Microbial Pathogenesis 1995; 19: 49–55

Opsonization of Staphylococcus aureus with afibronectin-binding protein antiserum inducesprotection in mice

Wubshet Mamo1*, Per Jonsson2 and Hans-Peter Muller3

1Section of Bacteriology and Epizootology, Department of Veterinary Microbiology,2Section of Clinical Microbiology and 3Department of Microbiology, SwedishUniversity of Agricultural Sciences, Uppsala, Sweden

(Received December 29, 1994; accepted in revised form June 27, 1995)

—––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––Mamo, W. (Section of Bacteriology and Epizoology, Department of Veterinary Microbiology,Swedish University of Agricultural Sciences, Uppsala, Sweden), P. Jonsson and H.-P. Muller.Opsonization of Staphylococcus aureus with a fibronectin-binding protein antiserum inducesprotection in mice. Microbial Pathogenesis 1995; 19: 49–55.

The virulence of Staphylococcus aureus opsonized with an antiserum raised against arecombinant fibronectin-binding protein (FnBP) was compared with homologous, non-opson-ized bacteria (treated with pre-immune serum) in a mouse mastitis model. Virulence wasevaluated comparing the number of bacteria recovered from the infected mammary glandsand according to the type of lesions produced. The average number of bacteria recoveredfrom the mammary glands inoculated with S. aureus opsonized with FnBP-antiserum wassignificantly lower (up to 107 cfu/ml) than the average number of bacteria recovered afterinoculation with non-opsonized bacteria (up to 1010 cfu/ml). Gross examination of infectedmammary glands showed that 65% of glands infected with opsonized bacteria developed lowgrade/or had no pathological changes, and 35% developed severe mastitis whereas, 75% ofglands inoculated with non-opsonized bacteria developed severe mastitis and 25% low grademastitis or had no pathological changes. According to the histopathological examination eightout of 10 glands inoculated with opsonized bacteria produced disseminated focal necrosis orhad no pathological changes and two glands produced non reactive necrotic lesions. Incontrast, only three out of 10 glands inoculated with non-opsonized homologous bacteriadeveloped disseminated focal necrosis and had no pathological changes while seven glandsdeveloped total necrosis. © 1995 Academic Press Limited

Key words: Staphylococcus aureus; Opsonization; Fibronectin-binding protein; Virulence;Animal Model; Mastitis; Antibodies; Fusion proteins; Vaccines.—––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Introduction

Adhesins present on Staphylococcus aureus surface are thought to mediate attach-ment of bacteria to host cells which may initiate of a disease.1,2 Thus, adhesins on S.aureus involved in pathogenesis could be important in the development of subunitvaccines against S. aureus infections.3 Fibronectin-binding proteins (FnBPs) presenton the surface of S. aureus have been reported to promote adhesion of bacteria tohost tissue and therefore have been considered as virulence factors.4,5

* Author to whom correspondence should be addressed: Dr Wubshet Mamo, Astra Arcus, Pre clinicalR & D , Section of Immunology, All, S-151 85 Sodertalje, Sweden. Phone: +46 8 553 28342; Fax: +46 8 55327060; E-mail: Wubshet.MamoÝarcus.se.astra.com.

0882-4010/95/070049+07 $12.00/0 = 1995 Academic Press Limited

W. Mamo et al.50

In case of intramammary infection (mastitis), adhesion of udder pathogens toepithelial cells in the mammary gland has been considered to be a prerequisite forthe initiation of the disease.6 Most S. aureus isolates specifically bind to extracellularmatrix fibronectin.7 FnBPs have been shown to mediate binding of S. aureus tosoluble fibronectin.8–10 Moreover, two genes for FnBPs of S. aureus have been cloned,sequenced and the encoded proteins designated as FnBPA and FnBPB have beenpurified and characterized.8,11 Applying a mouse mastitis model in recent studies,recombinant FnBPs from S. aureus (fusion proteins) used as immunogens inducedprotection in mice against challenge infection with S. aureus.12 However, in thesestudies, a relatively weak antibody response was recorded which could not be cor-related to the degree of protection observed.

In the present study the effect of opsonization of S. aureus with a rabbit hyp-erimmune serum against a recombinant FnBP on virulence in mice compared tobacteria treated with pre-immune serum is reported.

Results

In the present investigation a rabbit antiserum raised against a fusion protein rep-resenting the D-domain of the S. aureus FnBPA was used to opsonize S. aureus invitro. To determine the titre of this antiserum against native FnBPs without inter-ference of protein A, affinity purified FnBPs were used as antigen in microtitre plates.As demonstrated in Fig. 1a, the reaction to the antiserum against the fusion proteinwas very similar to the interaction of two rabbit antisera raised against native FnBPs.All three hyperimmune sera showed a specific titre of ×1: 6000 whereas the pre-immune serum was negative at all dilutions. As shown in Fig. 1b, the antiserumagainst the fusion protein used in this study had also a high titre against the b-galactosidase part encoded by a gene derived from Escherichia coli. Bacteria pre-treated with either anti gal-FnBPA serum or with pre-immune serum (control) wereadministered into the mammary glands of mice. Virulence was studied by comparingthe effects of opsonized bacteria and control bacteria in the glands.

