INFECTION AND IMMUNITY, Oct. 1996, p. 4324–4329 Vol. 64, No. 10 0019-9567/96/$04.00ϩ0 Copyright ᭧ 1996, American Society for Microbiology

Clusterin, a Putative Complement Regulator, Binds to the Surface of Staphylococcus aureus Clinical Isolates

1 1,2 1,2 1,2 SALLY R. PARTRIDGE, * MARK S. BAKER, MARK J. WALKER, AND MARK R. WILSON Institute for Molecular Recognition1 and Department of Biological Sciences,2 University of Wollongong, Wollongong, New South Wales 2522, Australia

Received 22 April 1996/Returned for modification 12 June 1996/Accepted 23 July 1996

The ability of Staphylococcus aureus Cowan I strain and a number of S. aureus clinical isolates to bind to the human blood clusterin was investigated. Binding of clusterin to these strains was tested by both

enzyme-linked immunosorbent assay and flow cytometry. All of the S. aureus strains examined appeared to bind Downloaded from clusterin to some extent, while nonpathogenic control strains Bacillus subtilis BR151 and Escherichia coli JM109 did not. Three S. aureus isolates were selected for more detailed study; binding of labeled clusterin was saturable, inhibited in the presence of excess unlabeled clusterin, and prevented by pretreatment of bacteria with proteases. From the saturation binding studies, estimates of the affinity constants for the binding of clusterin to the bacteria ranged from 31 to 57 nM. Addition of clusterin to S. aureus cultures was also found to result in aggregation of the bacterial cells; aggregation was not detected when clusterin was added to B. subtilis BR151 or E. coli JM109 cultures. These results suggest that at least some S. aureus strains possess specific proteinaceous receptors for clusterin. Such receptors may be an important new bacterial virulence determinant for S. aureus, as clusterin has been proposed to have a role in the regulation of complement http://iai.asm.org/ activity.

Many species of pathogenic bacteria have been found to similar-size subunits linked by five disulfide bonds (3, 5). Se- possess specific cell surface receptors for various host . quence analysis predicts that each chain contains amphipathic Interactions between host proteins and bacteria may contrib- ␣-helices and heparin binding sites (5). Clusterin is highly ute to bacterial pathogenicity in a number of ways. Staphylo- conserved among mammalian species (14) and is present in coccus aureus is a major human pathogen, associated with almost all tissues, including normal human plasma (at about 50 on October 12, 2015 by UQ Library conditions ranging from localized superficial infections to se- to 100 ␮g/ml [25]). This wide tissue and species distribution vere systemic such as endocarditis, osteomyelitis, and suggests that clusterin has important functions, but its precise toxic shock syndrome (34). Various S. aureus strains have been physiological role(s) remains unknown (14). Clusterin was found to bind different host proteins, and in some cases specific originally identified as a able to cause aggregation of receptor proteins have been identified. The best known of Sertoli cells in ram testis fluid (11). Isolation of a clusterin these receptors is protein A, which binds to the Fc regions of homolog in association with high-density lipoprotein particles antibodies (10), preventing their interaction with receptors on led to the suggestion of a role in transport (6). Increased phagocytic cells and thereby protecting the bacteria from clusterin levels also seem to be associated with tissue damage phagocytosis (7). S. aureus receptors for the extracellular ma- and cells undergoing () (14). trix proteins collagen (30), fibrinogen (23), and fibronectin (9, Human clusterin was first identified as a protein found in the 15, 35) have been identified, and receptor have been SC5b-9 complex of complement (25), and a role for clusterin in cloned and sequenced. Receptor proteins binding bone sialo- protecting cells from lysis by the terminal complement mem- protein (39) and (20) have also been isolated from brane attack complex has been proposed (13). Clusterin has S. aureus strains, and an S. aureus laminin-binding protein has also been demonstrated to interact with immunoglobulins by a been identified (21). Receptors for extracellular matrix pro- multivalent mechanism, binding to both the Fc and Fab re- teins appear to be important in initial adhesion of the bacterial gions (38). cells to tissues, and differential expression of receptors for Several of the properties and proposed functions of clusterin matrix proteins with specific tissue distribution may partly ex- suggest that possession of clusterin receptors could contribute plain the tissue tropism of infections caused by various staph- to bacterial pathogenicity. We have therefore investigated ylococcal strains (29). A plasminogen receptor has also been whether S. aureus Cowan I and a selection of clinical isolates identified on S. aureus Cowan I (16), and plasmin bound to S. are capable of binding human clusterin. The S. aureus strains aureus has been found to influence bacterial adherence (17). tested appear to express cell surface receptors for clusterin, Receptors for the iron-binding proteins and trans- which are proteins or are associated with proteins and which ferrin, which contribute to host defenses by limiting the avail- mediate clusterin-induced aggregation of bacterial cells. ability of an essential microbial nutrient, have also been iso- lated from S. aureus strains (24, 26). MATERIALS AND METHODS Clusterin is a 70- to 80-kDa glycoprotein composed of two Bacterial strains. Bacillus subtilis BR151 (19) and Escherichia coli JM109 (40) were used as negative controls in clusterin binding experiments. S. aureus Cowan I, a protein A-producing strain originally isolated from a patient with septic arthritis (2), was obtained from the Australian Collection of Microorganisms, * Corresponding author. Mailing address: Department of Biological University of Queensland. All other S. aureus strains were supplied by Southern Sciences, University of Wollongong, Northfields Avenue, Wollongong, Pathology, Wollongong, New South Wales, Australia. All of the strains were NSW 2522, Australia. Phone: (61-42) 214306. Fax: (61-42) 214135. obtained from swabs of various infections as follows: 19, skin infection; 20, Electronic mail address: [email protected]. abdominal wound infection; 21, infected leg; 22, infected toe; 36b, infected ear;

