Correspondence: Franklin D. Lowy, Department of Medicine and Pathology, College of Physicians and Surgeons, Columbia University, 630 West 168th Street, New York, NY 10032, USA. Tel.: +212 305 5787; Fax: +212 305 5794; e-mail: firstname.lastname@example.org
Staphylococcus epidermidis infections are common complications of prosthetic device implantation. SdrF, a surface protein, appears to play a critical role in the initial colonization step by adhering to type I collagen and Dacron™. The role of ionic interactions in S. epidermidis adherence to prosthetic material was examined. SdrF was cloned and expressed in Lactococcus lactis. The effect of pH, cation concentration, and detergents on adherence to different types of plastic surfaces was assessed by crystal violet staining and bacterial cell counting. SdrF, in contrast with controls and other S. epidermidis surface proteins, bound to hydrophobic materials such as polystyrene. Binding was an ionic interaction and was affected by surface charge of the plastic, pH, and cation concentration. Adherence of the SdrF construct was increased to positively charged plastics and was reduced by increasing concentrations of Ca2+ and Na+. Binding was optimal at pH 7.4. Kinetic studies demonstrated that the SdrF B domain as well as one of the B subdomains was sufficient to mediate binding. The SdrF construct also bound more avidly to Goretex™ than the lacotococcal control. SdrF is a multifunctional protein that contributes to prosthetic devices infections by ionic, as well as specific receptor–ligand interactions.
Infections are among the most common complications of prosthetic device implantation (Baddour et al., 2003; Gandelman et al., 2007; Wang et al., 2007). The capacity of bacteria to adhere to these devices through both specific and nonspecific interactions is a critical first step in the initiation of these infections (Broekhuizen et al., 2006; Tsapikouni et al., 2008; Otto, 2009). This problem is enhanced when the infection involves devices such as ventricular assist devices that are critical to patient survival (Rose et al., 2001). Infections involving these devices occur in 15–30% of patients and generally require either device removal or transplantation to affect a cure (Herrmann et al., 1997; Holman et al., 1997; Gordon et al., 2006) [INTERMACS (http://www.intermacs.org)].
Staphylococcus epidermidis remains the most common cause of prosthetic device-related infections (Simon et al., 2005; Gordon et al., 2006). As part of the commensal skin flora, staphylococci are uniquely situated to contaminate wounds when cutaneous barriers are breached. Surface proteins known as microbial surface components recognizing adhesive matrix molecules facilitate the initial colonization step (Patti et al., 1994; MacKintosh et al., 2006; Otto, 2009). SdrF, a S. epidermidis surface protein, appears to contribute to the initiation of prosthetic device infections. Previous studies showed that SdrF, a member of the serine–aspartate (SD) family of surface proteins, binds type I collagen and mediates adhesion of S. epidermidis to the ventricular assisted device (VAD) driveline (Bowden et al., 2005; Arrecubieta et al., 2007, 2009). Surprisingly, it was also found that SdrF also dramatically enhanced adherence to Dacron, the material that covers the driveline, presumably by ionic interactions (Arrecubieta et al., 2009). In light of the potential contribution of this ionic interaction to the initiation of infection, we further examined the nature of this process.
Materials and methods
Strains, plasmid, and growth conditions
Lactococcus lactis MG1363 (Wells, 1993) was grown at 30 °C in M17 media supplemented with 0.5% glucose. MG1363 strains containing the plasmid pOri23 (Que et al., 2000) expression vector were grown in media supplemented with erythromycin (5 μg mL−1). Escherichia coli XL-1 (Qiagen, CA) was grown at 37 °C in LB media. XL-1 containing his-tag expression plasmid pQE30 (Qiagen) were grown in media supplemented with Ampicillin (100 μg mL−1). Five different previously prepared E. coli constructs using the pQE30 expression plasmid were used in this study (Arrecubieta et al., 2007). These constructs expressed different components of the SdrF B domain including sdrFrB1-4, sdrFrB1, sdrFrB2, sdrFrB3, and sdrFrB4. Staphylococcus epidermidis strain 9491, a SdrF positive strain, was also used in this study (McCrea et al., 2000; Arrecubieta et al., 2007, 2009).
Recombinant strains, proteins, and antibodies
Staphylococcus epidermidis SdrF and subclones were cloned into the expression vector pOri23 and transformed into MG1363, as described (Arrecubieta et al., 2007, 2009). The same subclones were cloned into pQE30 his-tag expression system (Qiagen) and expressed from XL-1. These proteins were purified as previously described using His-trap columns (Pierce, IL; Arrecubieta et al., 2007). Purified proteins were biotinylated with EZ-Link NHS-LC-Biotin (Pierce). Polyclonal antibodies directed against the A and B domains of SdrF were used as previously described (Arrecubieta et al., 2007, 2009).
