method of Yatohgo et al. (42) and Tsp-1 (gift from Beate E. Kehrel, Department of Anaesthesiology and Intensive Care, Experimental and Clinical Haemostasis, University Hospital of Münster, Münster, Germany) were used. Fn, Fg, Cn, Vn, and Tsp-1 were labeled with biotin according to the instructions of the supplier (Roche, Mannheim, Germany). Blotted proteins on nitrocellulose were exposed with biotinylated ligands and subsequently detected using an avidin-alkaline phosphatase color reaction (Bio-Rad). Alternatively, Fn, Vn, Fg Tsp-1, and Cn I were labeled with DIG (digoxigenin-3-O-methyl-carbonyl-ɛ-aminocaproic acid-N-hydroxysuccinimide ester; Roche), blotted S. aureus surface proteins were incubated with DIG-labeled ligands, and blots were subsequently exposed to anti-DIG antibodies (Roche) and developed with a color reaction (Roche).

Cloning, expression, and purification of recombinant proteins and rEap tandem repeat domains.

Eap-N, Eap-7, and Eap-W (originating from strains Newman, CI-7, and Wood, respectively), all lacking the signal peptide, were expressed and purified as described earlier (16). Five domains of eap (eap1 to -5) of S. aureus CI-7 were amplified by PCR with a set of primers (Table 1). The PCR products were ligated into plasmid pQ30-UA (Qiagen, Hilden, Germany). The ligation mixture was transformed into freshly prepared competent cells of E. coli TG1 and the transformation mixture was plated on LB plates containing ampicillin. Representative plasmids containing the eap fragments were designated as pQeap1 to -5. Six-His-tagged rEap fusion proteins were expressed and purified according to the protocol provided by the manufacturer (Qiagen). The expression of His-tagged rEap by use of vector pQE30 in E. coli M15 allows single-step purification using Ni-nitrilotriacetic acid (NTA) affinity resin. E. coli strains containing the above-mentioned plasmids were grown in Luria broth containing ampicillin (100 μg/ml) overnight with shaking at 37°C. One liter of LB medium with ampicillin was inoculated with 50 ml of overnight culture. The culture was grown at 37°C to an optical density at 600 nm of 0.5. Isopropyl-β-d-thiogalactopyranoside (IPTG) (final concentration, 1 mM) was added, and the culture was incubated for 4 h at 37°C with shaking. The culture was centrifuged and the pellet was resuspended in lysis buffer (50 mM NaH₂PO₄, 300 mM NaCl, and 10 mM imidazole), pH 8, and then 1 mg/ml lysozyme was added to lyse the cell wall. After cold incubation for 1 h, RNase at 4 U/ml and DNase at 24 U/ml in 1 mM MgCl₂ were added, and the mixture was incubated for 30 min. The bacterial lysate (after centrifugation to remove cellular debris) was run through Ni-NTA resin (Qiagen) to purify the six-His-tagged proteins. Bound proteins were eluted with elution buffer (50 mM NaH₂PO₄, 300 mM NaCl, and 300 mM imidazole), pH 8. The eluted proteins were dialyzed against phosphate-buffered saline (PBS) overnight and analyzed by SDS-PAGE.

TABLE 1.

Primers used in this study to generate rEap tandem repeat domains

Tandem repeat domain	Primer: position in eap7a	Sequence	No. of amino acids/molecular mass (Da)
D1	F1: 133-159	GGA TAT TCT AAA ATA CAG ATT CCA TAT	116/13,017
	R1: 460-480	TTG AAC ATT TGC TTT TGC CTC
D2	F2: 469-494	GCA AAT GTT CAA GTG CCG TAT ACA AT	110/12,188
	R2: 774-798	TTT TGC TTC TTT ATC TTT CGC TTG C
D3	F3: 798-822	GTA AATAAT CAA GTG CCA TAT TCA	105/11,912
	R3: 1087-1114	TTT TAC TTT CGA AAC TGT TTT TAC AGT
D4	F4: 1114-1137	GCG GAG CGT TAT GTA CCA TAT ACA A	108/12,243
	R4: 1411-1437	AAG CGC TTT ATT AGT TTT AGT GTG TTG
D5	F5: 1458-1482	ACT AAA GTG AAG TTT CCA GTA ACG	98/11,031
	R5: 1732-1755	TTT AAA TTT AAT TTC AAT GTC TAC

^aEMBL accession number AJ243790.