4324 VOL. 64, 1996 S. AUREUS CLUSTERIN RECEPTORS 4325 and 55, infected wound. Strains were routinely grown at 37ЊCin2ϫLB medium or on LB plates (32). Preparation of bacterial cells. Single colonies were used to inoculate 10 ml of 2ϫ LB medium (32), and the cultures were incubated at 37ЊC in an orbital air shaker for approximately 16 h. Cells harvested from 1 to 10 ml of these cultures were washed three times in 1 ml of phosphate-buffered saline (PBS) containing

0.1% NaN3 and resuspended in 1 ml of PBS–0.1% NaN3. The A600softhe suspensions of washed bacteria were then determined. Immunoaffinity preparation of clusterin. Clusterin was prepared from frozen human serum packs (Red Cross Blood Bank, Sydney, New South Wales, Aus- tralia) by immunoaffinity chromatography, using a previously described mouse monoclonal anti-human clusterin antibody (G7 [25]). Serum was thawed rapidly, and EDTA, tetrasodium salt (to 10 mM), and phenylmethylsulfonyl fluoride (to at least 1 mM) were added. The chilled serum was then filtered first through Avantec Toyo no. 1 filter paper and second through a Whatman GF/C filter. The filtered serum was loaded onto a G7-Sepharose column preequilibrated with PBS containing 0.02% NaN3. After a wash with PBS–0.02% NaN3–0.5% Triton X-100 to remove and , clusterin was eluted from the col- umn with 2 M guanidine hydrochloride in PBS (pH 7.4). The eluted protein was Downloaded from dialyzed at 4ЊC overnight against PBS and passed over a protein G-Sepharose column to remove clusterin-immunoglobulin G complexes (38). The unbound fraction, which contained clusterin, was collected, concentrated by centrifugation through a Filtron Macrospec centrifugal concentrator (10-kDa cutoff), and checked for purity by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The concentration of clusterin was determined by using the bicinchoninic acid assay (36). Typically, about 10 to 12 mg of purified clusterin was obtained from 200 ml of serum. Biotinylation and FITC labeling of proteins. Clusterin preparations were di- FIG. 1. ELISA screening of bacterial strains for the ability to bind clusterin. alyzed against carbonate-bicarbonate buffer (0.1 M Na2CO3, 0.1 M NaHCO3 Samples of seven S. aureus strains, and B. subtilis BR151 and E. coli JM109 as

[pH 8.5]), while bovine serum albumin (BSA; Sigma) was dissolved directly in negative controls, bound to ELISA trays were incubated with biotinylated clus- http://iai.asm.org/ this buffer. Protein solutions were used at approximately 1 mg/ml. For biotiny- terin at 10 ␮g/ml and then with SA-HRP and OPD. Each bar represents the lation, biotinamidocaproate N-hydroxysuccinimide ester (Sigma) was added mean of six measurements, three from each of two separate cultures. The error from a stock solution (40 mg/ml in dimethyl sulfoxide) to approximately 200 ␮g bars represent 1 standard deviation of the mean. of the ester per mg of protein to be labeled. For fluorescein isothiocyanate (FITC) labeling, FITC (Sigma) was added from a solution freshly made in dimethyl sulfoxide to approximately 250 ␮g of FITC per mg of protein. For both labeling procedures, the reaction mixtures were incubated at room temperature E01, 9.000 log scale; SSC 366, 1.00, log scale; and FL1, 933, 1.00, log scale. Gating for1to2honarocking platform. Excess unbound label was removed by was used to collect data from bacterial cells and to exclude any background overnight dialysis against PBS at 4ЊC. In the case of FITC labeling, the reaction debris. The position, shape, and size of the gate were changed as required to mixture was protected from light during both the incubation and dialysis steps.

