α-Enolase, a Novel Strong Plasmin(ogen) Binding Protein on the Surface of Pathogenic Streptococci*

The plasmin(ogen) binding property of group A streptococci is incriminated in tissue invasion processes. We have characterized a novel 45-kDa protein displaying strong plasmin(ogen) binding activity from the streptococcal surface. Based on its biochemical properties, we confirmed the identity of this protein as α-enolase, a key glycolytic enzyme. Dose-dependent α-enolase activity, immune electron microscopy of whole streptococci using specific antibodies, and the opsonic nature of polyclonal and monoclonal antibodies concluded the presence of this protein on the streptococcal surface. We, henceforth, termed the 45-kDa protein, SEN (streptococcal surface enolase). SEN is found ubiquitously on the surface of most streptococcal groups and serotypes and showed significantly greater plasmin(ogen) binding affinity compared with previously reported streptococcal plasminogen binding proteins. Both the C-terminal lysine residue of SEN and a region N-terminal to it play a critical role in plasminogen binding. Results from competitive plasminogen binding inhibition assays and cross-linking studies with intact streptococci indicate that SEN contributes significantly to the overall streptococcal ability to bind plasmin(ogen). Our findings, showing both the protected protease activity of SEN-bound plasmin and SEN-specific immune responses, provide evidence for an important role of SEN in the disease process and post-streptococcal autoimmune diseases.

Streptococcus pyogenes is responsible for a wide variety of human diseases that range from suppurative infections of the throat (pharyngitis), skin (impetigo), and underlying tissues (necrotizing fasciitis), to an often fatal toxic shock syndrome, and the post-streptococcal sequelae, rheumatic fever, and acute glomerulonephritis. Bacterial surface proteins play a major role in these disease processes by exhibiting a wide range of functions. As data have become available, it is clear that most surface proteins found on Gram-positive bacteria, particularly those on group A streptococci, have a great deal of structural similarities (1,2). Proteins for which the function(s) has been defined have been found to be multifunctional, whereas in others a function has only been attributed to one of two or more domains (2,3). Thus, the multifunctional characteristics of these surface proteins increase the complexity of the Gram-positive surface beyond what has been previously imagined.
We recently described one such multifunctional protein, streptococcal surface dehydrogenase (SDH), 1 as a major surface protein on group A streptococci and other streptococcal groups which is structurally and functionally related to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (4). SDH also has an ADP-ribosylating activity (5) and exhibits multiple binding activities to several mammalian proteins such as fibronectin and cytoskeletal proteins (4). A structurally and enzymatically similar streptococcal protein, Plr, was also identified on group A streptococci, based on its ability to bind plasmin (6). SDH, however, is a weak plasmin-binding protein (4). During our studies characterizing the SDH molecule, we reported that a 45-kDa protein was also found in high amounts on the surface of group A streptococci (4). While determining the relative plasmin binding activity of SDH with respect to other streptococcal surface proteins, we found that the 45-kDa protein had in fact strong plasmin binding activity. 2 The plasmin(ogen) system displays a unique role in the host defense by dissolving fibrin clots and serving as an essential component to maintain homeostasis and vascular potency (7)(8)(9). Studies on the ability of Gram-positive bacteria to subvert the fibrinolytic activity of human plasmin(ogen) to their own advantage for tissue invasion have been largely focused on pathogenic streptococci and were described first as early as 1933 by Tillet and Garner (10). This property was subsequently attributed to the plasmin(ogen) activator, streptokinase (11), an extracellular 48-kDa protein secreted in culture supernatants (12). The role of pathogenic bacteria in tissue invasion utilizing this system has recently been reviewed (13).
In the present communication, we describe purification and characterization of the 45-kDa protein and show that it is the major plasmin(ogen) binding molecule on the surface of streptococci. We also show that this protein has significant sequence similarity with one of the important glycolytic enzymes, ␣-enolase, found generally in the cytoplasm. While bound on the surface of group A streptococci, this 45-kDa protein is found to retain its ␣-enolase activity, hence we named it SEN (streptococcal surface enolase). It is distinct from the 48-kDa streptokinase (12,14), the 35.8-kDa SDH (4), the 41-kDa Plr (6), or the 45-kDa plasminogen-binding protein, PAM (15), all of which have been reported to bind plasmin to varying degrees. ␣-Enolase has not been previously identified on the surface of bacteria; however, it has been shown to be expressed on the surface of neuronal (16), cancer (17), and some hematopoietic cells (18,19) as a novel plasmin(ogen) receptor. Here, in addition to the structural and functional characterization of SEN, we also describe the biological activity of SEN, the functional consequence of plasmin(ogen) binding to SEN, and the enzymatic activity of SEN-bound plasmin. . These strains were grown overnight in Todd-Hewitt broth (Difco) and washed once with 50 mM ammonium bicarbonate followed by two washes in 50 mM phosphate buffer, pH 6.1, to eliminate the presence of any soluble streptokinase, which may interfere with analysis. Type M6 (D471) streptococci (4) were used for the isolation of SEN, whereas the other strains were used to study the prevalence of the plasmin-binding 45-kDa related proteins in different streptococcal groups and group A serotypes.

Bacteria-Group
Human Plasminogen and Plasmin-Purified human plasminogen and plasmin (lysine-plasmin) were purchased commercially (Sigma). Plasmin was also generated from plasminogen by incubation with urokinase (20 units/ml, Sigma) in HBS gel buffer (50 mM HEPES/ NaOH, pH 7.4, containing 1 mM MgCl 2 , 0.15 mM CaCl 2 , and 0.1% of gelatin) containing 40 mM lysine. Conversion of plasminogen to plasmin was maximal after 1 h at 37°C. This method consistently converted more than 95% of the single chain zymogen molecule plasminogen to the heavy and the light chain of the plasmin molecule as reported earlier (20).
Radioiodination of Plasminogen and Plasmin-Purified plasminogen and plasmin were radioiodinated with Na 125 I (17 Ci mg Ϫ1 , NEN Life Science Products) by the chloramine-T method, using IODO-BEADs (Pierce) as described (4). The labeled proteins were separated from free iodine by passage over a G-25 column (PD-10, Amersham Pharmacia Biotech) and collection in HBS gel. The labeled proteins were stored at Ϫ70°C. Plasmin was also generated from the 125 Iradiolabeled plasminogen by incubation with urokinase (20 units ml Ϫ1 , Sigma) in HBS gel that contained 40 mM lysine (20). More than 95% of the radioactivity was found to be retained with the plasmin. Thus, the specific radioactivity of the labeled plasmin and plasminogen was found to be essentially the same. Furthermore, the specific radioactivity of commercially available purified plasmin (Sigma) and urokinase-generated plasmin was also the same. Typically, specific radioactivity of the 125 I-labeled plasmin/plasminogen was achieved in a range of 1.2-2.0 ϫ 10 6 cpm g Ϫ1 protein.
Blot Overlay System for Plasmin(ogen) Binding-Proteins in the bacterial cell wall extracts were resolved by 12% SDS-PAGE gels and blotted electrophoretically onto a PVDF membrane as described (21,22). Blots were incubated at room temperature for 3 h in a blocking HBST gel buffer (50 mM HEPES/NaOH, pH 7.4, containing 0.15 M NaCl, 1% acidified BSA, 0.5% gelatin, 0.5% Tween 20, 0.04% NaN 3 ) and probed for 4 h at room temperature in the HBST gel buffer containing 2.0 mM PMSF and 125 I-labeled human plasminogen or plasmin 3 ϫ 10 5 cpm ml Ϫ1 . The probed blots were washed several times with halfstrength HBST gel buffer containing 0.35 M NaCl, dried, and autoradiographed by exposure to Kodak X-OMAT AR film with an intensifying screen for 15 h at Ϫ70°C.