Biologic interaction analysis.

SPR measurements were performed using the BIAcore2000 instrument from BIAcore AB (Uppsala, Sweden). Sensor chip C1 (research grade), an amine coupling kit, surfactant P20, sample tubes, and caps were also obtained from BIAcore AB. The immobilization of proteins and an analysis of the interaction were carried out by an automatic method with BIAcore2000. The protein Eap was covalently coupled to sensor chip C1 via primary amine groups. After activation of the carboxylated matrix of sensor chip C1 with a single injection of 50 μl of 0.1 M N-hydroxysuccinimide-0.4 M N-ethyl-N′-(3-dimethylamino-propyl)-carbodiimide (NHS/EDC), 100 μl Eap (1:10 diluted in 10 mM sodium acetate buffer, pH 4.5) was injected over the activated surface. Excess activated esters were blocked by the injection of 55 μl of 1 M ethanolamine, pH 8.5. The immobilized amounts of Eap were in a range from 92 pg to 166 pg. Binding experiments were performed at 25°C in buffer of pH 7.4 containing 10 mM HEPES, 150 mM NaCl, 3.4 mM EDTA, and 0.005% surfactant P20. Sequential injection of Fn, Fg, Vn, Tsp-1, and Cn I allowed the determination of the respective kinetics of binding to Eap. The sensor chip was regenerated between each run with a pulse of 100 mM NaOH. The association and dissociation rate constants, k_on and k_off, respectively, were analyzed using BIAevaluation 3.1 software from BIAcore AB (Uppsala, Sweden).

Agglutination of bacteria by Eap or rEap tandem repeat domains.

S. aureus strain Newman was grown in LB broth overnight at 37°C. The bacteria were washed and suspended in PBS. A 40-μl bacterial suspension containing 2.3 × 10⁹ CFU/ml was placed on glass slides together with Eap or tandem repeat domains of Eap at final concentrations ranging from 0 μg/ml to 150 μg/ml. Agglutination was visible within 15 to 20 min at room temperature and was scored as follows: −, no agglutination; +, weak agglutination; and ++, strong agglutination.

Adherence and internalization of S. aureus strain Newman to fibroblasts and endothelial cells in the presence of Eap or rEap tandem repeat domains.

Fibroblasts (human fetal lung cells) were cultured in Dulbecco's modified Eagle's medium (Gibco-BRL) supplemented with 10% fetal calf serum (HyClone), HEPES buffer, α-glutamine, penicillin (100U/ml), sodium pyruvate, glucose, and pyridoxine. Human aortic endothelial cells (Clonetics, Walkersville, MD) were cultured using EBM-2 medium supplemented according to the instructions of the supplier (Clonetics). The cells were seeded (4 × 10⁴ cells/ml) in 24-well culture plates (Costar) and incubated at 37°C under 5% CO₂. The following standard procedure was followed. Upon reaching confluence, the cells were washed with the standard medium (Eagle's medium without supplements), and 900 μl of the standard medium was added to the cells. A mixture of 50 μl of strain Newman (2.3 × 10⁹ CFU/ml) and 50 μl of Eap tandem repeat domains or full-length Eap protein (final concentration, 30 μg/ml) was preincubated for 30 min at 37°C. Bacteria and PBS were used as the control. The mixture was then added to the cells in the wells and incubated for 2 h at 37°C. After incubation, wells were washed three times with PBS to remove nonadherent cells. A 200-μl volume of 10% trypsin was added to the wells to detach the cells, which were subsequently lysed by the addition of 800 μl of sterile water. Bacteria were then serially diluted and plated onto blood agar plates to determine viable counts. For the internalization assay, lysostaphin (final concentration, 20 μg/ml) was added for 20 min to kill extracellular bacteria before the trypsin step was performed.

Preparation of PBMCs and proliferation assay.