accommodate differences between the populations of S. aureus cells and those of on October 12, 2015 by UQ Library The final concentration of the labeled protein was determined by using the the control strains. Data from 10,000 events were collected for all samples. In bicinchoninic acid assay (36). some cases, samples of bacterial cells were incubated with unlabeled clusterin, ELISA. Appropriate amounts of bacterial cells were pelleted and resuspended then tissue culture supernatant from cells expressing G7 (a mouse anti-human to give an A of 3.0 in 0.3% methyl glyoxal (pH 8.0), which aids adhesion of 600 clusterin antibody [25]) or DNP9 (an isotype-matched control antibody raised bacterial cells to enzyme-linked immunosorbent assay (ELISA) plates (4). Ali- against the 2,4-dinitrophenyl group [38]), and finally FITC-conjugated sheep quots (50 ␮l) of bacterial suspensions were dispensed into ELISA plates (Dis- anti-mouse antibody (diluted 1:50 in PBS; Silenus, Melbourne, Victoria, Austra- posable Products, Adelaide, South Australia, Australia), and the plates were lia). In these cases, all of the solutions used contained rabbit immunoglobulin G incubated at 37ЊC for 90 min. After six washes with PBS in a Bio-Rad model 1250 at 250 ␮g/ml, which had been found to effectively block Fc binding of antibodies Immunowash plate washer, nonspecific binding to the wells was blocked by by S. aureus cell surface protein A (results not shown). Samples were then incubation with 1% heat-denatured casein in PBS (HDC) for 60 min at 37ЊC. analyzed as described above. Biotinylated or unlabeled clusterin in HDC was then added to the plates as Protease treatment of bacteria. Pronase (50 ␮g/ml) or trypsin (250 ␮g/ml) was required, and the plates were incubated for 60 min at 37ЊC. The plates were added to suspensions of washed bacterial cells (A adjusted to approximately washed as described above between incubations except that for competition 600 1.5 in PBS), and the suspensions were incubated at 37ЊC with shaking for 2 h. The experiments, biotinylated clusterin was added to unwashed wells preincubated cells were then pelleted, washed twice with 1 ml of PBS, resuspended in FITC- with unlabeled clusterin. Bound biotinylated clusterin was detected by addition labeled clusterin (20 ␮g/ml) to a calculated A of 3.0, incubated for 30 to 60 min of streptavidin-horseradish peroxidase conjugate (SA-HRP; 0.2 ␮g/ml in HDC; 600 at 37ЊC, washed, and analyzed by flow cytometry as described above. Sigma) and incubation for 30 min at 37ЊC. Plates were then washed once with Light microscopy. Bacterial cell samples for microscopy were resuspended in PBS–0.1% Triton X-100 (by shaking for 5 min on a plate shaker) and then 50 ␮l of PBS, BSA (100 ␮g/ml), or clusterin (5 to 100 ␮g/ml) to an A of 3.0 and washed five times with PBS–Triton X-100 (without shaking). Bound SA-HRP 600 incubated at 37ЊC for 60 min. The cells were then washed once with PBS, was detected by adding to each well 100 ␮lofo-phenylenediamine (OPD) resuspended in 100 ␮l of FACSfix, observed under a Leitz Dialux 22 microscope, reagent (2.5 mg of OPD per ml and 1 ␮l of 30% H O per ml in citrate buffer [0.1 2 2 and photographed. MNa2HPO4, 0.05 M citric acid, pH 5.0]), and the reaction was stopped after 3 to 10 min by addition of 50 ␮l of 1 M HCl per well. Plates were read immediately on a Bio-Rad model 3550 plate reader at 490 nm. Calculation of binding affinities and statistical analysis. The apparent disso- RESULTS ciation constants (Kds) for the binding of clusterin to the bacteria were estimated from curve fits obtained by using Kaleidagraph (Abelbeck Software), a nonlinear Screening of S. aureus strains for clusterin binding. Initial regression analysis program which uses the Levenberg-Marquardt algorithm. For screening of S. aureus strains for clusterin binding ability was the saturation curve, data were analyzed by fitting the equation of noncoopera- carried out by ELISA and flow cytometry. The nonpathogenic tive binding: absorbance ϭ m1/(1.0 ϩ m2/[clusterin]), where m1 is the amount of clusterin bound at saturating levels and m2 is the apparent affinity constant for bacterial strains B. subtilis BR151 and E. coli JM109 were used biotinylated clusterin in micrograms per milliliter. For the t test, the equation as negative controls, as these strains were not expected to bind used was t ϭ [m1(1) Ϫ m1(2)]/[sqrt(SE(1)∧2 ϩ SE(2)∧2)] for (n1 ϩ n2 Ϫ 4) human proteins. For ELISA, biotinylated clusterin (10 ␮g/ml degrees of freedom, where m1(1) and m1(2) are fitted values for the parameters in HDC) was added to wells containing bound bacteria or for data sets 1 and 2, SE(1) and SE(2) are the standard errors for the fitted values, and n1 and n2 are the numbers of datum points in data sets 1 and 2. previously incubated with methyl glyoxal alone. Initially, adhe- Flow cytometry. Bacterial cells prepared as for ELISA were resuspended in sion of bacterial cells to ELISA plates was checked microscop- 100 ␮l of PBS or a solution of the appropriate FITC-labeled protein to an A600 ically following staining of cells with carbol fuchsin (results not of 3.0 and incubated for 30 to 60 min at 37ЊC in the dark. The cells were then shown). Figure 1 summarizes the results of screening of the pelleted, washed twice with 1 ml of PBS, resuspended in 0.5 ml of FACSfix (1% paraformaldehyde in PBS [pH 7.4]), and stored at 4ЊC in the dark. Samples were negative control strains and seven S. aureus strains. All of the analyzed on a Becton Dickinson FACSort flow cytometer set as follows: FSC S. aureus strains tested except 55 bound much higher levels of 4326 PARTRIDGE ET AL. INFECT.IMMUN. Downloaded from http://iai.asm.org/

