The Bfp60 surface adhesin is an extracellular matrix and plasminogen protein interacting in Bacteroides fragilis

Abstract

Plasminogen (Plg) is a highly abundant protein found in the plasma component of blood and is necessary for the degradation of fibrin, collagen, and other structural components of tissues. This fibrinolytic system is utilized by several pathogenic species of bacteria to manipulate the host plasminogen system and facilitate invasion of tissues during infection by modifying the activation of this process through the binding of Plg at their surface. Bacteroides fragilis is the most commonly isolated Gram-negative obligate anaerobe from human clinical infections, such as intra-abdominal abscesses and anaerobic bacteraemia. The ability of B. fragilis to convert plasminogen (Plg) into plasmin has been associated with an outer membrane protein named Bfp60. In this study, we characterized the function of Bfp60 protein in B. fragilis 638R by constructing the bfp60 defective strain and comparing its with that of the wild type regarding binding to laminin-1 (LMN-1) and activation of Plg into plasmin. Although the results showed in this study indicate that Bfp60 surface protein of B. fragilis is important for the recognition of LMN-1 and Plg activation, a significant slow activation of Plg into plasmin was observed in the mutant strain. For that reason, the possibility of another unidentified mechanism activating Plg is also present in B. fragilis can not be discarded. The results demonstrate that Bfp60 protein is responsible for the recognition of laminin and Plg-plasmin activation. Although the importance of this protein is still unclear in the pathogenicity of the species, it is accepted that since other pathogenic bacteria use this mechanism to disseminate through the extracellular matrix during the infection, it should also contribute to the virulence of B. fragilis.

INTRODUCTION

Plasminogen (Plg), an abundant glycoprotein found in the blood plasma, is a central component of the fibrinolytic system. It is converted to plasmin, a broad-spectrum serine protease, by the action of urokinase (uPA) or the tissue-type Plg activator (tPA). Since plasmin can degrade structural proteins and activate other proteolytic enzymes, it is highly important for the degradation of the tissue barriers and cell migration. The recruitment and activation of Plg must be well controlled. However, several pathogens can interfere with the host Plg-plasmin system by expressing Plg receptors or activators (Lähteenmäki et al., 1998 ; Lähteenmäki, Kuusela & Korhonen, 2001a; Lähteenmäki et al., 2005; Stie et al., 2009). The surface molecules bind to Plg and enhance its activation by tPA on the bacterial surface, providing proteolytic activity to a nonproteolytic bacterium with the help of a host-derived proteolytic system. Since metastatic tumor cells utilize Plg activation to penetrate the basal membrane (BM), the term bacterial metastasis was coined to describe the analogous mechanism that bacteria employ to cross this barrier (Läteenmäki et al., 2005). Furthermore, Plg activators efficiently enhance bacterial adherence to human extracellular matrix (ECM) and mouse BM (Lähteenmäki et al., 1998). Several human invasive pathogenic bacteria, such as, Neisseria meningitidis (Ullberg et al., 1992), Salmonella typhimurium(Lähteenmäki et al., 1995), Yersinia pestis (Kukkonen et al., 2001), Staphylococcus aureus (Mökänen et al., 2002) and Helicobacter pylori (Jönsson et al., 2004) are examples of human pathogens that interact with the host Plg system to enhance their pathogenicity.

The phylum Bacteroidetes is a predominant component of the gastrointestinal tract microbiota and by far the most-well studied is the Bacteroides fragilis species which have shown to prime T cells responses in animals models via the capsular polysaccharide PSA. In human stool samples this taxon represents a level of at least 0.1% in 16% of samples (over 1% abundance in 3%) (Huttenhower et al., 2012; Patrick et al., 2011) and is the anaerobic pathogen most frequently isolated from endogenous infections (Finegold & Wexler, 1996), especially from patients with intra-abdominal infections (Gibson et al., 1998) and bacteraemia (Brook & Frasier, 2000). Several virulence factors have been described for this microorganism, such as, proteases (Patrick et al., 1996), enterotoxin (Wu et al., 1998) and lipopolysaccharide (Pumbwe et al., 2006), but their roles in pathogenicity are still not well elucidated. Without a doubt, the most studied virulence factor in the species is the capsular polysaccharide complex (CPC), which can modulate its antigenicity by expressing at least eight distinct polysaccharides (Gibson et al., 1998; Krinos et al., 2001).