Purification of the 45-kDa Protein-The dialyzed and concentrated cell wall extracts were sequentially precipitated with ammonium sulfate at 40, 60, and 80% saturation. The precipitated proteins were then dialyzed against 50 mM Tris/HCl, pH 8.0, and concentrated to an appropriate volume. The proteins in the dialyzed preparations were resolved by SDS-PAGE, electroblotted onto a PVDF membrane, and probed with labeled plasmin(ogen). A strong plasmin(ogen) binding activity was found to be mainly associated with a 45-kDa protein of the sequentially fractionated cell wall extract with 40 -60% saturation of ammonium sulfate (Fig. 1, A and B). For further purification, 40 -60% ammonium sulfate precipitates were used as starting material. The dialyzed precipitate was concentrated (Centriprep-10, Amicon) and stored at Ϫ70°C until further use. The concentrated sample was applied to a Mono Q FPLC column (HR10/10, Amersham Pharmacia Biotech) pre-equilibrated with 50 mM Tris/HCl buffer, pH 8.0. After washing with 5-column volumes of this buffer, bound proteins were eluted with a 70-ml linear NaCl gradient from 0 to 700 mM and then with a 20-ml linear NaCl gradient from 700 mM to 1 M. Protein elution profile in each fraction was determined by SDS-PAGE and by Coomassie stain. A duplicate gel was Western blotted and probed with 125 Iplasmin(ogen) as described above. The 45-kDa protein eluted at 630 mM NaCl. The pooled fractions containing the 45-kDa protein and exhibiting plasmin(ogen) binding activity were dialyzed against the starting buffer and re-chromatographed on the Mono Q column using the same conditions. The positive fractions were again pooled and concentrated to a volume of Ͻ1.0 ml, using Centriprep-30 and Centricon-30 concentra-

FIG. 1. Identification of group A streptococcal plasminogenbinding protein.
Group A streptococcal cell wall-associated proteins were concentrated and separated by step-wise ammonium sulfate precipitation. The proteins therein were resolved by 12% SDS-PAGE and electroblotted onto PVDF membranes in duplicate. They were visualized by Coomassie Blue stain, and duplicate membranes were blocked and probed separately each with 125 I-labeled plasminogen and 125 Ilabeled plasmin as described under "Experimental Procedures." The position of a plasmin(ogen)-binding 45-kDa protein was identified (arrow). The position of SDH, a 35.8-kDa protein (4), is also highlighted (arrow). Numbers on left indicate the position of prestained molecular mass markers in kDa.
tors (Amicon). The concentrated sample was applied to a Superose-12 FPLC column (Amersham Pharmacia Biotech) pre-equilibrated with 50 mM Tris/HCl, pH 8.0. Fractions containing both the 45-kDa protein and plasmin(ogen) binding activity were pooled. These fractions were then concentrated, mixed with an equal volume of 4 M (NH 4 ) 2 SO 4 , and applied to a Poros BU/M hydrophobic column (Perspective Biosystems, Cambridge, MA) pre-equilibrated with 50 mM Tris/HCl buffer, pH 8.0, containing 2 M (NH 4 ) 2 SO 4 . The proteins were eluted with a 20-ml decreasing linear gradient of (NH 4 ) 2 SO 4 from 2.0 to 0.0 M. The 45-kDa protein was eluted in one fraction at 1.32 M (NH 4 ) 2 SO 4 . The eluted protein was then dialyzed and stored at a concentration of 250 g/ml at Ϫ70°C until further use. Protein concentration was determined by the BCA (bicinchoninic acid) method (Pierce).
N-Terminal Sequencing and Peptide Mapping-N-terminal amino acid sequence of the purified 45-kDa protein was determined as described (4,23). Briefly, the purified 45-kDa protein was resolved by SDS-PAGE and electroblotted onto a PVDF membrane. The protein was visualized by staining with 0.1% Ponceau S (Sigma) in 1% acetic acid. Plasminogen and plasmin binding activity was confirmed by autoradiography. The section of the membrane containing the protein band of interest was excised, destained with double distilled water, and subjected to automated Edman degradation. Each sample contained approximately 5 g of the protein as determined by the BCA protein estimation method (Pierce). A duplicate sample of PVDF membrane was digested with lysine-specific endopeptidase (Lys-C, sequencing grade, Boehringer Mannheim), and the resulting peptide fragments were separated by capillary electrophoresis interphased with the matrix-assisted laser desorption ionization time-of-flight mass spectrometer (Perspective Biosystems). N-terminal sequences of the two internal peptide fragments were then determined as described above. All microsequence analyses were performed at the Protein/Biotechnology Facility of the Rockefeller University.
␣-Enolase Activity and Enzyme Kinetics-The strong N-terminal sequence homology of the 45-kDa protein with ␣-enolase prompted us to investigate whether this protein is enzymatically active. ␣-Enolase activity was measured essentially as described earlier by both the coupled assay (24) and the direct assay at A 240 (25).
In the coupled assay, ␣-enolase activity was determined by measuring the transformation of NADH⅐H ϩ to NAD ϩ . The enzymatic reactions were performed at 37°C in 100 mM HEPES buffer, pH 7.0, containing 3.3 mM MgSO 4 , 0.2 mM NADH, 0.3 mM 2-phosphoglycerate (2-PGE), 1.2 mM ADP, 10.3 IU of lactate dehydrogenase, and 2.7 IU of pyruvate kinase in a final reaction volume of 1.0 ml. The reaction was started by adding 100 l of the test solution containing ␣-enolase. The ␣-enolase activity was measured in terms of the rate of reduction in the absorbance at 340 nm (i.e. increase in the production of NAD from NADH⅐H ϩ ). The decrease of the extinction at 340 nm was recorded as the change in (⌬)A 340 nm min Ϫ1 , using a Spectronic 3000 spectrophotometer (Milton Roy, Rochester, NY).
For kinetic studies, a single enzyme assay was used, involving only the transformation of 2-PGE to phosphoenolpyruvate (PEP) by ␣-enolase, thus avoiding interactions of the effectors with other enzymes. This reaction was performed at 37°C in 100 mM HEPES buffer, pH 7.0, containing 10 mM MgSO 4 and 7.7 mM KCl and different concentrations of 2-PGE (9 -35 mM) in a final volume of 1.0 ml. Change in absorbance/ min was monitored spectrophotometrically at 240 nm as described above. The results were recorded as the rate of PEP release at 5-s intervals for a period of 3 min at 240 nm. To determine the extinction coefficient of PEP, the absorbance of different concentrations of PEP (32.5-1040 g ml Ϫ1 ) was measured at 240 nm in 1-cm cuvettes. By using the Lambert-Beer formula (A ϭ a m cl), the extinction coefficient (a m ) for PEP (disodium salt) was calculated as 1.164 ϫ 10 Ϫ3 M Ϫ1 . By using 5 g of SEN and different concentrations of 2-PGE, the kinetic coefficients, K m and V max , were calculated from the values obtained for the intercepts and slopes of the double-reciprocal plots of Lineweaver-Burk (26).