Peripheral blood mononuclear cells (PBMCs) were isolated from heparinized blood of healthy donors by Ficoll-Hypaque gradient centrifugation. The PBMCs were cultured in RPMI 1640 medium supplemented with 25 mM HEPES, 4 mM l-glutamine, 100 U/ml penicillin-streptomycin, and 5% heat-inactivated fetal calf serum. PBMCs (2 × 10⁶ cells/ml) were cultured for 72 h, after which they were pulsed for 6 h with 1 μCi per well of [³H]thymidine (specific activity, 5.0 Ci/mmol; Amersham Pharmacia Biotech, United Kingdom). Phytohemagglutinin-L (Sigma, St. Louis, MO) was used as a positive control for stimulation at a concentration of 2 μg/ml. All samples were assayed in triplicate, and the data are presented as counts per minute. The experiments were performed twice using cells from different individuals. For stimulation of proliferation, PBMCs were cultured with Eap or Eap tandem repeat domains at concentrations ranging from 0 to 81 μg/ml (final concentration). After 72 h, cells were pulsed for 6 h with 1 μCi per well of [³H]thymidine (specific activity, 5.0 Ci/mmol).

Statistical method.

An unpaired two-sided Student t test was used, with a threshold of statistical significance assumed at P values of <0.05 and <0.01.

RESULTS

Interaction of rEap and rEap tandem repeat domains with extracellular matrix ligands. (i) rEap.

Coomassie blue-stained SDS-PAGE of rEap-N Eap-7, and Eap-W (originating from strains Newman, CI-7, and Wood 46, respectively) revealed proteins of sizes similar to those of to the Eap analogues extracted with 2% SDS from the same strains. In ligand overlay assays, the recombinant proteins also showed binding to biotin-labeled Fn, Fg, Vn, Tsp-1, and Cn I (data not shown). The specific interactions between rEap and either Fn, Fg, Vn, or Tsp-1 were evaluated by using SPR (BIAcore). For this purpose, rEap isolated from E. coli was attached to a C1 sensor chip, a solution containing either Fn, Fg, Vn, or Tsp-1 was perfused over the surface, and the interaction was analyzed (Fig. 1). All four ligands interacted with immobilized rEap-N; however, the affinity of ligand interaction varied greatly. For Fg, the mean K_d (dissociation constant) value was 173 nM. For Vn (assuming a mean molecular mass of the native [i.e., polymeric] Vn molecule of 1,200 kDa), the K_d value was 3.2 nM. The K_d value for Tsp-1 was 0.36 nM. rEap-W expressed in E. coli interacted with Fg, Tsp-1, and Vn in SPR experiments at K_d values of 2.36 nM, 0.504 nM, and 2.98 nM, respectively. There was a specific binding of Fn to immobilized Eap-N and Eap-W. However, a quantitative evaluation of this binding was not possible due to the fact that no available model in the BIAevaluation software fitted the obtained curves well enough. The K_d values suggest that Eap-W binds more strongly to Fg and Vn than to Eap-N. Eap-N and Eap-W did not show binding to BSA, used as a control protein.

FIG. 1.

Sensorgrams (SPR, BIAcore analysis) showing binding of Fn, Fg, Vn, and Tsp-1 to, along with dissociation from, Eap-N. Eap (160 pg) was immobilized on sensor chip C1 as described in Materials and Methods. Sensor chip surfaces were subsequently exposed to various concentrations of purified extracellular matrix proteins, and SPR was determined as described in the text. Mean values of the rate constants and of the equilibrium K_d are given below the respective sensorgrams. The following conditions and concentrations were used (concentrations correspond in order to the curves from the bottom to the top for each of the respective panels): Vn at 47.7, 49.5, 57.5, 66, 91.6, 100, 125, and 150 nM; Tsp-1 at 17.78, 22.22, 26.67, 35.5, 31.11, 44.44, 53.33, 71.11, and 88.89 nM; Fg at 590, 1,180, 1,770, 2,360, 2,950, 3,540, 4,130, and 5,020 nM; and Fn at 114, 227, 341, 454, 558, 681, 795, 908, 1,022, 1,135, 1,249, 1,362, and 1,476 nM.

(ii) Characterization of rEap tandem repeat domains.