The ability of B. fragilis to strongly adhere to laminin-1 (LMN-1) (Ferreira et al., 2006) and to convert Plg into plasmin (Ferreira et al., 2009) has been associated with an outer membrane protein (OMP) of approximately 60 kDa, Bfp60. Bfp60 was previously identified by Sijbrandi and colleagues (2005; 2008) as cell-surface Plg binding protein. However, its involvement in the pathogenicity of B. fragilis remains unclear. The purpose of this study was, therefore, to characterize the function of Bfp60 protein as a cell surface in B. fragilis 638R by constructing the bfp60 defective strain and comparing the properties of the wild type and mutant strains to bind LMN-1 and activate Plg into plasmin.

MATERIALS AND METHODS

1-Bacterial strains and growth conditions

B. fragilis strains used in this study are listed in Table 1. Strains were routinely grown anaerobically in brain heart infusion broth supplemented (BHIS) with hemin (5 mg/mL) and L-cystein 0.5 g/L) (Jousiemies Somier et al., 2002). Escherichia coli strains were cultured in Luria-Bertani Broth (LB) or agar. Rifampicin_(20 μg/mL),_gentamicin (100 μg/mL) and erythromycin (10 μg/mL) were added to the media when required.

Table 1.

Strains and plasmids used in this study

Strain or plasmid	Phenotype and/or Genotypea	Reference or source
B. fragilis strains
638R	Clinical isolate Rif r	Privitera et al., 1979
Bf6m	638R Δbfp60R:: pFD516 ermr	This study
Bfp6c	638R bfp60 (Con) – constitutively activating the outer membrane protein Bfp60	This study
E. coli DH10B	Cloning host strain	Invitrogen
Plasmids
pGem T easy	Cloning vector, (Ampr)	Promega
pFD340	Bacteroides –E. coli expression shuttle vector, (Spr) Ermr	Smith et al., 1992
pFD540	Suicide vector, derivated from deletion of pBI143 in pFD288 (SpR) Ermr	Smith et al., 1995
ATCC 968 S. aureus	Positive for the plasminogen activation	Ferreira et al., 2009

^aErm^r – Erythromycin resistance; Rif^r – Rifamycin resistance; Sp^r – spectinomycin resistance; Amp^r – Ampicillin resistance. Parentheses indicate antibiotic resistance expression in E. coli.

2-Construction of bfp60 insertional mutant and constitutive expressing strains

All DNA modifications and manipulations were carried out according to standard protocols (Sambrook et al., 1989). To construct of a bfp60 insertion mutant a DNA fragment of 600 bp, corresponding to an internal fragment of the bfp60 gene of B. fragilis, was used. The following two oligonucleotide primers were designed based on the whole nucleotide sequence of the bfp60 gene (1700 bp) available in the GenBank (EMBL/GenBack nucleotide database accessing number: AJ786264.1; Sijbrandi et al., 2005). The sense and antisense oligonucleotide sequences are as follows: Bfp60F 5′ CTTTACTTATGGCATTGG 3′(Forward) and Bfp60R 5′ GTTGTCCGTTGTAGGC 3′ (Reverse). The DNA fragment was amplified by PCR using Taq polymerase (Invitrogen) in 50 μL containing 1x Taq polymerase buffer, 1 mM MgCl₂, 0.3 mM dNTP, 0.5 pmol of each primer and 2 μL of DNA (20 μg/μL). The following cycle program was used: 95°C for 1 min, 35 cycles at 94°C for 30 s, 40°C for 30 s and 72°C for 30 sec; and a final extension at 72°C for 10 min. The amplified fragment was cloned into the pGEM-T easy vector (Promega) according to manufacturer’s instructions. The plasmid was sequenced to confirm the cloned DNA fragment. The bfp60 fragment was excised from the pGEM-T easy vector with SphI and PstI enzymes and ligated into the suicide vector pFD516, previously digested with the same enzymes. The new construct was transformed by electroporation into E. coli DH10B. The suicide vector containing the 600 bp bfp60 fragment was mobilized into B. fragilis by the triparental filter mating protocol (Shoemaker et al., 1986). The transconjugants with insertional mutations were selected on BHIS agar plates containing 20 μg/mL rifampicin, 100 μg/mL gentamicin and 10 μg/mL erythromycin. The B. fragilis bfp60::pFD516 construct, Bfp6m, was confirmed by PCR using the following oligonucleotides: Bfp60Fmut 5′ CCAAAGCGATCCAGGAAATA 3′ and Bfp60mut TGCACCTGTCATAGCCTCAG 3′.