␣-Enolase Activity of Intact Streptococci-To determine whether the 45-kDa protein is functionally active as an ␣-enolase on the streptococcal surface, an overnight culture of group A streptococci (D471) was washed (3 ϫ) with 100 mM HEPES/NaOH buffer, pH 7.0, and different concentrations of streptococci were incubated with and without 3 mM 2-PGE in 100 mM HEPES buffer, pH 7.0, containing 10 mM MgCl 2 and 7.7 mM KCl as described above. The reaction was allowed to occur in a final volume of 1.0 ml for a period of 3 min at room temperature, after which the bacteria were removed by centrifugation (4000 ϫ g for 10 min). The supernatants were analyzed by measuring absorbance at 240 nm as described above. For the remaining portion of the "Experimental Procedures," the 45-kDa protein will be referred to as SEN (surface enolase).
Production and Purification of Rabbit Polyclonal Antisera Against SEN-Polyclonal antibodies to SEN were prepared in New Zealand White rabbits immunized subcutaneously with 150 g of purified SEN emulsified in complete Freund's adjuvant (1:1) at multiple sites. Rabbits were boosted twice, each time with 150 g of the purified protein in incomplete Freund's adjuvant (1:1) at 3-week intervals. The rabbits were bled 10 days after the third immunization. All sera were filtersterilized and stored at 4°C. To prepare SEN-specific IgG, the polyclonal serum was subjected to sequential purification on protein-A Sepharose CL-4B (Amersham Pharmacia Biotech) and SEN affinity columns. The affinity column was made by covalently linking approximately 2 mg of purified SEN to 0.5 g of affinity matrix (Ultralink 3M-Carboxy beads, Pierce) with 200 l of 1-ethyl-3-[3-dimethylaminopropyl]carbodiimide-HCl (Pierce, 120 mg ml Ϫ1 ), in a final volume of 2.0 ml of 100 mM MES buffer, pH 4.7, at room temperature for 3-4 h. The protein-coupled matrix was then washed, and finally the affinity column was equilibrated with 50 mM HEPES/ NaOH buffer, pH 7.4. The polyclonal serum (2-3 ml) was first adsorbed to the protein A-Sepharose CL-4B column using 50 mM Tris/HCl, pH 8.0, as the initial buffer. After washing, the bound IgG was eluted with 200 mM glycine/HCl buffer, pH 2.5. Eluted IgG was dialyzed against the starting buffer, concentrated, and affinity purified on the SEN affinity column using a strategy similar to that as described above. SEN-specific IgG was then used for immunochemical analyses.
Production and Purification of Mouse Monoclonal Antibodies Against SEN-BALB/c ϫ SJL-F1 mice were subcutaneously immunized with 30 g of purified SEN in complete Freund's adjuvant (1:1 v/v). After 3 weeks, mice were bled and tested for antibodies to SEN by enzymelinked immunosorbent assay and Western blot analysis using a crude cell wall extract of group A M6 strain D471. Affinity purified rabbit polyclonal antibodies against SEN were used as a positive control. Mice with high antibody titers were given a second dose of antigen intraperitoneally in distilled water. Mouse spleens were excised 3-3.5 days after the last booster. The spleen cell fusion to P3-NS1/1AG4 -1(NS-1) myeloma cells was performed as described (27,28). Hybridomas cloned by limiting dilution were grown in 2-liter rolling tissue culture flasks. From these cultures, secreted monoclonal antibodies were precipitated at 50% ammonium sulfate saturation. The precipitates were then dialyzed and purified using a protein A-Sepharose affinity column.
Location of SEN in Streptococci-To determine the location of SEN in streptococcal cells, an overnight culture of strain D471 was subjected to lysine digestion in 30% raffinose buffer to extract streptococcal cell wall-associated proteins as described above. From the resulting protoplasts, the membrane and cytoplasmic fractions were separated as described previously (5,22). Proteins from each cellular fraction were resolved by SDS-PAGE and electroblotted onto a PVDF membrane. The presence of SEN in different cell fractions was monitored by affinity purified anti-SEN polyclonal (75.0 ng ml Ϫ1 ) and monoclonal antibodies (12 ng ml Ϫ1 ) as described (4,5).
Immune Electron Microscopy-Group A streptococci (D471) from the overnight TH broth cultures were harvested, washed, and adjusted to a concentration of 10 9 cfu/ml. An aliquot of 200 l of the bacterial suspension was incubated with 4 g of affinity purified anti-SEN(1A10) or anti-SDH (4F12) monoclonal antibodies for 4 h followed by a 2-h incubation with colloidal gold (5-and 10-nm sized beads for anti-SEN and anti-SDH labeled bacteria, respectively) anti-mouse IgG (Amersham Pharmacia Biotech, 1:25) at room temperature. The labeled bacteria were then fixed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer, pH 7.4, for 4 h at 4°C. The fixed labeled bacteria were then processed for transmission electron microscopy as described (29).

Prevalence of SEN in Various Group A Streptococcal M Types and Streptococcal
Groups-Proteins from the cell wall extract of several M serotypes after lysin digestion and those of various streptococcal grouping strains (group A-H, L, and N) after mutanolysin digestion were resolved by SDS-PAGE and transferred to a PVDF membrane. The blots were blocked, probed with affinity purified anti-SEN rabbit polyclonal antibodies (75.0 ng ml Ϫ1 ) for 3-4 h, and the reactive protein bands were visualized as described (4,5).
Location and Nature of Plasmin(ogen) Binding Domain of SEN-Purified SEN (10 g) was treated with carboxypeptidase Y (Boehringer Mannheim) at a substrate to enzyme concentration ratio of 13.5:1 in 50 mM HEPES buffer, pH 7.0, at 37°C for 6 h. Equal amounts of the enzyme-treated and untreated SEN were resolved by SDS-PAGE, and their plasminogen binding activity was determined by the blot overlay method as described above. In another set of similar experiments, plasmin(ogen) binding activity was measured in the presence of 0.1 M EACA or 0.1 M lysine.
Ligand Binding Assays for Plasmin(ogen) Binding to Immobilized SEN-The ligand binding analysis was carried out using 96-well mi-crotiter plates (C8 Maxi Break-apart, Nalge Nunc International, Naperville, IL). The plates were coated with 100 l of purified SEN in 0.05 M carbonate buffer, pH 9.6 (5 g ml Ϫ1 ), for 3 h at 37°C and were then kept at 4°C overnight. The plates were washed and blocked with HBST gel blocking buffer for 4 h at room temperature. A serial 2-fold dilution of aprotinin and PMSF-treated 125 I-plasminogen (3.8 pmol) or 125 Iplasmin (2.2 pmol) in a final volume of 100 l of HBST gel buffer containing 2 mM PMSF was added to the SEN-coated wells and incubated for 4 h at room temperature on a shaker. The plates were then washed three times with HBST gel buffer, and individual wells were counted in a (␥-counter for quantitative analysis of the bound plasmin-(ogen). Each dilution was tested in triplicate wells. Nonspecific binding was measured after the addition of a 200 molar excess of unlabeled plasminogen/plasmin or 0.1 M EACA. Nonspecific binding was also evaluated in BSA-coated plates. Nonspecific binding contributed between 4 and 10% of the total counts without these agents. Specific binding was calculated by subtracting nonspecific binding (10%) from the total binding. The amount of free plasmin(ogen) was calculated by subtracting specifically bound plasmin(ogen) from the total amount of labeled plasmin(ogen) added. To determine the equilibrium dissociation constant (K D ), a nonlinear least square analysis of the total count offered versus the count bound was carried out using the curve fitting computer program from Sigma Plot. The values of bound plasmin(ogen) versus bound/free ratio were plotted, and the slope representing Ϫ1/K D was determined by linear regression analysis using the formula of Scatchard (30). The specific binding of plasmin(ogen) to SEN was also determined in a competition assay in which a constant amount of 125 I-plasminogen (0.77 pmol) and 125 I-plasmin (1.37 pmol) was mixed with decreasing concentrations of free plasmin(ogen) (5 ng to 50 g, i.e. up to 380 pM excess). The amount of plasmin(ogen) bound to immobilized SEN in wells, in the absence of any competitor, was treated as the maximum binding value. Percentage of the maximum binding at different concentrations of the competitors was plotted.