FIG. 2. Screening of bacterial strains by flow cytometry for the ability to bind clusterin. Samples of B. subtilis BR151 (A), E. coli JM109 (B), and S. aureus Cowan I (C), 21 (D), 36b (E), and 55 (F) were treated with BSA-FITC (20 ␮g/ml; thin lines) or clusterin-FITC (20 ␮g/ml; thick lines), washed, and analyzed by flow cytometry. on October 12, 2015 by UQ Library biotinylated clusterin than either of the two negative control strains. For flow cytometry, samples of bacteria cells were incubated in either PBS, BSA-FITC (20 ␮g/ml), or clusterin-FITC (20 FIG. 3. Binding of clusterin to strains Cowan I, 21, and 36b is saturable and ␮g/ml). Treatment of all samples with the negative control inhibited by excess unlabeled clusterin. (A) Samples of S. aureus Cowan I protein BSA-FITC resulted in levels of fluorescence similar to (squares), 21 (circles), and 36b (triangles) bound to ELISA plates were incubated with increasing concentrations of biotinylated clusterin in the range 0 to 25 those seen following incubation in PBS alone (results not ␮g/ml. Bound clusterin was detected with SA-HRP and OPD. Each datum point shown). Treatment of the B. subtilis and E. coli control strains represents the mean of three measurements, and error bars represent 1 standard with clusterin-FITC also resulted in essentially the same level deviation of the mean. (B) Samples of S. aureus Cowan I, 21, and 36b bound to of fluorescence as treatment with PBS or BSA-FITC (Fig. 2A ELISA plates were incubated with HDC (filled columns) or unlabeled clusterin (2 mg/ml; open columns). Biotinylated clusterin was then added to a final con- and B), confirming that these strains do not bind clusterin. In centration of 10 ␮g/ml. After incubation and washing, bound biotinylated clus- contrast, S. aureus Cowan I, 21, and 36b (Fig. 2C to E, respec- terin was finally detected with SA-HRP and OPD. Each datum point represents tively) and 19, 20, and 22 (results not shown) incubated with the mean of three measurements, and error bars represent 1 standard deviation clusterin-FITC all showed greater fluorescence than those in- of the mean. cubated with PBS or BSA-FITC, in agreement with the ELISA results indicating that these strains bind clusterin. S. aureus 55 appeared to bind a low level of clusterin, much less than the other S. aureus strains (Fig. 2F). The results shown are repre- Affinity of clusterin binding to selected S. aureus strains. sentative of several repeated experiments. The affinity of clusterin binding to selected S. aureus strains To confirm that it was clusterin, rather than a colabeled was investigated by ELISA. Increasing concentrations of bio- minor contaminant, binding to the bacteria, unlabeled clus- tinylated clusterin in the range 0 to 25 ␮g/ml were added to terin in conjunction with a clusterin-specific monoclonal anti- samples of strains Cowan I, 21, and 36b bound to ELISA body (G7 [25]) was used. Specific binding of unlabeled clus- plates. Figure 3A shows that all three strains show saturable terin to S. aureus Cowan I and other S. aureus strains which binding of clusterin, suggesting that clusterin binding to S. bound biotinylated clusterin was confirmed (results not aureus strains is mediated by specific receptors. The results shown). shown are representative of several repeated experiments. These results indicate that some S. aureus strains are capable Analysis of the data indicated that there is a single class of of binding clusterin and that ELISA and flow cytometry are clusterin binding sites on each strain, and estimated Kd values suitable methods for screening strains of pathogenic bacteria were 31.50 Ϯ 13.00, 38.75 Ϯ 19.25, and 56.13 Ϯ 22.00 nM for for the ability to bind host proteins. Strains Cowan I, 21, and strains Cowan I, 21, and 36b, respectively. When analyzed 36b were selected for further study of clusterin binding inter- statistically, there appeared to be no significant difference be- actions. tween the Kd values for the three strains, although strains VOL. 64, 1996 S. AUREUS CLUSTERIN RECEPTORS 4327 Downloaded from http://iai.asm.org/