Five tandem repeat domains of rEap of S. aureus strain CI-7 were expressed as six-His-tagged recombinant protein and each tandem repeat domain was purified in a single step on Ni-NTA resin. Coomassie blue-stained SDS-PAGE confirmed purity (Fig. 2) and the expected deduced sizes (Table 1) of each tandem repeat domain. The five tandem repeat domains also showed binding to biotin-labeled Fn, Fg, Vn, Tsp-1, and Cn I (Fig. 2). Tandem repeat domains applied in equal amounts on a gel (SDS-PAGE) stained alike with Coomassie blue. Tandem repeat domain 5 (D5) showed weaker binding with all tested ligands, whereas D3 showed strong binding with Fg, Tsp-1, and Cn I. D1 exhibited a strong binding of Fn and Vn but a weak binding of Fg, Tsp-1, and Cn I. Results of interaction of all five tandem repeat domains of Eap7 with Fn and Fg in SPR are summarized in Table 2. Tandem repeat domains D1 to D3 showed increasing binding to Fg, with K_d values of 61.9 nM, 27 nM, and 6.78 nM for D1, D2, and D3, respectively. Tandem repeat domains D4 and D5 did not show binding to Fg. The binding of D1, D2, and D3 to Fn was specific, but a quantitative evaluation of this binding was not possible due to the fact that no appropriate model in the BIAevaluation software was available, similar to what was the case for the full-length proteins. D4 did not bind to Fn, and Fn binding of D5 was inconclusive and thus was recorded as absent. However, D4 showed binding to biotinylated Fn and Fg, and the weak binding of D5 to Fn and Fg corresponded to what was found in ligand overlay assays.

FIG. 2.

Analysis of rEap tandem repeat domains. (A) Schematic representation of the domain constructs used in this study. (B) Constructs were expressed using vector pQE30UA and purification was achieved in a single step on Ni-NTA resin. Shown are Coomassie blue-stained SDS-PAGE and ligand overlay assay membranes as indicated. For ligand overlay assays, nitrocellulose membranes with blotted proteins were probed with biotinylated Fn, Fg, Vn, Tsp-1, and Cn I, and protein-protein interaction was detected with avidin in an enzymatic color reaction as detailed in Materials and Methods.

TABLE 2.

Summary of binding characteristics of Fn and Fg to the recombinant tandem repeat domains D1 to D5

Tandem repeat domain (immobilized amt [pg])	Binding to binding partner:		Net charge (% KR-EDb)
	Fna	Fg (Kd [nM])
D1 (40)	Yesa	Yes (61.9)	8 (7.4)
D2 (75)	Yesa	Yes (27)	12 (11.1)
D3 (219)	Yesa	Yes (6.78)	14 (13.6)
D4 (93)	No	No	11 (10)
D5 (91)	No	No	5 (5.2)

^aNo model available to fit the data. Fn showed the strongest binding to peptide D3. Fn at a concentration of 454 nM gave a signal of approximately 22 response units (RU) during binding to D1 (immobilized amount, 40 pg) and one of approximately 37 RU during binding to D2 (immobilized amount, 75 pg). However, Fn at a concentration of 477 nM gave a signal of approximately 280 RU during binding to D3 (immobilized amount, 219 pg).

^bKR, lysine and arginine; ED, glutamic acid and aspartic acid.

Effect of Eap and rEap tandem repeat domains on agglutination, adherence, and invasion of host cells and proliferation of human leukocytes.

In order to determine the minimum number of tandem repeat domains required for the interaction of staphylococci and host cells, we examined three properties known to be at least partially mediated by Eap, namely, staphylococcal agglutination and adherence to and invasion of host cells.

(i) Agglutination.

The ability of Eap to rebind to S. aureus has been shown to cause bacterial aggregation (28). To investigate the effect of individual tandem repeat domains on the agglutination of intact staphylococci, S. aureus strain Newman was incubated with Eap or rEap tandem repeat domains at various concentrations (0 to 150 μg/ml). E. coli was used as a negative control (data not shown). Control experiments performed with native Eap and rEap did not show differences between the two preparations (data not shown). The addition of Eap, but not the addition of individual Eap tandem repeat domains, promoted the agglutination of strain Newman. Agglutination by full-length Eap was observed starting from 20 μg/ml and was maximal at the highest concentration used (150 μg/ml). By contrast, no agglutination was visible for any of the single tandem repeat domains, regardless of the concentration used (up to 150 μg/ml). The two-domain constructs D1-2 (tandem repeat domains 1 and 2 in cis) and D3-4 (tandem repeat domains 3 and 4 in cis) resulted in a weak agglutination of strain Newman. The three-domain construct D1-3 (tandem repeat domains 1, 2, and 3 in cis) caused the agglutination of strain Newman (to an extent of approximately 70% of full-length Eap). Thus, two-domain constructs had a weak effect on agglutination, whereas the three-domain construct led to a stronger agglutination, which was still surpassed by that seen for full-length Eap.