A plasmid constitutively expressing bfp60 was constructed by PCR amplification of a 1692 bp promoterless bfp60 gene containing 53 bp upstream of the ATG start codon using the following primers Bfp60XmaI 5′ ATGCCCGGGGAATAGACAAAATTTCTC 3′ and Bfp60SacI 5′ GAGCTCTCGCTTATTTTACGATGC 3′ containing recognition sites for XmaI and SacI enzymes, respectively. The promoterless bfp60 gene fragment was cloned into the expression vector pFD340 in the same orientation as the IS4351 promoter. The new construct pFD340bfp60 was mobilized in the B. fragilis 638R strain as previously described. Transconjugants were selected on BHIS containing 20 μg/mL rifampicin, 100 μg/mL gentamicin and 10 μg/mL erythromycin and the resulting strain was designated Bf6c.

3-Adhesion and inhibition assays

Laminin from Engelbreth-Holm Swarm tumor (LMN-1; Sigma) or human plasminogen (Plg; Sigma) were used throughout this study. LMN-1 and Plg were immobilized onto glass coverslips as described previously (Ferreira et al., 2009) and placed into 24-well culture plates. Briefly, to prepare the plates, 15 μg/mL or 10 μg/mL of Plg or LMN-1 were suspended in 0.01 M PBS and immobilized onto glass coverslips for 1 h at room temperature. Immediately, LMN-1 or Plg-coated coverslips were carefully washed with PBS containing 0.1% bovine serum albumin (BSA) (w/v) to remove unbound LMN-1, avoiding non-specific association, and blocked with PBS containing 2% BSA (PBSB) for 1h at room temperature. Coverslips incubated with 2% BSA alone were used as the control.

For the adhesion assay, 200 μL of a suspension of cells of B. fragilis 638R wild type, bfp60 mutant or constitutively expressing bfp60 strains (10⁹ CFU/mL in 0.1% PBSB), were placed into 24-well culture plates containing the coverslips covered with either LMN-1 or Plg.

For the inhibition assay, B. fragilis 638R was mixed with 10 μg/mL Bfp60r protein (Ferreira et al., 2009) or with 1:100 dilution of anti-Bfp60 polyclonal antibody 0.1 M PBSB for 1h at room temperature before adding the suspension to respective coverslips placed in the microplate wells. After the 1h incubation at room temperature, the coverslips were washed twice with 0.1 M PBS, and allowed to interact with the coverslips covered with LMN-1, for 1h at 37°C under anaerobic conditions. After the incubation, the coverslips were washed again with 0.1 M PBSB, fixed with methanol for 5 min, washed with 0.01 M PBS, and then stained with 0.1% crystal violet. Three independent experiments were carried out in triplicate and the association of the bacteria (ratio of adherence) to the coverslips was established by counting 10 random fields in each coverslip using an optical microscope (x1000 magnification). Statistical analysis was carried out using Student’s unpaired t test unless otherwise specified (Prism 5.0; Graphpad Ltd. La Jolla, CA). All results were considered significant at a 95% confidence level (p ≤ 0.05). Data are expressed in Figures as the mean ± standard deviation (SD).

4-Plasminogen activation assay

The indirect assay (Jönsson et al., 2004) was used to measure the conversion of Plg into plasmin to determine if Bfp60 and Bfp60 constitutive strains were capable of activating the Plg. The 638R wild type strain preincubated with 1:100 dilution of antibodies against the Bfp60 (638R + Abbfp60), as described above, was also included in the test. For the assay, bacterial suspensions were incubated with 4 μg of Plg in 200 μL 0.02 M HEPES buffer (pH 7.4) in 1.5 mL Eppendorf tubes. Then, 30 μL of the chromogenic substrate S-2251 (0.45 mM; Val-Leu-Lys-p-nitrolinine dihydrochloride; Chromogenix) was added and incubated at 37°C. At 1h intervals, a 200 μL sample was transferred to a 96 wells microplate (TPP®) and the colorimetric reaction was measured at 405 nm. Streptokinase (STK; 1 U- used according to the manufactures’ specifications), plasmin alone, the 638R wild type strain and ATCC 968 Staphylococcus aureus were used as positive controls. The Streptokinase was used as a control of the capacity of an enzyme convert Plg into plasmin; S. aureus ATCC 968 strain was included to show that bacteria could convert the Plg used in the test into plasmin; the purified plasmin was a control of the chromogenic substrate. Plg in HEPES buffer with the chromogenic substrate was used as negative controls. Statistical analysis was carried out using Student unpaired t test unless otherwise specified (Prism 5.0; Graphpad Ltd. La Jolla, CA). All results were considered significant at a 95% confidence level (p ≤ 0.05). Data are expressed in Figures as the mean± standard deviation (SD).