FIG. 4. Binding of FITC-labeled clusterin to selected S. aureus strains is inhibited by pretreatment of bacteria with proteases. Samples of S. aureus Cowan I (A and B), 21 (C and D), and 36b (E and F) were treated with trypsin (250 ␮g/ml; A, C, and E) or pronase (50 ␮g/ml: B, D, and F). After being washed, these samples were then incubated with clusterin-FITC (dotted lines) and com- pared with untreated samples incubated with clusterin-FITC (20 ␮g/ml; thick lines) or BSA-FITC (20 ␮g/ml: thin lines) by flow cytometry. on October 12, 2015 by UQ Library

Cowan I and 21 appear to have more clusterin binding sites per cell than strain 36b. To confirm that interaction of clusterin with these staphylo- FIG. 5. Clusterin-mediated aggregation of S. aureus cells. Samples of cells coccal strains is specific, cells of strains Cowan I, 21, and 36b from cultures of S. aureus Cowan I were incubated in clusterin (100 ␮g/ml in bound to ELISA plates were preincubated with excess unla- PBS; A) or BSA (100 ␮g/ml in PBS; B) and examined microscopically, using a beled clusterin, and then biotinylated clusterin was added to a 100ϫ objective lens and 12.5ϫ eyepiece. final concentration of 10 ␮g/ml. Figure 3B shows that excess unlabeled clusterin inhibits binding of biotinylated clusterin to these three strains. The results shown are representative of cytometry, we observed that samples of bacterial cells incu- several repeated experiments. bated with various concentrations of clusterin contained par- Effect of protease treatment on binding of clusterin to se- ticles with increased forward scatter and side scatter compared lected S. aureus strains. The results of saturation and compe- with samples incubated in PBS or similar concentrations of tition studies suggested that clusterin binds to S. aureus Cowan BSA (results not shown). A possible explanation for this effect I, 21, and 36b via specific receptors. The effect of protease was that clusterin causes aggregation of cells of these strains. treatment on binding of FITC-clusterin to these three strains To further investigate this effect, bacteria that had been incu- was therefore investigated to determine whether these recep- bated in PBS or in various concentrations of BSA or clusterin tors could be proteins. A similar method has previously been were examined by light microscopy. Cell aggregates were seen used to demonstrate that a proteinaceous receptor is respon- in S. aureus Cowan I samples that had been incubated in sible for plasminogen binding to S. aureus Cowan I (16). Sam- clusterin (Fig. 5A) but not in samples incubated in PBS alone ples of bacterial cells that had been preincubated with pronase (results not shown) or in PBS containing similar concentra- or trypsin, washed, and incubated with FITC-labeled clusterin tions of BSA (Fig. 5B). The aggregating ability of clusterin were analyzed by flow cytometry. Figure 4 shows that such appeared to correlate with clusterin binding ability of the cells, treatment of S. aureus Cowan I, 21, and 36b with either pro- as B. subtilis BR151 and E. coli JM109 did not form clumps, tease virtually abolished clusterin binding. Pretreatment of even at high clusterin concentrations (up to 2 mg/ml; results these same strains with PBS without added protease had no not shown). significant effect on clusterin binding (results not shown). These results suggest that the putative clusterin receptors ex- DISCUSSION pressed by these strains are proteins or are associated with proteins. Using both biotinylated clusterin in an ELISA and FITC- Clusterin-mediated aggregation of S. aureus cells. While labeled clusterin in flow cytometry, we have shown that at least studying the binding of clusterin to S. aureus strains by flow some S. aureus strains are able to bind human clusterin, 4328 PARTRIDGE ET AL. INFECT.IMMUN. whereas two nonpathogenic species do not. For three strains IV collagen can also mediate aggregation of S. aureus cells (Cowan I, 21, and 36b) investigated in more detail, this binding (37). Such aggregation may allow bacteria to elude phagocytic appears to be mediated by specific receptors, which are pro- host defenses, and it has been demonstrated that the ability of teins or are associated with proteins. The apparent dissociation S. aureus strains to aggregate in the presence of fibronectin constant for the binding of biotinylated clusterin to these re- correlates with invasiveness (31). ceptors was estimated to be in the range 31 to 57 nM. The It is clearly possible that the ability of S. aureus strains to levels of clusterin found in normal human plasma (approxi- bind clusterin could contribute to the virulence of these strains. mately 625 to 1,250 nM [25]) are thus well above the concen- Binding of other host proteins via specific receptors has been tration required to saturate the receptors, indicating that clus- found to correlate with increased virulence in a number of terin binding could be physiologically relevant. Recently, a cases. For example, inactivation of a S. aureus encoding a Streptococcus pyogenes secreted protein (SIC protein) which collagen receptor has been found to result in a dramatic re- binds to clusterin has been identified (1); dot blot analysis duction in virulence in a mouse model of septic arthritis (28). indicates that some clusterin binding activity may also be Furthermore, a strain of S. aureus which demonstrated de- present in media in which S. aureus Cowan I has been grown creased adherence to immobilized fibronectin, as a result of (27). transposon mutagenesis, was found to be less virulent than the Previous workers have generally investigated binding of host parental strains in a rat model of infective endocarditis (18). In Downloaded from proteins to bacterial cells by incubating the cells with 125I- the case of the streptococcal SIC protein, it was found that only labeled proteins and measuring the radioactivity associated certain group A streptococcal strains, those that are associated