Cross-linking Studies-Sulfosuccinimidyl-2-[p-azido-salicylamido] ethyl-1-3Ј-dithiopropionate (SASO) (Pierce, 300 g) was iodinated (0.5 mCi 125 I-Na) with IODO-GEN (100 g), and conjugation of [azidosalicylamido]ethyl-1-3Ј-dithiopropionate to plasmin(ogen) (400 g) was carried out in the dark essentially as described before (31) with minor modifications. 125 I-labeled plasminogen-ASD was purified on a PD-10 column as described above. Cross-linking of 125 -labeled plasminogen-ASD with intact group A streptococci was performed as follows. Overnight TH broth culture of streptococci (D471) was centrifuged, washed once with phosphate-buffered saline, and resuspended to a concentration of 2 ϫ 10 10 cfu/ml. An aliquot of (100 l) of this streptococcal suspension was incubated with 125 I-labeled plasminogen-ASD (3 ϫ 10 5 cpm) in the dark under constant rotation at 37°C for 1 h in a final volume of 150 l. The reaction mixtures were then irradiated for an additional 30 min with UV 350 nm light. The labeled bacteria were then resuspended in 50 mM phosphate buffer containing 30% raffinose, digested with mutanolysin, and further fractionated into cell walls, cytoplasm, and membranes as described (4,5). Samples were then analyzed by SDS-PAGE under reducing conditions, followed by autoradiography.
Plasminogen Binding Activity of Intact Streptococci-Streptococci (5 ϫ 10 9 cfu/ml, 50 l/well) were fixed in a 96-well poly-L-lysine-coated microtiter plate with 0.2% glutaraldehyde for 1 h at room temperature. Unoccupied sites were blocked by 0.1 M lysine. The plates containing streptococci were then treated overnight at 4°C with 2% BSA in HBST gel buffer. To determine the role of the C-terminal lysine residue in plasminogen binding, some of wells with fixed streptococci were treated with carboxypeptidase B (5 g/well, Boehringer Mannheim) for a period of 4 h at 37°C on a microtiter plate shaker. Dose-dependent plasminogen binding activities of intact and carboxypeptidase B-treated streptococci were measured using serial 2-fold dilutions of 125 I-plasminogen in a final volume of 100 l of HBST gel buffer, as described above for plasminogen binding activity of purified SEN. The plates were then washed three times with HBST gel buffer, and individual wells were counted in a ␥-counter for a quantitative analysis of the bound plasmin(ogen). Each dilution was tested in triplicate.
Competitive inhibition of plasminogen binding to intact streptococci was measured in the presence of varying amounts (molar excesses) of purified SEN or SDH and a fixed amount of 125 I-plasminogen (0.77 pmol) in a final volume of 100 l of HBST gel buffer using microtiter plates precoated with group A streptococci as described above.
In Vitro Proteolytic Activity of Plasmin Bound to SEN-␣ 2 -Antiplasmin is a fast-acting plasmin inhibitor of plasma (32). Thus, the proteolytic activity of plasmin was evaluated in terms of its inhibition in the presence of ␣ 2 -antiplasmin when bound to various substrates. To de-termine the proteolytic activity of plasmin bound to SEN, purified SEN was immobilized onto Ultralink 3M-Carboxy beads (Pierce), using a method similar to that as described above for the preparation of monospecific anti-SEN antibodies. Ultralink beads containing 1 g of SEN were mixed with 125 I-labeled plasmin (0.375 g) or plasminogen (0.39 g) for 4 h in a final volume of 100 l. Unbound plasmin(ogen) was then removed by three washes with HBST gel buffer. The amount of bound plasmin(ogen) was determined on the basis of radioactive counts. The proteolytic activity of bound plasmin was determined by measuring the cleavage of the chromogenic substrate Val-Leu-Lys-para-nitroanilide (Sigma, 23.8 g, 3.4 mg ml Ϫ1 ) in a final reaction volume of 200 l of HBST gel. The change in absorbance at 405 nm was determined spectrophotometrically (MR 4000, Dynatech Laboratories, Inc., Chantilly, VA). The inhibitory effect of ␣ 2 -antiplasmin on equivalent amounts of SEN-bound and free plasmin was determined by measuring the cleavage activity on the chromogenic substrate. Similarly, SEN-bound and free plasminogen were activated with either tPA (600 units) or streptokinase (100 units), and the subsequent proteolytic activity of the generated plasmin was measured in the presence and absence of ␣ 2antiplasmin as described above. Blanks were run with buffer containing only substrate.
Opsonophagocytosis of Group A Streptococci in the Presence of Anti-SEN Antibodies-In vitro phagocytosis of group A type M6 streptococci (strain D471) and heterologous type 14 strain (T/14/46) by human phagocytes in the presence of polyclonal anti-SEN antibodies was measured as described (33). Briefly, 0.4 ml of freshly drawn heparinized blood was mixed with 0.1 ml of appropriately diluted logarithmic phase culture streptococci (100 -300 cfu ml Ϫ1 ) in the presence of different concentrations of anti-SEN antibodies. The mixtures were incubated at 37°C for 3 h either with constant slow rotation or under a stationary condition. At the end of the incubation, an aliquot was plated on proteose peptone blood agar and incubated at 37°C overnight. Surviving bacteria were determined from the number of ␤-hemolytic colonies. The experiments carried out under stationary conditions served as an internal control.

Identification of a Novel 45-kDa Plasmin(ogen) Binding Protein in the Group A Streptococcal Cell Wall Extracts-By
Western blot, we examined the plasmin(ogen) binding activity of proteins in a crude streptococcal cell wall extract using 125 Ilabeled plasmin and plasminogen (Fig. 1A). The results showed that, in addition to the weak plasminogen binding to the 39-kDa SDH molecule (4), a significantly stronger plasminogen binding occurred with a 45-kDa protein present in the streptococcal cell wall extract. We also found similar binding activities of SDH and the 45-kDa protein with 125 I-labeled plasmin (Fig.  1B). These findings identify a new protein with strong plasmin-(ogen) binding activity in the cell wall extract and confirm our previous report of the weak plasmin(ogen) binding activity of SDH (4).
Purification of the 45-kDa Protein-The 45-kDa protein was partially purified from the cell wall extract by 40 -60% ammonium sulfate precipitation. The protein was further purified by ion-exchange chromatography on a Mono Q column followed by a Superose-12 molecular sieve. With the latter, the peak elution volume having plasmin(ogen) binding activity corresponded to that of a standard 150-kDa IgG molecule, suggesting that the native form of the 45-kDa protein is likely a multimer. Final purification was achieved on a Poros BU/M hydrophobic column (Fig. 2A). The average yield of purified 45-kDa protein from a total of 10 liters (1.5 ϫ 10 8 cfu/ml) of bacterial culture was 1.128 mg.
N-terminal Amino Acid Sequence and Identification of the 45-kDa Protein as an ␣-Enolase Enzyme-N-terminal amino acid sequence of the 45-kDa protein revealed a single amino acid in the first 50 residues (Fig. 2B). N-terminal sequences of two internal peptides (Pep-1, molecular mass 1712.1 Da; and Pep-2, molecular mass 1683.5 Da) obtained after cleavage with a lysine-specific endopeptidase were also determined for 15 and 11 residues, respectively. The presence of a single amino acid at each sequence cycle for the intact 45-kDa protein and each internal peptide verified the homogeneity of these molecules.