5-Antiserum production against recombinant protein

Rabbit polyclonal antibodies against purified Bfp60r protein (Ferreira et al., 2009) was produced in the Laboratory Célula B in the Biotechnology Center at the Universidade Federal do Rio Grande do Sul -UFRGS University. The Montaine 888 and Marcol 52 were used as adjuvants. One milligram of the Bfp60r protein (Bfp60r) was subcutaneously administered to rabbits. Five subsequently booster injections were given with 2-week intervals using the same protein preparation and amount as above. Non-immunized control rabbits were injected with PBS. Five days after the last dose the animals were bled and the sera were analyzed by dot blotting(Ferreira et al., 2009).

6-Western- blotting analysis

For the immunoblotting analysis, the purified Bfp60 (Ferreira et al., 2009) was applied to a standard sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) with a discontinuous bis-acrylamide gel (4% stacking; 12% separating) in Tris-glycine running buffer (Tris 3 g/L, glycine 72 g/L and SDS 5 g/L) (Laemmli, 1970) and transferred to a nitrocellulose membrane overnight at 4°C (30 V; 40 mA). The membrane was washed 3x with 0.01 M PBS and blocked with PBST (150 mM NaCl, 0.03% Tween 20 and 5% skim milk) overnight at 4°C. After 3 washings with PBS (0.3% Tween 20) the membrane was incubated with the primary antibody anti-Bfp60 (1:100 for 2 h at room temperature. The membrane was washed in 0.01 M PBS (0.3% Tween 20) 3x again and the membrane incubated with the secondary antibody anti-rabbit IgG conjugated with peroxidase (1:5000; Sigma) in 0.01 M PBS (0.3% Tween 20) for 1 h at room temperature. The membrane was then washed (3x in PBS) and the reactivity was developed with 50 mg DAB (3,3′ Diaminobenzidine), 30% H₂O₂ in 0.01 M PBS. To stop the reaction the membrane was washed with distilled water.

RESULTS

Laminin-1 adhesion assay: The adhesion of B. fragilis strains to immobilized LMN-1 is shown in Figure 1. The adhesion of the mutant strain bfp6m to LMN-1 was significantly lower than the parent strain (p<0.05; p-value: 0.0001). A residual basal level of adhesion was still observed. The Bfp6c also showed strong adherence to LMN-1, but its adhesion was not statistically different (p>0.05; p-value: 0.09) from the wild type strain.

Figure 1.

Adhesion assay of Bacteroides fragilis 638R strain and its derivatives to Laminin-1. 638R- wild strain; bfp60m- strain with the gene for the Bfp60 protein mutated; Bfp60c- strain expressing the Bfp60 protein constitutively. Coverslips with BSA (2%) were used as negative controls. All tests were performed in triplicate and data are expressed as the mean ± standard deviation (SD).

The Bfp60 protein expressed and purified in E. coli is shown in the SDS-PAGE (Figure 2A) and a Western blot (Figure 2B) shows the reactivity of the anti-Bfp60 serum against the Bfp60r protein. The antibodies produced in rabbit were polyclonal and a dot-blot and an ELISA assays were also used to confirm the recognition (data not shown). The antibodies against the Bfp60r protein were used for the inhibition assay to LMN-1 and in the activation of the Plg into plasmin.

Figure 2.