with cell pellets. The ELISA and flow cytometric methods that with severe and toxic infections, contained and expressed the we have used have certain advantages over this method: bio- sic gene, suggesting that the SIC protein may influence patho- tinylated and FITC-labeled proteins are simpler and safer to genicity (1). prepare than iodinated proteins, the 96-well layout of ELISA In addition to investigating the possible roles of cell surface- plates allows easy simultaneous screening of many strains, and bound clusterin in bacterial pathogenicity, experiments aimed flow cytometry gives information on individual cells rather than at identifying and isolating clusterin receptor proteins and cell populations, allowing the possibility of analyzing subpopu- cloning and sequencing the corresponding genes are currently http://iai.asm.org/ lations of cells. We have also used these methods to demon- under way. The availability of cloned clusterin receptor genes strate that some bacterial clinical isolates bind plasminogen, in and corresponding sequence data would provide the basis for agreement with previous results (22), and to screen strains the creation of mutations in clusterin receptor genes and for other than S. aureus for clusterin binding ability (27). studying the effects of these mutations on pathogenicity. The Although these studies indicate that some S. aureus strains results of these studies, plus the availability of purified clus- appear to express clusterin receptors, we have not yet deter- terin receptor proteins, may also provide information useful mined how possession of such receptors could influence bac- for designing new strategies to combat bacterial infections. The terial pathogenicity. An obvious starting point would be the characterized receptors could provide potential targets for the on October 12, 2015 by UQ Library proposed role of clusterin as a regulator of complement activ- action of new drugs (29). Alternatively, the receptor itself ity. The recently identified clusterin-binding SIC protein was could be used as a vaccine component, which not only would found to be incorporated into the terminal C5b-C9 comple- act as an antigen in the normal immune response but also ment complexes formed in serum and to inhibit complement- could block receptor-ligand binding. Such strategies have al- mediated lysis of erythrocytes (1). As gram-positive strains, ready been tested by using other bacterial receptors for host such as S. aureus and S. pyogenes, are resistant to lysis by the proteins, particularly receptors for extracellular matrix pro- complement membrane attack complex, clusterin bound to the teins (e.g., reference 33). In the face of increased bacterial bacterial cell surface apparently would have to influence other resistance to traditional antibiotics, these therapeutic strate- aspects of clusterin action to affect pathogenicity. Åkesson et gies are likely to become increasingly important in fighting al. (1) suggested that SIC could influence the host-pathogen bacterial infections. relationship by affecting other functions of the membrane at- tack complex, such as stimulation of the production and re- ACKNOWLEDGMENTS lease of inflammatory mediators. Another aspect of antibacte- rial complement activity which could be influenced by clusterin S.R.P. was supported by a postdoctoral fellowship from the Institute is opsonization. Deposition of complement components such for Molecular Recognition, University of Wollongong. We thank David Jackson of Southern Pathology, Wollongong, for as C3b on the bacterial surface results in enhanced phagocy- supplying strains of pathogenic bacteria and Simon Easterbrook- tosis of both gram-negative and gram-positive bacteria. Clus- Smith, University of Sydney, for help in calculating binding affinities. terin bound to specific receptors on the bacterial cell surface could prevent deposition of opsonizing molecules on the bac- REFERENCES teria, providing protection against complement-mediated op- 1. Åkesson, P., A. G. Sjo¨holm,and L. Bjo¨rck. 1996. Protein SIC, a novel sonization. Pathogenic bacteria are known to have mechanisms extracellular protein of Streptococcus pyogenes interfering with complement for resisting the opsonizing effects of complement, for exam- function. J. Biol. Chem. 271:1081–1088. ple, the antiphagocytic M proteins of streptococci (8). 2. Ankerst, J., P. Christensen, L. Kjelle´n, and G. Kronvall. 1974. A routine diagnostic test for IgA and IgM antibodies to rubella virus: absorption of IgG It is also possible that the cell aggregating action of clusterin with Staphylococcus aureus. J. Infect. Dis. 130:268–273. can influence bacterial pathogenicity. Clusterin has previously 3. Choi-Miura, N. H., Y. Takahashi, Y. Nakano, T. Tobe, and M. Tomita. 1992. been shown to cause aggregation of various types of eukaryotic Identification of the disulphide bonds in human plasma protein SP-40, 40 cells when suspended in PBS, but the effect was not reproduc- (Apo J). J. Biochem. 112:557–561. 4. Czerkinsky, C., A. S. Rees, L. A. Bergmeier, and S. J. Challacombe. 1983. ible when the cells were suspended in solutions containing The detection and specificity of class specific antibodies to whole bacterial other serum proteins in addition to clusterin (11). It has been cells using a solid phase radioimmunoassay. Clin. Exp. Immunol. 53:192–200. known for some time that S. aureus cells form clumps in the 5. de Silva, H. V., J. A. K. Harmony, W. D. Stuart, C. M. Gil, and J. Robbins. presence of plasma, and this effect can mainly be attributed to 1990. J: structure and tissue distribution. Biochemistry 29: 5380–5389. interaction of fibrinogen with an S. aureus fibrinogen receptor 6. de Silva, H. V., W. D. Stuart, Y. B. Park, S. J. T. Mao, C. M. Gil, J. R. known as clumping factor (12). Fibronectin, laminin, and type Wetterau, S. J. Busch, and J. A. K. Harmony. 1990. Purification and char- VOL. 64, 1996 S. AUREUS CLUSTERIN RECEPTORS 4329