When the three sequences were compared with known sequences in the translated GenBank TM data base, 81% identity and ϳ90% homology was found with the first 50 N-terminal residues of ␣-enolases of Bacillus subtilis and those of prokaryotic (Fig. 2B) and eukaryotic origin (comparison not shown). Similarly, the two internal peptide sequences (Pep-1 and Pep-2) also revealed Ͼ90% identity with corresponding regions of ␣-enolase from various sources, placing Pep-1 in the center of the molecule and Pep-2 toward the C terminus (Fig. 2B). These findings confirm the identity of the 45-kDa plasmin(ogen)binding protein as ␣-enolase.
␣-Enolase Activity and Enzyme Kinetics-By establishing that the sequence of the 45-kDa protein was that of ␣-enolase, we investigated whether it also possessed the activity of this glycolytic enzyme. In a coupled-enzyme assay, the 45-kDa protein converted terminal NADH to NAD in a dose-dependent manner (Fig. 3A). This indicated the conversion of pyruvate to lactate by lactate dehydrogenase and NADH, confirming the conversion of phosphoglycerate to phosphoenolpyruvate by ␣-enolase and further to pyruvate in a sequential manner in the presence of externally furnished pyruvate kinase and ADP. A similar dose-dependent conversion of 2-phosphoglycerate to phosphoenolpyruvate was exhibited by the 45-kDa protein in a direct enzyme assay (not shown) also confirming that it is an enzymatically active ␣-enolase. The latter assay was used to determine the kinetic properties of the 45-kDa protein.
The results of enzyme reaction rates, using 5 g of purified protein with various concentrations of 2-phosphoglycerate (0.19 -7.5 mM) measured at 240 nm, were analyzed as Michaelis-Menten and double-reciprocal Lineweaver-Burk plots as shown in Fig. 3, B and C. The analysis revealed a K m value of 1.492 mM for 2-phosphoglycerate. From these plots, a V max of 31.25 mM phosphoenolpyruvate min Ϫ1 mg Ϫ1 of the 45-kDa protein was determined. Considering 1 IU as the enzyme activity necessary to transform 1 mol of substrate per min, the purified 45-kDa protein yielded a total of 70 IU g Ϫ1 .
␣-Enolase Activity of Intact Streptococci-To determine whether the ␣-enolase activity of the 45-kDa protein is in fact displayed on the streptococcal surface, enzymatic activity was carried out using intact group A streptococci by the single enzyme assay method in the presence of 3 mM 2-phosphoglycerate. The results shown in Fig. 3D revealed a dose-dependent ␣-enolase activity catalyzed by the intact streptococci. In the absence of 2-phosphoglycerate, intact streptococci did not catalyze any detectable enzyme reaction, further suggesting that the 45-kDa protein is expressed on the surface. In addition, to rule out the possibility that this enzymatic activity was not due to cell lysis-related release of enzyme, we assayed for the presence of the cytoplasmic enzymes lactate dehydrogenase and pyruvate kinase and found no activity (not shown). Based on these results, the 45-kDa protein is hereafter referred to as SEN.
Subcellular Location and Prevalence of SEN in Other M Types and Streptococcal Groups-By using affinity purified rabbit anti-SEN antibodies and monoclonal antibodies, we examined the distribution of SEN in various subcellular fractions of group A streptococci by Western blot analysis. The results showed that SEN is found in both the cell wall and cytoplasmic fractions, with negligible amounts in the membrane fraction (Fig. 4 A). Furthermore, SEN was found to be present in comparable quantities in all M types examined and in all streptococcal groups except group N (Fig. 4B). The uniform antibody reaction that showed no obvious size heterogeneity among the SENs in different M types indicates that SEN is a conserved protein in all streptococci tested.
Plasmin(ogen) Binding Activity of SEN and Its Comparison with That of SDH-To compare the plasmin(ogen) binding activities of purified SEN and SDH, equal quantities of the purified proteins were separately resolved by SDS-PAGE and electro-blotted onto PVDF membranes. The blots were then probed with either 125 I-plasminogen, 125 I-plasmin, or 125 I-plas-

FIG. 2. Purification and characterization of the 45-kDa protein.
A, the presence of the 45-kDa protein was identified in 40 -60% ammonium sulfate precipitates (see also Fig. 1). This was further purified by ion-exchange (Mono Q), molecular sieve (Superose-12), hydrophobic (Bu/M) column chromatography. Purified 45-kDa protein retained both plasminogen and plasmin (not shown) binding activity. B, the 45-kDa protein was subjected to N-terminal amino acid sequencing of the intact molecule and internal peptides obtained by digestion with Lysspecific endopeptidase. The first 50 amino acid residues of the intact molecule and 10 residues each of the two internal peptides (Pep-1 and Pep-2) were compared with sequences in the GenBank TM data base. The results show a strong sequence identity (87.5-100%) with the N-terminal and peptide fragments of B. subtilis ␣-enolase (Bst-␣-enolase). The location of the Pep-1 corresponds to the middle of the molecule whereas that of Pep-2 corresponds to the C-terminal region of the molecule. min derived from 125 I-plasminogen by urokinase. The results showed that plasmin(ogen) bound weakly to SDH compared with SEN (Fig. 5A). The results are in agreement with our previous results (4) and with those shown in Fig. 1. Furthermore, SEN consistently showed significantly higher binding affinity for plasminogen than plasmin (Fig. 5A).
The plasmin(ogen) binding activity of SEN was found to be significantly decreased in the presence of 0.1 M lysine or 0.1 M EACA (not shown), suggesting that the exposed lysine residue(s) in SEN is likely responsible for this binding activity (Fig.  5B). Similarly, carboxypeptidase Y-treated SEN also showed significant decrease in binding activity to plasmin(ogen), indicating that the lysine-binding residues are likely located at the C-terminal end of the SEN molecule (Fig. 5C).
To determine the specific binding activity of plasmin(ogen) to SEN, a quantitative solid phase assay was carried out using 96-well microtiter plates and different concentrations of 125 Iplasminogen and 125 I-plasmin. Nonspecific binding of 125 I-labeled plasminogen/plasmin to only BSA-coated wells or with SEN-coated wells in the presence of 0.1 M EACA or 200 molar excess of unlabeled plasmin(ogen) did not exceed Ͼ10% of the bound counts in the absence of these agents. Thus, to determine the specific binding of plasmin(ogen) to SEN, only 90% of the total bound counts were considered for the calculation. 125 I-Labeled plasminogen and plasmin reacted specifically and strongly with SEN-coated wells. Scatchard analysis of the binding of plasminogen and plasmin to SEN was nonlinear, indicating the presence of more than one site for this interaction. Thus, for plasminogen, two equilibrium dissociation constants (K Dpg1 ϭ 1.3 nM and K Dpg2 ϭ 7.4 nM) were recorded. Although plasmin bound to SEN with relatively lower affinity, Scatchard analysis revealed a similar nonlinear curve with two distinct equilibrium dissociation constants (K Dpl1 ϭ 2.2 nM, K Dpl2 ϭ 27 nM) (Fig. 5D). The reaction between 125 I-plasminogen/plasmin and immobilized SEN (solid phase on microtiter plates) was inhibited by unlabeled plasminogen/plasmin in a dose-dependent manner. At 156 molar excess of unlabeled plasminogen, approximately 90% inhibition in the binding of the labeled plasminogen to SEN was recorded. At a similar concentration of free plasmin, relatively less inhibition (75%) in the binding of the labeled plasmin was recorded. Together, these results revealed the specificity of the interactions and high affinity of plasmin(ogen) to SEN, indicating that SEN may have more than one site for the binding of plasmin(ogen) and may serve as the major plasmin(ogen) binding molecule on the surface of group A streptococci.