Bacterial adherence assays
Adherence assays were carried out in 96-well plates as previously described (Arrecubieta et al., 2007, 2009). Mid-log phase MG1363 cells were suspended in phosphate buffer saline (PBS) to a final OD600 nm = 0.1. Aliquots of 100 μL were added to the wells and incubated for 1 h at 37 °C. Wells were washed with PBS, and attached cells were stained with crystal violet for direct cell counting or were recovered by three sequential 5-min treatments with Trypsin/EDTA at 37 °C and then plated on GM17 agar for cell counts (Arrecubieta et al., 2009).
Three differently charged 96-well plastic plates were studied: Tissue Culture (TC), Primaria, and Polysterene (Becton Dickinson, NJ). A second type of prosthetic material frequently used in prosthetic devices, Goretex™, was also used in adherence assays. To further examine the nature of the ionic interaction, different environmental conditions were studied including pH (4.5, 7.2, and 9.5), cations (calcium, lithium, magnesium, sodium), and disruptive agents (Tween20; Sigma, St. Louis, MO) prepared in PBS (Sigma).
Each experiment was performed at least three times, and each time point was performed in triplicate. Data were analyzed using an unpaired Student's t-test. A value of P < 0.05 was considered to be statistically significant.
Protein surface interaction
Purified His-tagged proteins were biotinylated and attached to 96-well plates as previously described (Arrecubieta et al., 2007). Briefly, adherent biotin-labeled proteins were incubated with HRP-avidin (DakoCytomation, Glostrup, Denmark) for 30 min at 22 °C. After PBS washing, binding of the HRP-avidin was quantified by adding the substrate o-phenylenediamine dihydrochloride and measuring the resulting absorbance at 490 nm in a microplate reader (Bio-Rad, CA). Attachment assays were carried out in three different 96-well plate materials described above. All reagents were purchased from Sigma.
Adherence of L. lactis expressing SdrF to plastic
Adherence of L. lactis expressing the S. epidermidis surface protein SdrF to three different types of plastic was examined at three different initial bacterial concentrations. The Primaria plates are positively charged, while the polystyrene plates have a net neutral charge and the TC plates are negatively charged. The SdrF expressing constructs showed increased attachment to the three different plastic plates tested when compared with the lactococcal controls at the two higher initial bacterial inocula (ODs 0.5 and 1.0; P < 0.01; Fig. 1; P < 0.05). Attachment was highest to the Primaria™ plates, an increase of 70%, when compared with either the polystyrene or TC plates.
SdrF has two ligand-binding domains, the A and B domains. The B domain, composed of four structurally similar subdomains, mediates binding to collagen. Previous studies found that the B4 subdomain was sufficient to mediate this binding interaction (Arrecubieta et al., 2007). The SdrF B4 subdomain was also capable of mediating attachment to the Primaria™ plates, although adherence to the other two types of plastic was reduced when compared with SdrF (Fig. 2). Antibodies targeting the SdrF B domain, but not the A domain, reduced adherence of SdrF-lactis to the polystyrene plates (Fig. 3; P < 0.05) further suggesting that the SdrF interaction with plastic is through its B domain.
Binding to Goretex (polytetrafluoroethylene), a second hydrophobic material frequently used in prosthetic material, was also assessed. While the lactococcal expressing SdrF construct demonstrated enhanced binding to the material (P <0 .05), neither the isolated A or B domains of SdrF or a SdrF positive S. epidermidis, 9491, demonstrated enhanced binding when compared with the controls (Fig. 4).
Effect of cations, pH, and detergents on adherence
Increasing concentrations of the cations sodium, lithium, and calcium reduced the attachment of L. lactis expressing SdrF (Fig. 5a). Similar effects were observed when the B4 subunit of the SdrF was challenged with increasing cation concentration. Ca2+ and Na+ cations showed the largest effect on SdrF expressing clones to the polystyrene surface reducing adherence to plastic by 53% and 60%, respectively (Fig. 5b and c).
Effect of pH and detergents on adherence
Lactococcus lactis expressing SdrF bound to plastic most efficiently at pH 7.4 (Fig. 6a). Increasing concentrations of the detergents Tween20 and beta-D-octylglucoside reduced adherence of L. lactis expressing SdrF (Fig. 6b), the SdrF B domain, and SdrF B4 subdomain to polystyrene plastic (Fig. 6c and d). Beta-d-octylglucoside produced a greater effect than Tween20 with the SdrF B1,4 and SdrF B4 interaction with polystyrene (P < 0.05; Fig. 6c and d). The protein denaturing agents urea and guanidine chloride also affected the adherence of the SdrF B domain and its subdomain B4 to the polystyrene wells (Fig. 6c and d). Guanidine chloride caused a larger reduction in binding by the SdrF B domain and its subdomain B4 (P < 0.05).
Staphylococcus epidermidis is one of the primary pathogens responsible for prosthetic device infections (von Eiff et al., 2002). In a previous study, we utilized the lactococcal heterologous expression system to demonstrate that SdrF mediates bacterial adherence to the ventricular assist device extracutaneous Dacron covered drivelines.(Arrecubieta et al., 2009). This suggested that SdrF–Dacron surface interactions contributes to the initiation of prosthetic device infections. This study further explored the nature of this interaction.