acterization of apolipoprotein J. J. Biol. Chem. 265:14292–14297. transferrin-binding protein. Infect. Immun. 62:3850–3858. 7. Dossett, J. H., G. Kronvall, R. C. Williams, and P. G. Quie. 1969. Antiph- 25. Murphy, B. F., L. Kirszbaum, I. D. Walker, and A. J. F. d’Apice. 1988. SP-40, agocytic effects of staphylococcal protein A. J. Immunol. 103:1405–1410. 40, a newly identified normal human serum protein found in the SC5b-9 8. Fischetti, V. A. 1991. Streptococcal M protein. Sci. Am. 264(June):32–40. complex of complement and in the immune deposits in . 9. Flock, J.-I., G. Fro¨man,K. Jo¨nsson,B. Guss, C. Signa¨s, B. Nilsson, G. J. Clin. Invest. 81:1858–1864. Raucci, M. Ho¨o¨k,T. Wadstro¨m,and M. Lindberg. 1987. Cloning and ex- 26. Naidu, A. S., J. Miedzobrodzki, J. M. Musser, V. T. Rosdahl, S.-Å. Hed- pression of the gene for a fibronectin-binding protein from Staphylococcus stro¨m,and A. Forsgren. 1991. Human lactoferrin binding in clinical isolates aureus. EMBO J. 6:2351–2357. of Staphylococcus aureus. J. Med. Microbiol. 34:323–328. 10. Forsgren, A., and J. Sjo¨quist. 1966. “Protein A” from S. aureus. I. Pseudo- 27. Partridge, S. R. Unpublished data. immune reaction with human ␥-globulin. J. Immunol. 97:822–827. 28. Patti, J. M., T. Bremell, D. Krawjewska-Pietrasik, A. Abdelnour, T. 11. Fritz, I. B., K. Burdzy, B. Setchell, and O. Blaschuk. 1983. Ram rete testes Tarkowski, C. Ryde´n, and M. Ho¨o¨k. 1994. The Staphylococcus aureus colla- fluid contains a protein (clusterin) which influences cell-cell interactions in gen adhesin is a virulence determinant in experimental septic arthritis. In- vitro. Biol. Reprod. 28:1173–1188. fect. Immun. 62:151–161. 12. Hawiger, J., S. Timmons, D. D. Strong, B. A. Cottrell, M. Riley, and R. F. 29. Patti, J. M., and M. Ho¨o¨k. 1994. Microbial adhesins recognizing extracellular Doolittle. 1982. Identification of a region of human fibrinogen interacting matrix macromolecules. Curr. Opin. Cell Biol. 6:752–758. with staphylococcal clumping factor. Biochemistry 21:1407–1413. 30. Patti, J. M., H. Jonsson, B. Guss, L. M. Switalski, K. Wiberg, M. Lindberg, 13. Jenne, D. E., and J. Tschopp. 1989. Molecular structure and functional and M. Ho¨o¨k. 1992. Molecular characterization and expression of a gene characterization of a novel human cytolysis inhibitor (CLI) found in blood encoding a Staphylococcus aureus collagen adhesin. J. Biol. Chem. 267:4766– and seminal plasma. Identity to sulfated glycoprotein 2 (SGP-2), a constit- 4772. Downloaded from uent of rat testis fluid. Proc. Natl. Acad. Sci. USA 86:7123–7127. 31. Proctor, R. A., G. Christman, and D. F. Mosher. 1984. Fibronectin-induced 14. Jenne, D. E., and J. Tschopp. 1992. Clusterin: the intriguing guises of a agglutination of Staphylococcus aureus correlates with invasiveness. J. Lab. widely expressed glycoprotein. Trends Biochem. Sci. 17:154–159. Clin. Med. 104:455–469. 15. Jo¨nsson,K., C. Signa¨s, H. P. Muller, and M. Lindberg. 1991. Two different 32. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a genes encode fibronectin binding proteins in Staphylococcus aureus. The laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold complete nucleotide sequence and characterization of the second gene. Eur. Spring Harbor, N.Y. J. Biochem. 202:1041–1048. 33. Schennings, T., A. Heimdahl, K. Coster, and J. I. Flock. 1993. Immunisation 16. Kuusela, P., and O. Saksela. 1990. Binding and activation of plasminogen at with fibronectin binding protein from Staphylococcus aureus protects against the surface of Staphylococcus aureus. Increase in affinity after conversion to experimental endocarditis in rats. Microbiol. Pathol. 15:227–236. the Lys form of the ligand. Eur. J. Biochem. 193:759–765. 34. Sheagren, J. N. 1984. Staphylococcus aureus. The persistent pathogen. 17. Kuusela, P., M. Ullberg, G. Kronvall, T. Tervo, A. Tarkkanen, and O. N. Engl. J. Med. 310:1368–1373. Saksela. 1992. Surface-associated activation of plasminogen on gram-posi- 35. Signa¨s, C., G. Raucci, K. Jo¨nsson,P.-E. Lindgren, G. M. Anantharamaiah, http://iai.asm.org/ tive bacteria. Effect of plasmin on the adherence of Staphylococcus aureus. M. Ho¨o¨k,and M. Lindberg. 1989. Nucleotide sequence of the gene for a Acta Ophthalmol. 70:42–46. fibronectin binding protein from Staphylococcus aureus: use of this peptide 18. Kuypers, J. M., and R. A. Proctor. 1989. Reduced adherence to traumatized sequence in the synthesis of biologically active peptides. Proc. Natl. Acad. rat heart valves by a low-fibronectin-binding mutant of Staphylococcus au- Sci. USA 86:699–703. reus. Infect. Immun. 57:2306–2312. 36. Smith, P. K., R. I. Krohn, G. T. Hermanson, A. K. Mallia, F. H. Gartner, 19. le Grice, S. F. J. 1990. Regulated promoter for high-level expression of M. D. Provenzano, E. K. Fujimoto, N. M. Goeke, B. J. Olson, and D. C. heterologous genes in Bacillus subtilis. Methods Enzymol. 185:201–214. Klenk. 1985. Measurement of protein using bicinchoninic acid. Anal. Bio- 20. Liang, O. D., J.-I. Flock, and T. Wadstro¨m. 1995. Isolation and characteri- chem. 150:76–85. sation of a vitronectin-binding surface protein from Staphylococcus aureus. 37. Vercellotti, G. M., J. B. McCarthy, P. Lindholm, P. K. Peterson, H. S. Jacob,