Role of SEN in Plasminogen Binding Activity of Streptococci-In addition to enzymatic and biochemical properties suggesting that SDH (4) and SEN are located on the streptococcal surface (Figs. 2-4) and to provide additional proof of their presence on the streptococcal surface, SEN-and SDH-specific mouse monoclonal antibodies (1A10 and 4F12, respectively) were used in indirect immune electron microscopy. As shown in Fig. 6A, both SEN and SDH molecules, reacting with their specific monoclonal antibodies (4 g/10 6 streptococci), were found on the surface of streptococci. Their binding patterns suggest that the distribution of these proteins is either in the form of a cluster or the epitopes recognized by specific monoclonal antibodies are not uniformly exposed on the cell surface. The latter argument is supported by the fact that even at higher concentrations of both monoclonal antibodies (up to 20 g of IgG/10 6 streptococci), of the distribution of gold particles was the same as that seen with lower concentrations of monoclonals (data not shown).
By having confirmed the location of SEN and SDH on the streptococcal surface, the role of these proteins in plasminogen binding by streptococci was further investigated. Since plasminogen binding activity is dependent on the C-terminal lysine residue, we subjected intact streptococci to carboxypeptidase B digestion to remove the C-terminal lysine residues of the plas-FIG. 3. ␣-Enolase activity associated with group A streptococci. A, ␣-Enolase activity of the purified 45-kDa protein was measured, using a coupled enzyme assay catalyzing 2-phosphoglycerate to phosphoenolpyruvate as described under "Experimental Procedures." The decrease in absorbance at 340 nm depicting the catalytic rate of conversion of NADH to NAD min Ϫ1 in the presence of lactate dehydrogenase, ADP, and pyruvate kinase was measured as A 340 nm min Ϫ1 . B, enzyme kinetics of the purified 45-kDa protein was determined, using 5 g of the purified 45-kDa protein as the enzyme source. The rate of conversion of 2-PGE (S) to PEP min Ϫ1 (V, absorbance at 240 nm) in the presence of various concentrations of 2-phosphoglycerate was plotted by the method of Michaelis-Menten. C, Lineweaver-Burk double-reciprocal kinetic analysis of SEN was derived from the Michaelis-Menten plot (see C). Each data point is the average of triplicate enzyme assays. D, ␣-enolase activity of intact streptococci. Enzyme activity of intact streptococci was measured as described in A. Unlike A, the enzyme activity in the reaction mixture was measured at the end of 2 min incubation only as the end point decrease in absorbance at 340 nm of the bacteria-free supernatant rather than as a rate analysis. minogen-binding proteins. These results were compared with those obtained for the plasminogen binding activity of streptococcal cell wall proteins extracted from intact untreated and carboxypeptidase B-treated streptococci (Fig. 6B). In this analysis, carboxypeptidase B-treated streptococci showed a significant reduction in plasminogen binding activity. Similarly, by phosphor image analysis, the cell wall-associated SEN from the carboxypeptidase B-treated streptococci showed 20 -30% less binding as compared with SEN associated with untreated streptococcal cell walls (Fig. 6B). In both treated and untreated streptococci, plasminogen binding activity of cytoplasmic SEN was found to be the same. The fact that carboxypeptidase B did not completely remove the plasminogen binding activity suggests that whereas the C-terminal lysine residue of a protein molecule may play an important role in its ability to bind plasminogen, the region N-terminal to the C-terminal lysine residue may also play a role in this binding. As described above, plasminogen binding activity of SDH was found to be significantly less than that of SEN, and carboxypeptidase B treatment did not seem to affect plasminogen binding activity.
In the blot overlay plasminogen binding assay, both SEN and SDH were found to bind plasminogen with a distinct high and low affinity; however, it is not clear whether SEN or SDH or both these molecules, while bound on the streptococcal surface, participate in plasminogen binding. To understand the individual roles of these molecules in streptococcal plasminogen binding activity, a cross-linking study was carried out using 125 Ilabeled plasminogen-ASD and intact streptococci. The presence of the labeled proteins was detected in the streptococcal cell wall extract and not in the corresponding cytoplasmic preparation, indicating that 125 I-labeled plasminogen-ASD came in contact with only surface-exposed molecules. Furthermore, SEN, but not SDH, was found to be labeled, suggesting that SEN, because of its high affinity with plasminogen may be cross-linked with 125 I-labeled plasminogen-ASD more efficiently as compared with SDH (Fig. 6C). These results were further confirmed by a competitive inhibition binding assay in which the plasminogen binding activity of intact streptococci was measured in the presence of purified SEN or SDH (Fig.  6C). The results showed that in the presence of low concentrations of purified SEN (Յ1.6 M), up to 50% of the plasminogen binding activity was inhibited and at higher concentrations (Յ12 molar excess) up to 75% inhibition could be achieved (Fig.  6C). At similar low concentrations of SDH, plasminogen binding activity of streptococci was inhibited only to a minimal level (6%). Plasminogen binding activity of streptococci could be inhibited up to 50% by 12 molar excess of free SDH (Fig. 6C). Together these data indicate that SEN may play an important  5. Plasmin(ogen)-binding properties of SEN. A, comparison between the plasmin(ogen) binding activity of purified SEN and that of SDH (4) (see also Fig. 1). One g each of SEN, SDH, and cytoplasmic SDH (4, 5) were electroblotted in replicates onto a PVDF membrane. One gel was stained with Coomassie Blue. The PVDF membranes were each blocked and probed with either 125 I-labeled plasminogen, 125 I-plasmin, or role in streptococcal plasminogen binding and is the functional plasminogen-binding receptor protein on the surface of group A streptococci.
Protease Activity of Plasmin and Activated Plasminogen Bound to Purified SEN-The proteolytic activity of free and SEN-bound plasmin was determined by measuring their ability to cleave the chromogenic substrate Val-Leu-Lys-para-nitroanilide in the presence and absence of ␣ 2 -antiplasmin, a fast acting plasmin inhibitor. The results show that plasmin, either bound to SEN or free in solution, exhibits the same proteolytic activity (Fig. 7). However, SEN-bound plasmin was not inactivated as rapidly in the presence of ␣ 2 -antiplasmin as free plasmin. After 4 h, SEN-bound plasmin still retained activity whereas free plasmin was nearly completely inactivated (Fig. 7).
We next studied the proteolytic activity of SEN-bound plasminogen to plasmin after tPA activation in the presence and absence of ␣ 2 -antiplasmin. As expected, plasminogen (either SEN-bound or free) exhibited no proteolytic activity (Fig. 7), whereas tPA-activated plasminogen showed high activity, with the SEN-bound form exhibiting somewhat higher activity (23% more) than an equal amount of free plasminogen. However, after 4 h in the presence of ␣ 2 -antiplasmin, SEN-bound tPAactivated plasminogen exhibited a significantly elevated proteolytic activity when compared with the activated form of free plasminogen (Fig. 7). In contrast, ␣ 2 -antiplasmin could not inactivate streptokinase-activated free or bound plasminogen as reported earlier (13,34). Taken together, these results indicate that plasmin bound to SEN is significantly resistant to inactivation by ␣ 2 -antiplasmin.