Recombinant protein analysis by SDS-PAGE coomassie blue-stained gel and Western blotting. (A)- SDS-PAGE analysis of Bfp60 protein. 1-Molecular weight standards; 2- Whole cell lysate proteins of the E. coli Rosetta cells containing the recombinant Bfp60 protein without the IPTG-induction; lane 3- whole cell lysate proteins of E. coli Rosetta cells containing the recombinant Bfp60 protein induced by IPTG; 4- Recombinant Bfp60 protein after purification with Ni-resin (Ferreira et al., 2009). (B) – Western blotting of the recombinant protein probed with anti-Bfp60 antibodies produced in rabbit. The sign (➔) indicates the band detected by the antibodies anti- Bfp60 of approximately 60 kDa.

Inhibition of laminin-1 adhesion: The inhibition assay showed that when 638R wild strain was pre-incubated with the antibodies against the Bfp60r protein, the adhesion to LMN-1 decreased considerably (p<0.05; p-value: 0.003) (Figure 3). The antibodies were titrated and the inhibition was dose-dependent, although after 1:100 no inhibitory effect was observed (data not shown). A decrease in the adhesion also occurred when the 638R wild type strain was pre-incubated with 10 μg/mL of the purified Bfp60r protein (p<0.05; p-value: 0.002). To check if the inhibition was dose dependent, lower (1.5 μg/mL, 2.5 μg/mL, 5 μg/mL and 10 μg/mL) and higher (15 μg/mL, 20 μg/mL and 25 μg/mL) concentrations of the Bfp60r protein were assayed (Figure 4). A plateau of inhibition was achieved after 10 μg justifying this the use of this concentration for further assays.

Figure 3.

Inhibition assay of Bacteroides fragilis 638R strain to Laminin-1. 638R - wild strain; 638R + Abbfp60 − wild strain incubated with the antibody against the outer membrane protein Bfp60; 638R + Bfp60r protein. − wild strain incubated with Bfp60 recombinant protein. BSA (2%) was used as negative control. All tests were performed in triplicate and data are expressed as the mean ± standard deviation (SD).

Figure 4.

Adherence assay of Bacteroides fragilis 638R strain to different concentrations of the recombinant protein Bfp60. All tests were performed in triplicate and data are expressed as the mean ± standard deviation (SD).

An indirect assay was performed to determine whether the mutant lost the ability to activate the Plg into plasmin. Figure 5 shows the conversion of Plg into plasmin, with the S-2251 substrate when the strains were incubated with Plg. Both the 638R strain and Bfp6c converted the Plg into plasmin at the levels similar to S. aureus used as positive control after 4 h incubation. No significant differences in Plg activation were observed between the parent and the Bfp60 constitutively expressing strains (p>0.05; p-value: 0.07). However, the bfp6m mutant and the parent strain incubated with anti-Bfp60 antibodies had a significant reduction in the formation of plasmin compared to 638R strain (p<0.05; p-value: 0.025). These findings confirm that Bfp60 is involved in Plg activation in B. fragilis

Figure 5.

Plasminogen activation assay of Bacteroides fragilis. Lane A- 638R- wild strain; bfp60m- strain with the gene for the Bfp60 protein mutated; Bfp60c- strain expressing the Bfp60 protein constitutively; lane B- 638R + Abbfp60 − wild strain incubated with the antibody against the outer membrane protein Bfp60; lane C- S. aureus (ATCC 968) and STK (Streptokinase) were used as positive controls for the assay and HEPES buffer with the S-2251 substrate was used as negative control. All tests were performed in triplicate.

A significant slow activation of Plg into plasmin was observed in the mutant strain compared to negative control after 4 h of incubation (p<0.05; p-value: 0.035). This suggests that another unidentified mechanism for activating Plg is also present in B. fragilis.

DISCUSSION

Membrane associated or secreted bacterial proteases are essential to many bacterial species during vertebrate infections. Bacterial proteases can target host proteins to provide essential nutrients for the pathogen, to disarm components of the host immune response and to remove obstructions to efficient dissemination within the host (Goguen et al., 1995; Travis et al., 1995). Some pathogens mobilize host-derived PLG to its outer membrane. In some cases, certain pathogens immobilize host-derived Plg to their outer membranes in order to overcome the absence of endogenous extracellular proteases necessary for efficient migration across tissue barriers (Coleman et al., 1995; Grab et al., 2005).