Biochim. Biophys. Acta 1250:110–116. and L. T. Furcht. 1985. Extracellular matrix proteins (fibronectin, laminin on October 12, 2015 by UQ Library 21. Lopes, J. D., M. dos Reis, and R. R. Brentani. 1985. Presence of laminin and type IV collagen) bind and aggregate bacteria. Am. J. Pathol. 120:13–21. receptors in Staphylococcus aureus. Science 229:275–277. 38. Wilson, M. R., and S. B. Easterbrook-Smith. 1992. Clusterin binds by a 22. Lottenberg, R., D. Minning-Wenz, and M. D. P. Boyle. 1994. Capturing host multivalent mechanism to the Fc and Fab regions of IgG. Biochim. Biophys. plasmin(ogen): a common mechanism for invasive pathogens? Trends Mi- Acta 1159:319–326. crobiol. 2:20–24. 39. Yacoub, A., P. Lindahl, K. Rubin, M. Wendel, D. Heinegård, and C. Ryde´n. 23. McDevitt, D., P. Francois, P. Vaudaux, and T. J. Foster. 1994. Molecular 1994. Purification of a bone sialoprotein-binding protein from Staphylococ- characterization of the clumping factor (fibrinogen receptor) of Staphylococ- cus aureus. Eur. J. Biochem. 222:919–925. cus aureus. Mol. Microbiol. 11:237–248. 40. Yanisch-Peron, C., J. Vieira, and J. Messing. 1985. Improved M13 phage 24. Modun, B., D. Kendall, and P. Williams. 1994. Staphylococci express a cloning vectors and host strains; nucleotide sequence of the M13mp18 and receptor for human transferrin: identification of a 42-kilodalton cell wall pUC19 vectors. Gene 33:103–119.

Editor: V. A. Fischetti