(ii). Adherence and invasion.

Externally added Eap can enhance the binding and internalization of S. aureus strain Newman into fibroblasts and epithelial cells (9, 17). The role of Eap tandem repeat domains on fibroblast and endothelial cell adherence and invasion by S. aureus was examined. A confluent layer of human fibroblasts was inoculated with strain Newman after Eap or Eap tandem repeat domains had been added and incubated at 37°C for 2 h. As expected, Eap enhanced S. aureus binding to (P < 0.05) (Fig. 3) and invasion of (P < 0.05) (Fig. 4) human fibroblasts and human aortic endothelial cells significantly. The presence of the monomeric Eap tandem repeat domains had no effect on either the binding to or the invasion of fibroblasts by S. aureus. To determine if invasion is dependent on the number of tandem repeat domains, a confluent layer of human aortic endothelial cells was incubated with tandem repeat domain construct D1-3, consisting of the first three domains of Eap, native Eap, and rEap domain monomers, respectively. Invasion of S. aureus in the presence of D1-3 was increased significantly compared with what was seen for control wells (endothelial cells in the presence of strain Newman but without any exogenous protein added) (Fig. 4B) (P value of <0.05 versus controls).

FIG. 3.

Adherence of strain Newman to human fibroblasts and endothelial cells. (A) Confluent layers of fibroblasts were inoculated with strain Newman without the addition of protein (control wells) or with strain Newman supplemented with either Eap protein or rEap tandem repeat domains and incubated for 2 h. Bacteria were detached, and the viable count was determined. Results are expressed as means; error bars show standard deviations (n = 3; *, P value of <0.05 versus the control). (B) Microtiter plate wells were coated with endothelial cells and subsequently blocked with 1% BSA. Bacteria were grown in BHI broth overnight, washed with PBS, and added to microtiter wells in the presence of Eap constructs. After 1 h, wells were washed and antistaphylococcal antibody was added, followed by the addition of alkaline phosphatase-conjugated goat anti-rabbit antibody was added. Alkaline phosphatase color substrate was used, and optical density was determined at 405 nm.

FIG. 4.

Invasion of human fibroblasts and human endothelial cells by strain Newman. Confluent layers of fibroblasts (A) and endothelial cells (B) were inoculated with strain Newman without the addition of protein (control well) or with strain Newman supplemented with either Eap protein or Eap tandem repeat domains and incubated for 2 h. Wells were further incubated with lysostaphin to kill extracellular bacteria. Host cells were lysed, and the viable count was determined. Data are presented as mean CFU of three and two experiments in panels A and B, respectively. Error bars show standard deviations. P values versus the control were <0.05 (*) and <0.01 (**).

We tested the adherence of strain Newman to EA.hy 926 cells in the presence of single tandem repeat domains D1 to D5, D3-4, D1-3, and full-length Eap. The adherence of strain Newman to EA.hy 926 cells in the presence of single tandem repeat domains was only moderately enhanced (up to 10% of that in the presence of full-length Eap). By contrast, the presence of constructs D3-4 and D1-3 caused a substantial increase in adherence (up to 57% of that seen with full-length Eap) (Fig. 3B).

(iii) Proliferation of human PBMCs.

To assess the effect of Eap on human immune cells, PBMCs were cultured for 72 h in the presence of different concentrations of Eap or rEap tandem repeat domains (0 to 81 μg/ml). As in a previous study (10), Eap showed a stimulatory effect at concentrations of 0 to 9 μg/ml and an inhibitory effect at higher concentrations with regard to PBMC proliferation. Single tandem repeat domains D1 to D3 and D5 had a stimulatory effect at concentrations of 9 to 27 μg/ml (Fig. 5A). D1-3 (in cis) displayed a dose profile similar to that of full-length native Eap (Fig. 5B).

FIG. 5.