Plasminogen is a single-chain glycoprotein that is inactive until cleaved by Plg activators to form plasmin. The active enzyme consists of five Kringle domains, each with three disulfide bonds that contain the lysine binding sites and the catalytic domain (Toledo et al., 2011). Binding of Plg to a mammalian receptor, fibrin clot, or a bacterial cell, facilitates its activation to plasmin and makes the molecule less susceptible to inactivation by α antiplasmin (Boyle and Lottenberg, 1987; Kukkonen et al., 2001).

As a commensal microorganism, B. fragilis must be able to survive in the intestinal mucosa. Conversely, as a pathogen, it must be able to attach to the site of infection, evade the host defense and produce factors that will, most of the time, damage the host tissue. In fact, B. fragilis is frequently isolated from opportunistic infections, including peritonitis, soft tissue abscesses and bacteraemia, with an estimated mortality rate of 30% (Cheng et al., 2009). Once B. fragilis has the chance to adhere to the mucosa and cross the epithelial barrier it has a long way to go until it reaches the blood stream. For that, the bacteria must migrate through the tissue layers and ECM, which are considered to be a major challenge for most bacterial pathogens and their dissemination within the host body (Lähteenmäki et al., 2001). During peritonitis, intra-abdominal abscesses composed primarily of thick fibrin walls are formed from the conversion of fibrinogen, a key structural component of the blood coagulation system, to fibrin. The fibrin formation not only helps to contain the bleeding, but also helps to trap any bacteria in the clot, preventing their dissemination. On the other hand, soluble plasmin is formed to dissolve the clot, but its activity has an extremely short half-life in plasma due to rapid inactivation by α antiplasmin (Lahteenmäki, Kuusela and Korhonen, 2001b). The broad spectrum proteinase activity of plasmin can directly mediate breakdown of a variety of extracelullar matrix constituents of the ECM, including fibronectin, laminin and vitronectin. Cell-bound plasmin also activates other matrix-degrading proteinases, such as collagenases. Together, these proteolytic functions facilitate the migration of cells through extracellular matrices and basement membrane barriers (Lahteenmäki, Kuusela and Korhonen, 2001b). The relative importance of Plg in infectious disease is indicated by the surface associated Plg binding properties manifested by several human pathogens, including Gram-positive bacteria, such as Staphylococcus and Streptococcus (Coleman et al., 1999; Gebbia, Coleman and Benach, 2001), Yersinia pestis, Listeria monocytogenes, Pseudomonas aeruginosa and at least four pathogenic fungal species (Coleman et al., 1999; Sun et al., 2004; Stie, Bruni and Fox, 2009). Borrelia burgdorferi, for instance, is known to bind host Plg, which can be converted to plasmin and it is able to degrade fibronectin, penetrate the endothelium, and activate matrix metalloprotease-9 and 1 (Coleman et al., 1995; Coleman et al., 1999).Plg is required for efficient disseminat ion in ticks and laboratory animals, enhancing their transmission and dissemination (spirochemia) (Coleman et al., 1997). Invasive bacterial pathogens that immobilize plasmin/plasminogen on their surface localized are turned into a proteolytic active organism capable of degrading tissue barriers, such as, basement membranes and extracellular matrices (Hu et al., 1995; Klempner et al., 1996; Coleman et al., 1999). Bacterial proteases can also activate latent procollagenases or inactivate protease inhibitors of human plasma, and thus contribute to tissue damage and bacterial spread across tissue barriers (Lahteenmäki, Kuusela and Korhonen, 2001a).