Attachment assays to polystyrene showed that L. lactis strains expressing SdrF adhered better to polystyrene, especially to the Primaria™ plates, than did the plasmid controls. Both TC and Primaria™ plates are modified polystyrene plastic. In the case of TC plastic, the addition of COOH groups to the polystyrene polymer confers a net negative charge to the surface of the polymer. On the other hand, Primaria™ plates are modified by the incorporation of NH2 groups, which makes the plates positively charged. The higher attachment observed in the Primaria™ suggests that SdrF, a negatively charged molecule, preferentially binds the positively charged plate via ionic interactions.
Antibodies targeting the B, but not the A, domain showed a reduction in bacteria expressing SdrF attachment to polystyrene, suggesting that the interaction occurs via the negatively charged B domain and also that its subdomains are sufficient to mediate attachment (McCrea et al., 2000).
The cation concentration (ionic strength) of a solution also affects protein–surface interactions. Cations can interfere with the hydrostatic and electrostatic forces that operate in the adsorption of proteins to surfaces (Agnihotri & Siedlecki, 2004; Tsapikouni et al., 2008). Increasing concentrations of several ions (Ca2+, Li1+, Na1+, Mg2+) reduced the attachment bacteria expressing SdrF and the B domain and subunit to polystyrene. These results add further support to the observation that the attachment of SdrF to polystyrene is ionic and are perturbed by increasing concentrations of ions in the solution.
Calcium cations caused a greater reduction in attachment with a lower concentration than any of the other ions assayed. Sequence analysis of SdrF B domain revealed high sequence similarity with another staphylococcal surface protein, clumping factor A (ClfA; O'Connell et al., 1998). Like SdrF, ClfA is also a binding protein that, in the case of CflA, binds fibrinogen (O'Connell et al., 1998). ClfA–fibrinogen binding is localized to a region where the sequence resembles the Ca2+-binding EF-hand motif often found in eukaryotic binding proteins (D'Souza et al., 1990; O'Connell et al., 1998). In these proteins, Ca2+ interferes with protein–ligand interaction either by occupying the ligand-binding site or binding to another site and causing a conformational change in the protein that prohibits the binding of the ligand. The steep reduction observed with increased Ca2+ concentration suggests that SdrF–polystyrene ionic interaction may depend on the conformational state of the protein.
The pH value of the surrounding solution affects the properties of both, the polymer and the protein. Our results suggest that at values close to physiological pH, the interaction between SdrF and polystyrene surfaces was optimal. The pH affects the protonation of proteins and surfaces (Matsumoto et al., 2003). Preliminary predictions made using Protean (DNASTAR Lasergene 8) suggest that at physiological pH (7.4) SdrF has an overall negative charge (near-324.4) with the B domain concentrating most of that negative overall charge. These preliminary predictions might help explain the ionic nature of the SdrF–polystyrene interaction and its preference for slightly positively charge surfaces.
Detergents (i.e. Tween20 and beta-d-octylglucoside) and disruptive agents (i.e. urea and guanidine chloride) are also known to perturb protein–surface interactions, as these molecules denature or perturb the protein structure (Boks et al., 2008). Increasing concentrations of the nonionic surfactant Tween20 reduced the interaction between SdrF as well as the B domain constructs and the polystyrene surface. Both of these detergents are used in the pharmaceutical industry and contact lenses to avoid protein and microbial adsorption to the material (Santos et al., 2007) due to their amphiphilic properties.
The effect of guanidine chloride on SdrF B4-polystyrene interaction was higher than the effect of urea. Although still controversial, these two disruptive agents appear to denature proteins in different ways (Lim et al., 2009). While urea seems to create hydrogen bonds to the peptide group, guanidine chloride appears to disrupt the main backbone of the peptide (Lim et al., 2009). Guanidine chloride is usually more effective than urea when the peptide contains helices stabilized by planar residues (Lim et al., 2009). This indicates that the SdrF–polystyrene interaction depends on the tertiary structure of the peptide, specifically the SdrF B4 subdomain.
A limitation of the study is that we were unable to create S. epidermidis strains that were isogenic for SdrF. The availability of an isogenic pair would have added further information regarding the role of SdrF in these binding interactions.
SdrF appears to be a unique protein, capable of helping to initiate prosthetic device infections in different ways. It is present on the vast majority of S. epidermidis strains, can bind to Dacron or other prosthetic materials via ionic interactions and is also capable of adhering to matrix molecules such as collagen that coat internal portion of these devices via specific receptor–ligand interactions. Further investigation of this and other S. epidermidis surface proteins is warranted.
This work was supported in part by the National Heart, Lung and Blood Institute-Specialized Center for Clinically Oriented Research (grant HL 077096-01). Thoratec Corporation (Pleasanton, CA) kindly provided the Dacron™ material currently used on the exterior surface of the Heartmate VAD DLs. None of the authors have a conflict of interest with any of the material in this manuscript.