Opsonophagocytosis of Group A Streptococci in the Presence of anti-SEN Antibodies-To understand the biological role of SEN as an important streptococcal surface protein, we determined whether antibodies to SEN were able to opsonize these organisms in vitro. The opsonic activity of anti-SEN antibodies was measured in terms of the ability of streptococci to survive in blood from a nonimmune individual who lacks type-specific anti-M antibodies against the test strain. The results, as shown in Fig. 8, A and B, revealed that affinity purified anti-SEN IgG antibodies effectively opsonized and enhanced the phagocytosis of group A streptococci of two different serotypes (types 6 and 14). These results not only confirm the surface location of SEN on streptococci but also suggest that anti-SEN IgG antibodies may foster non-type-specific protection against streptococcal infection.

DISCUSSION
The plasmin(ogen) binding property of pathogenic bacteria in general, and of streptococci in particular, is suggested to be one of the characteristics that may contribute to tissue invasion and the overall pathogenicity of group A streptococci (13,34). Plasminogen activation is responsible for the degradation of intravascular clots and extracellular proteolysis in a wide variety of physiological and pathological processes (7)(8)(9). In this report, we identify and characterize a novel plasmin(ogen)binding protein, SEN, on the surface of group A streptococci. Structurally and functionally this protein is an ␣-enolase, one of the key glycolytic enzymes, and is the second glycolytic enzyme that we have identified on the surface of group A streptococci. We reported earlier that streptococcal surface dehydrogenase (SDH), a major protein on the surface of these organisms, is structurally and functionally related to the glycolytic enzyme glyceraldehyde-3-phosphate dehydrogenase and is able to bind plasmin weakly (4). On the basis of latter activity, another group has found a structurally and functionally similar protein, Plr (plasmin receptor), from a clinical group A streptococcal strain (6). The identification of this 45-kDa plasminogen-binding protein as ␣-enolase was based on its strong sequence identity, both at its N-terminal end and within two different internal peptides to other reported ␣-enolases. We have identified and isolated SEN from the M6 streptococcal strain from which SDH was originally identified and purified (4).
Functional identity of SEN in its purified form, as well as on the streptococcal surface, was confirmed by its ability to catalyze the conversion of 2-PGE to PEP in both direct and coupled enzyme assays. The values for the enzyme kinetic constants, K m and V max , for purified SEN (Fig. 3) are comparable with those of other reported ␣-enolase enzymes (35,36). Furthermore, both monospecific polyclonal and monoclonal antibodies against SEN were found to be important tools for its cellular location in group A streptococci and its presence on the surface of other streptococci (Fig. 4).
From this and several other reports (6,14,15), it is apparent that group A streptococci express more than one type of plasmin(ogen) binding receptor which acquires plasmin(ogen) through various mechanisms (34). Clinical isolates and animalpassaged strains have more plasmin(ogen) binding capacity then the same strains passaged in the laboratory (37). Thus, group A streptococci expressing several low affinity plasmin(ogen) binding molecules may bind the equivalent amount of plasmin(ogen) as strains that express fewer high affinity molecule(s). In addition to strain specificity, M type specificity has also been reported for the plasmin(ogen)-binding protein, PAM, in group A streptococci (15). Although not characterized, at least two types of plasmin(ogen) receptors (low and high affinity) have been reported in group G streptococci (38), a characteristic that may well be found on other streptococcal groups. We report here that in group A streptococci, surface glyceraldehyde-3-phosphate dehydrogenase (SDH/Plr) is a weak plasminogen-binding protein and that SEN is a strong plasminogen-binding protein (Fig. 5). By using anti-SEN antibodies, we have identified ␣-enolase-like molecules on the surface of both encapsulated and unencapsulated strains of Streptococcus pneumoniae, but not on staphylococci, 2 suggesting that surface ␣-enolase (plasmin(ogen) binding) may be an important virulence determinant for pathogens of the respiratory mucosa. the labeled product (urokinase-plasmin) obtained from the urokinase-treated 125 I-labeled plasminogen. The plasminogen/plasmin binding activity of SEN and SDH was then visualized by autoradiography. B, effect of the presence of lysine on the plasminogen binding activity of SEN. Different concentrations (0.04 -1.0 g) of purified SEN was Western-blotted on PVDF membranes, and plasminogen binding activity as described above in A was carried out in the presence and absence of 0.1 M lysine and visualized by autoradiography. C, effect of carboxypeptidase Y treatment on the plasminogen binding activity of SEN. 50 g of purified SEN was treated with carboxypeptidase Y (3.7 g) for 6, 12, and 18 h. Enzyme treated (ϩ) and untreated (Ϫ) preparations were electroblotted onto PVDF membranes, reacted with 125 I-labeled plasminogen, washed, and autoradiographed. A duplicate PVDF membrane containing the enzyme-treated SEN was also stained with anti-SEN polyclonal antibodies (75 ng ml Ϫ1 ) as described in Fig. 4. D, to determine specific binding of plasminogen and plasmin to SEN. Microtiter plate (break-apart) wells coated with 0.5 g/well SEN were incubated with increasing concentrations of aprotinin/PMSF-treated 125 I-plasminogen and -plasmin. Nonspecific binding was assessed as binding to BSA-coated wells and subtracted from the total binding. Nonspecific binding contributed ϳ10% of the total binding. Scatchard analysis of the specific binding data both for plasminogen and plasmin is shown as an inset in the corresponding figure, respectively. Competitive inhibition of the interaction between plasminogen/plasmin and SEN was assessed by determining the binding of the indicated amount of 125 I-plasminogen/plasmin to immobilized SEN in the presence of increasing molar excess of unlabeled plasminogen or plasmin. Each data point represents the mean of three different experiments. In each experiment, an average reading for each parameter was calculated from three individual wells.
␣-Enolase is one of the key glycolytic enzymes found generally in the cytoplasm; nevertheless, its presence on the surface of cells is not without precedent. Several eukaryotic studies have provided evidence that ␣-enolase-related molecules are expressed on the surface of several cell lines such as U937 human monocytoid (19), human breast tumor (17), peripheral blood monocytes, and neutrophils (18) and that these molecules contribute about 10% of the plasminogen-binding capacity of the cells. Recently, ␣-enolase has also been shown to be present as an abundant immunodominant antigen in the cell wall of Candida albicans (39,40). In prokaryotes, however, the presence of cell-surface ␣-enolase has not been previously reported. Our present findings on the plasmin(ogen) binding activity of SEN (Fig. 5) differ from those of the reported eukaryotic plasminogen binding ␣-enolases (18) in two major respects. (i) SEN exhibits significantly higher affinity for plasmin(ogen) (K D ϭ 1-4 nM, Fig. 5D) as compared with that of eukaryotic enolase (K D ϭ 0.1-2 M). (ii) In contrast to eukaryotic enolase, SEN exhibited more than one interaction site for plasminogen and plasmin. Although plasminogen and plasmin showed comparable binding affinity to SEN in a solid phase assay, the latter consistently bound less efficiently on Western blots. We speculate that in addition to the C-terminal lysine binding site of the SEN molecule for plasmin(ogen), the region upstream of this site may also be responsible for the binding. This is supported by the fact that plasmin(ogen)-binding proteins that do not possess lysine residues at their C-terminal ends also show appreciable affinity for plasmin(ogen), possibly through other exposed lysine residue(s) (38,41).