Several proteins have been found to play a major role in microbial recruitment of Plg, including enolase, glyceraldehyde-3phosphate dehydrogenase (GAPDH) and phophoglycerate kinase (PGK). Enolases are cytolic metaloenzymes that catalyze the conversion of 2-phospho-D-glycerate to phosphoenolpyruvate (Pancholi, 2001). Despite the lack of classical protein sorting machinery or cell membrane anchoring moieties, enolases are expressed on the surface of a variety of eukaryotic cells where they can function as Plg receptors (Pancholi, 2001). Incredibly, enolases are also found on the surface of Gram-positive and Gram-negative bacteria, fungi and protozoa (Toledo et al., 2011) where they can similarly function as Plg receptors (Lähteenmäki, Edelman and Korhonen, 2005). The surface location of enolase in several types of Prokaryotic and Eukaryotic cells is intriguing, since this enzyme does not have known cell surface protein motifs, such as signal peptidase cleavage site, cell wall anchors or sequences, or membrane-spanning domains. Sijbrandi and colleagues (2005) discovered an α-enolase, named P46, in B. fragilis mainly located in the cytoplasm and partially associated with the inner membrane (IM). Under iron-restricted conditions, however, P46 is localized in the IM fraction. Although the authors observed that Plg-binding to B. fragilis did occur, it was not P46 dependent. Instead, a 60kDa protein was identified as a putative Plg binding protein and named Bfp60. In 2008, the same group of researchers (Sijbrandi et al., 2008) located, cloned and sequenced the Bfp60 and the sub-cellular localization of the protein determined that Bfp60 was in the outer membrane protein of B. fragilis, more specifically, it was expressed at the cell surface. Even though, the authors demonstrated that Bfp60 was sufficient and required for Plg binding to B. fragilis and could be found in other Bacteroides species, the role of this potential virulence factor in its pathogenicity is still unclear. Our previous work (Ferreira et al., 2009) demonstrated that the protein involved in the LMN-1 adhesion in B. fragilis correspond to Bfp60 and it was shown that recombinant protein could interfere with the adhesion to the LMN-1 by performing inhibition assays.

We studied the function of Bfp60 protein in B. fragilis by constructing and characterizing strains with defective expression of Bfp60 and comparing them with a parental strain when assayed for LMN-1 binding and Plg activation. Antibodies against the Bfp60r protein were also used to check if they would block both characteristics. In the adhesion assay of the strains, we proved that Bfp60 is responsible for the adhesion to LMN-1 since the mutant strain Bfp6m reduced considerably the adhesion to LMN. When the Bfp6c was assayed, the strain strongly adhered to LMN-1 again. Although we have proved that the Bfp60 is responsible for the recognition of the LMN-1, we expected that Bfp6m strain would completely loose the capacity of adhesion when compared to the wild strain, 638R, and negative control. The same happened in the inhibition assay when the wild strain, 638R, was pre-incubated with the antibody against the Bfp60 protein and the inhibitory effect of purified Bfp60. A basal adhesion to LMN-1 was still observed. Other structures apart from surface proteins can also play a role in the recognition of LMN-1 and PLG, such as fimbriae and outer membrane vesicles (OMV), respectively. In 1998, Kukkonen and colleagues identified two LMN-binding fimbriae, the type 1 in Salmonella enterica Serovar Typhimurium and type fimbriae G of E. coli, as Plg receptors. Concerning the OMV, recently, it was found out that the OMV released by B. burgdorferi have an enolase as Plg receptor, which can facilitate the proteolysis in the peribacterial environment and promote further degradation of matrix proteins (Toledo et al., 2011). OMV are released naturally by Gram-negative bacteria and they represent a considerable portion of the bacterial cell and have been considered a part of the stress response components (Beveridge, 1999). B. fragilis OMVs were also observed and were shown to posses several degradation enzymes (Patrick et al., 1996). Pumbwe and coworkers (2007) exposed B. fragilis strains to 0.15% of conjugated and non conjugated bile salts and this resulted in an overproduction of fimbriae-like appendices and OMVs, leading to an increase in the resistance to antimicrobial agents, co-aggregation, adhesion to intestinal epithelial cells and biofilm formation. During the preparation of our bacterial cells the presence for those structures were not checked, so it is not possible to insure if any of those structures are responsible for the basal levels of adherence still observed. When the Plg activation assay was performed again, we were expecting the Bfp6m to behave as the negative control, but after 2h assay a slight increase in activity was observed, becoming higher with 4h incubation. The pre-treatment of 638R with anti-Bfp60 presented the same result. The Bfp6c strain activated as much as the wild strain.

The Plg-plasmin system has been exploited by bacteria to facilitate infection (Gladysheva et al., 2003). The action of bacterial Plg activators on the host system can cause the degradation of ECM and facilitate the spread and invasion of bacteria. In addition to Plg activators, bacteria also produce plasmin-binding molecules that have been implicated in the pathogenesis of infection. Altogether the data presented here through the generation of mutants for the Bfp60 surface protein of B. fragilis show that this molecule is important for the recognition of laminin-1 and Plg activation. The results represent not only an important contribution to the understanding of B. fragilis pathogenesis and interaction with the host, but also provide further insights into the mechanism of action of Plg activators that may result in new therapeutic agents targeted to infections caused by the species.