Effect of Eap full-length protein and Eap domains on PBMCs. PBMCs were stimulated with indicated concentrations of Eap or Eap tandem repeat domains for 72 h (A) and Eap or D1-3 in cis (B). PBMCs without the addition of any exogenous protein but culture medium served as negative controls. Data are presented as mean counts per minute of triplicate determinations. Single representative experiments out of two are presented in panels A and B.

DISCUSSION

In this study, we have shown that single tandem repeat domains of Eap have different characteristics in terms of binding to a number of soluble host ligands, despite a relatively high sequence similarity. Two different methods, namely, ligand overlay assays and SPR (BIAcore) analysis, yielded compatible results for most of the ligand-domain pairs tested. However, for D4 the interaction with Fn and Fg had to be interpreted as absent, whereas both ligands showed strong binding in ligand overlay assays. Conversely, D1 displayed only very weak binding of Fg in ligand overlay assays, whereas SPR indicated a moderate binding. Whereas the agglutination of staphylococci, as well as the adherence to and invasion of host cells, required more than one Eap tandem repeat domain, the modulation of PBMC proliferation was possible with single Eap tandem repeat domains.

Eap has a broad binding spectrum, and at least seven plasma proteins have been found to bind Eap (28). Results of the SPR study clearly indicate the specific nature of binding of Eap to most of the tested ligands. Eap did not show binding to BSA used as a control, and it did show different affinities for different ligands. Therefore, the previous view that Eap can interact with almost every host ligand, due to the net positive charge, is no longer supported in the light of these data.

eap genes from different strains revealed a high degree of overall similarity (74 to 96%) at the nucleotide level (16). S. aureus may be able to produce shorter or longer forms of Eap depending on a point mutation in an adenine-rich region of the eap gene [poly(A) stretch], which may cause premature translational termination. In one case, the PCR products of strains Newman D2C and Wood 46 were 2,056 and 2,364 bp, resulting in predicted mature proteins of 77 and 85.55 kDa, with six and seven tandem repeat domains, respectively. However, in another case, proteins of 65.5 and 74 kDa for Newman D2C and Wood 46 were observed instead, with five and six tandem repeat domains, respectively. The reason for this difference is a stop codon preceded by nine adenine bases in the second of these two cases (starting at nucleotide 1740 in Newman D2C and nucleotide 2049 in Wood 46). Such phase variation for Eap, which has been described for several strains (2), might enable S. aureus to differentially modulate the host immune system (16, 22).

The Eap structure has been solved with single-domain constructs (6). It has been shown that four-domain constructs of Eap adopt an elongated conformation in solution, a situation in which the single domains appear to be connected only by the linker region between the tandem repeat domains (11). Most of the identified Eap homologues comprise five or six tandem repeat domains of 93 to 110 amino acids. In Eap-7, D1 exhibited the highest alignment score to other tandem repeat domains, and D5 showed a lower alignment score to the other four tandem repeat domains. The intra-Eap alignment score between D1 and D4 in Eap-7 ranged between 45% and 73%. Surprisingly, D5 of Eap-7 did not show binding to Fn and Fg. This may be explained by the fact that D5 showed very little homology with the other four tandem repeat domains: the intra-Eap alignment scores of D5 with D1 to D4 were 29%, 31%, 23%, and 29%, respectively. This low homology explains the inability of D5 to bind to Fn and Fg in BIAcore studies and the weak reactivity in ligand overlay assays. D3 of Eap-7 reacted strongly with Fg in ligand overlay assays and also showed the strongest binding to Fn and Fg in BIAcore studies. These data suggest that the binding specificity for different soluble ligands is apparently not necessarily based on recognition by linear Eap epitopes, as presented by blotted Eap in ligand overlay assays, e.g. This difference might explain the divergent results for SPR and ligand overlay assays which were observed for a minority of the tandem repeat domains.

Eap is able to form oligomers, and these direct Eap-Eap interactions cause bacterial aggregation due to the surface association of Eap (28). In the present study, the full-length Eap caused the agglutination of bacteria, but none of the rEap tandem repeat domain monomers (D1 to D5) were able to enhance the agglutination of S. aureus.