The high affinity of intact group A streptococci for plasmin-(ogen) (K D ϭ 0.2 nM (37)) has recently been attributed to the Plr protein, a member of GAPDH family (6); however, this activity may actually be due to the combined activity exhibited by several such binding proteins on the streptococcal surface. Recently, glyceraldehyde-3-phosphate dehydrogenase isolated and purified from Streptococcus equisimilis has been shown to have an equilibrium constant in the range of 220 -260 nM for plasminogen and about 25 nM for plasmin (41). We reported earlier that SDH, also a member of GAPDH family, is a weak plasmin(ogen)-binding protein (4). In view of these reports and of other published reports showing low affinity plasminogenbinding proteins of group A streptococci (14,15,38), SEN may be the major plasmin(ogen)-binding protein on the surface of group A streptococci. Based on its ubiquitous presence on the surface of a variety of group A streptococcal serotypes and streptococcal groups (Fig. 4), we suggest that SEN, or an SENlike molecule, may serve as a major plasmin(ogen)-binding molecule/receptor on the surface of nearly all pathogenic streptococci and would therefore play an important role in disease outcome.
Earlier reports that the Plr molecule (6,42,43) or other plasminogen-binding proteins (15,41) are the streptococcal plasminogen-binding receptors, are based on the plasminogen binding activity of the purified natural or recombinant proteins using different methods. It is not clear, however, if Plr (6) or SDH (4) is in fact the streptococcal plasminogen-binding receptor and, if so, whether this binding activity represents the observed plasminogen binding activity exhibited by intact streptococci. Hence, in the present communication we investigated whether plasminogen binding activities of SEN and SDH are relevant to the observed streptococcal plasminogen binding activity. We first confirmed their surface location using immune electron microscopy with SEN-and SDH-specific monoclonal antibodies (Fig. 6A). Furthermore, we found a significant reduction in plasminogen binding activity of carboxypeptidase B-treated intact streptococci indicating that the structural domain of SEN and/or SDH which contains the C-terminal lysine residue is exposed to the surface. In conjunction with these results, we found that the plasminogen binding activity of the surface-bound SEN, but not that of cytoplasmic SEN, was reduced after treatment with carboxypeptidase B, suggesting that SEN was accessible to this enzyme. The observation that FIG. 7. Functional consequences of plasminogen bound to SEN and protection of SEN-bound plasmin activity. One g of SEN was bound to 3M-Emphase ultralinked Carboxy beads. SEN beads were then separately mixed with 125 I-labeled plasminogen (Pg) and -plasmin (Pl) and washed, and the amount of bound labeled plasmin(ogen) was calculated from their specific radioactivity g Ϫ1 . Plasminogen was activated either by tissue plasminogen activator (tPA, 600 units, i.e. 0.02 g) or streptokinase (Sk, 100 units). Protease activity of free or bound plasmin or the one derived from plasminogen was determined spectrophotometrically at 405 nm in the presence or absence of ␣ 2 -antiplasmin (Apl) using the chromogenic substrate Val-Leu-Lys-p-nitroanilide (23.8 g/reaction). Each bar represents the average value of three individual readings. by human phagocytes was carried out in the presence of affinity purified rabbit polyclonal (10 g/reaction) or mouse monoclonal (7.5 g/reaction) anti-SEN IgG as opsonic source under either rotation or stationary conditions at 37°C for 3 h as described under "Experimental Procedures." The number of surviving streptococci in the tube were enumerated as cfu by plate count. The inset in each figure shows the cfu obtained under stationary conditions which served as internal control for bacterial growth. Each bar represents an average of three independent experiments with S.D. ϳ10 -15% of the average value. the enzyme-treated streptococci also showed significant plasminogen binding activity raises the possibility that the region which is N-terminal to the C-terminal lysine residues of SEN (and also that of SDH) may also play a role in plasminogen binding activity. From the cross-linking studies that were designed to determine whether SEN, SDH, or both the proteins play a role in streptococcal plasminogen binding, it was possible to determine that SEN is probably more exposed to the surface as compared with SDH, since the labeled ASD from the 125 I-labeled plasminogen-ASD complex was found to cross-link to SEN rather than to SDH. A dose-dependent inhibition of the streptococcal plasminogen binding activity, even in the presence of low concentrations of free SEN, further confirmed that the high affinity of SEN for plasminogen inhibits the ability of the bacteria to bind to plasminogen. These results together indicate that SEN serves as a primary receptor for plasminogen binding. The question of how glyceraldehyde-3-phosphate dehydrogenase (SDH/Plr) (4, 6) and ␣-enolase (SEN) are transported through the cell membrane and sorted onto the cell surface without the presence of a signal sequence remains intriguing. Whether SEN, like eukaryotic surface ␣-enolases, is transported by internal signal sequences like that found with plasminogen activator inhibitor 2 (44) or by the post-translational acylation method (45) needs further investigation. It is possible that SEN and SDH are transported to the cell surface as a complex in conjunction with a secreted protein utilizing a specialized secretory system. What is clear, however, is that SEN and SDH, are members of the growing group of proteins which lack signal sequences but are transported to the surface and anchored to cells by an as yet undefined mechanism (44).
Like GAPDH/SDH (4,46), eukaryotic ␣-enolase has also been shown to be a multifunctional protein presenting a variety of activities besides its native glycolytic activity. Functions such as being a structural component of turtle lens, J-crystallin (47), a neurotropic factor (48), the ability to form a stable complex with Clostridium difficile toxin B (49), and the ability to bind to polynucleotides (50) have been recently reported for ␣-enolase. If SEN, like eukaryotic ␣-enolases (47-50) and SDH (4,5), is also a multifunctional molecule, its physiological implications are at present equivocal.
Our finding that plasmin bound to SEN retains its proteolytic activity even in the presence of ␣ 2 -antiplasmin indicates that SEN may be an important streptococcal virulence determinant. Particularly in infected tissues, such a characteristic would enable the streptococcus to become more invasive by evading localization by the host's clotting pathway (34). The fact that ␣-enolase has recently been found to be secreted in the growth medium (39,51) and that increased levels of fungalspecific ␣-enolase have been found in patients with invasive candidiasis (52) suggest that a similar phenomenon may exist in cases of invasive streptococcal infection. Understanding the properties of this molecule during growth and infection may prove useful in sorting out the complex pathogenic properties displayed by group A streptococci.
␣-Enolase present in the cell wall of C. albicans (40) has been designated as an abundant immunodominant antigen in cases of invasive candidiasis (39,51). The fact that antibodies to SEN enhance the phagocytosis of group A streptococci of heterologous M types (Fig. 6) lends further support to the surface location of SEN and indicates that antibodies to surface molecules other than M protein are opsonic for group A streptococci, although M-specific antibodies are more effective (33). It is likely that the immune response to streptococcal ␣-enolase may also play an important role in the final outcome of a streptococcal infection in view of the fact that ␣-enolase is also present in several hematopoietic cells as a surface-expressed molecule (18). However, whether such an immune response does occur in humans during streptococcal infection is presently unknown.
The presence of ␣-enolase on the surface of streptococci and also on the surface of a variety of mammalian tissues including brain (neuron-specific enolase) (8,18) adds new insight in the role of SEN-specific antibodies in post-streptococcal autoimmune diseases such as glomerulonephritis and neurological disorders such as Sydenham's chorea, a major manifestation of rheumatic fever (53). Since anti-enolase-specific antibodies have been reported in systemic rheumatic diseases (54), and autoimmune polyglandular syndrome (55), the latter as a result of C. albicans-specific enolase, as well as cell-mediated immunity to enolase in schizophrenia (56), a possible role of SEN in post-streptococcal autoimmune diseases cannot be ruled out.
On the basis of biochemical and biological properties, SEN together with SDH/Plr (4, 6) develop an emerging theme for a new class of bacterial surface proteins. In vitro cross-protective nature of the anti-SEN antibodies, coupled with the ubiquity of SEN on the streptococcal surface, may prove useful for an immunotherapeutic intervention against streptococcal diseases. In addition to this, its potential role in autoimmune disease suggests that SEN is an important biologically active molecule with a substantial role in streptococcal pathogenesis.