S. aureus possesses the ability to adhere to and invade nonprofessional phagocytes (32, 36, 37, 39). Previous studies by us (17) and others (21) showed a role of Eap in the adherence of S. aureus to eukaryotic cells. In these studies, an eap-negative mutant adhered less well to both fibroblasts and epithelial cells. The addition of exogenous Eap increased the adherence both of the wild type and of the eap-negative mutant to fibroblasts (17). In the present study, full-length Eap was able to enhance the adherence of S. aureus to fibroblasts, but the rEap tandem repeat domain monomers (D1 to D5) were not, demonstrating the need for at least two tandem repeat domains of Eap for host cell interaction with S. aureus. We reported earlier that Eap plays an important role in the internalization of S. aureus strain Newman (9). The addition of exogenous Eap increased the internalization of both the parent strain and the mutant strain by fibroblasts, and the addition of antibodies against Eap blocked this effect. Strain Newman is defective in FnBPs (7), resulting in poor invasiveness. In order to dissect the role of Eap in defined functions, strain Newman may be the best choice for this reason. In addition, it is cna negative, and thus the role of Eap in adherence and invasion is easier to identify. The most straightforward approach would be to use heterologous expression, e.g., in S. carnosus. Unfortunately, this has not been successful to date for any group working in this area, since S. carnosus, unlike S. aureus, appears not to be able to bind Eap on its surface. Hence, the use of strain Newman in this context is probably one of the best options currently available. This approach has been chosen in most of studies dealing with Eap. Additionally, we have studied invasion for different clinical strains and found a great variation between strains. Two selected strains have been further tested. Both the weakly (L12) and the highly (U35) invasive strain could be enhanced by the addition of external Eap (9).

Full-length Eap was able to enhance the invasion of fibroblasts by S. aureus, but the rEap tandem repeat domain monomers (D1 to D5) were not. On the other hand, D1-3, consisting of the first three tandem repeat domains of Eap as one polypeptide (in cis), enhanced the invasion of endothelial cells by S. aureus strain Newman significantly. This is in accordance with the observation that Eap has a nonspecific cross-linking activity that is dependent upon having at least two of its six 110-amino-acid tandem repeat domains for blocking major histocompatibility complex-T-cell receptor interactions (25).

An Eap analogue designated p70 from S. aureus Wood 46 is capable of inducing a time- and dose-dependent increase in immunoglobulin M and immunoglobulin G synthesis in PBMCs (19). PBMCs stimulated with Eap display increased interleukin 4 (IL-4) synthesis (18). It is interesting to speculate that the activation of IL-4 by Eap can modulate the immune response to S. aureus infection by interfering with the interactions of activated T cells and major histocompatibility complex class II-bearing antigen-presenting cells. However, the effect of IL-4 on the interactions between activated T cells and antigen-presenting cells is most likely additional to the effect of the blockage of ICAM-1 interactions by Eap (3). We have shown that the proliferation of PBMCs had the same dose profile with single Eap tandem repeat domains as it did with the whole Eap protein; that is, at a low concentration, the proliferation of PBMCs was stimulated, whereas high concentrations inhibited PBMC proliferation.

For S. aureus aggregation, adherence, and invasion, multiple binding sites on the protein are required. On the other hand, monomeric tandem repeat domains of Eap had the same effect on PBMC proliferation as did full-length Eap. Eap has a structural homology with the C-terminal domains of bacterial superantigens (toxins such as toxic shock syndrome toxin 1, staphylococcal enterotoxin A, and staphylococcal enterotoxin B) from S. aureus and streptococcal pyrogenic exotoxin C from Streptococcus pyogenes (6). However, Eap does not act as a superantigen (25). Nevertheless, TSST-1 has an activity profile similar to the one we present here for Eap tandem repeat domains: at low concentrations, TSST-1, like Eap, stimulates PBMCs from normal subjects, and at high concentrations TSST-1 induces B-cell apoptosis (10, 14).

Taken together, our data indicate that at least some of the biologically diverse functions of Eap have different structural requirements. Thus, the superficially nondiscriminate appearance of Eap with regard to its broad spectrum of activities starts to give way to a more differentiated view of different functions. This is in accordance with another surprisingly specific recognition of monomeric versus aggregated Cn I, which we have shown earlier (12). In addition, it is tempting to speculate that for both TSST-1 and Eap, the effect elicited by them on PBMCs is due to the structural similarity that these proteins share with each other, although any other similarity between Eap and superantigens remains to be demonstrated.