Table 1 shows the average number of bacteria recovered in glands infected withthree different concentrations (104, 105, and 106 cfu/ml) of opsonized or non-opson-ized bacteria after 18 h of incubation. The number of bacteria recovered from theglands inoculated with non-opsonized bacteria was up to 100-fold higher comparedto the number of bacteria recovered from glands inoculated with opsonized bacteria.

Forty eight hours after inoculation of opsonized or non-opsonized bacteria, mam-mary glands of mice were dissected and subjected to histopathological exam-inations. As shown in Table 2, inoculation of opsonized bacteria induced about 35%gross/medium grade mastitis whereas up to 75% of glands inoculated with non-opsonised bacteria developed gross/medium grade mastitis. Histopathologicalexamination revealed that only two out of 10 glands inoculated with opsonizedbacteria had developed non-reactive total necrosis while eight glands showed dis-seminated focal necrosis or had no pathological changes. In contrast, seven out of10 glands inoculated with non-opsonized bacteria showed non-reactive totalnecrosis.

Discussion

Adhesion of bacteria to host epithelial cells is widely accepted as a primary step inthe successful colonization of a host.1,2 Most S. aureus strains bind specifically tofibronectin and FnBPs on the surface of the bacteria have been shown to be respon-

Opsonization of Staphylococcus aureus 51

Fig. 1. Evaluation of the antibody titre of rabbit anti gal-FnBP serum against native FnBP from Staphy-lococcus aureus (A) and against b-galactosidase (B). Closed circles: gal-FnBP antiserum; open trianglesand squares; antiserum raised against native FnBP from S. aureus; cross symbols: rabbit preimmuneserum, included as control.

W. Mamo et al.52

Table 1 Average number of S. aureus recovered1 in the micemammary gland inoculated with bacteria opsonized2 with FnBPantiserum or non-opsonized.—––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

No. of S. aureus No. of S. aureusS. aureus inoculated recovered Significancestrain SA 113 (log cfu/ml) (log cfu2SD) level3

—––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––Opsonized 4.0 5.220.62 NS

5.0 6.020.80 NS6.0 7.020.42 NS

Non-opsonized 4.0 6.020.22 *5.0 8.620.35 *6.0 9.820.22 **

—––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––1 Each value represents the mean2SD of five glands.2 Opsonization of bacteria was done as described in Materials and Methods.3 NS = non significant, * = P ¾ 0.05; ** = P ¾ 0.01

Table 2 Gross and histopathological examination of lesions in the mammary glandsof mice inoculated with S. aureus opsonized with anti-FnBP antiserum (or non-opsonized).—–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Gross examination2 Histopathological examination3

— —–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––Type of lesions Type of lesions

Inoculating No. of (%) No. of (No)Inoculated doses glands —–––––––––––––––––– glands —––––––––––––––––––––––––S. aureus (bacteria/ml) examined (+++/++) (+/0) examined A B C 0—–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––Opsonized1 105 18 35 65 10 2 0 2 6Non-opsonized 105 20 75 25 10 7 0 2 1—–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

1 Opsonization of bacteria was done as described in Materials and Methods.2 Gross examination was scored as described by Haraldsson and Jonsson.25 (+++/++ = gross/medium

mastitis; +/0 = low grade/or no pathological change).3 Histopathological examination was recorded according to Haraldsson and Jonsson.25 (A = non-reactive

total necrosis; B = advanced regressive and mild inflammatory reaction; C = disseminated and focal necrosisand 0 = no pathological reaction).

sible for the binding. Studies by Kyupers and Proctor4 revealed that fibronectin couldmediate adhesion of S. aureus to traumatized rat heart valves. Mutants, low infibronectin-binding were recovered in significant lower numbers from the heartvalves compared with the parent strain and suggested that the ability of S. aureusto bind to fibronectin was an important factor in establishing adherence in vivo. Luket al.,13 indicated that antibodies against recombinant fibronectin-binding proteinscan block binding of bacteria to immobilized fibronectin in vitro. Attachment of S.aureus to fibronectin-containing plasma clots could also be inhibited by recombinantproteins containing the fibronectin-binding region of FnBP from S. aureus as well asby synthetic peptides mimicking the 38 amino acid repeats of the fibronectin-bindingregion.14

We have recently shown in a mouse mastitis model that active immunization ofmice with recombinant FnBPs12 as well as with fibrinogen binding proteins (FgBPs)15

evoked a protective immunoresponse resulting in significant reduction of mastitis.While these results supported the important role of FnBps and FgBPs of S. aureus

Opsonization of Staphylococcus aureus 53

as immunogens, the relatively low antibody titres against FnBPs recorded in activelyimmunized mice could not be related to the protection observed.12

In the present study the significantly higher number of bacteria recovered in themammary glands of mice inoculated with non-opsonized S. aureus compared to thenumber recovered in the glands of mice inoculated with opsonized bacteria indicatethat antibodies against FnBP can provoke protection against S. aureus. The mech-anism by which protection was achieved is not quite clear. The lower number ofbacteria recovered from mammary glands after pretreatment of the bacteria withhyperimmune serum suggests that opsono-phagocytosis may play an importantrole. Antibodies against FnBPs may, however, also directly block the adherence ofthe bacteria to host tissue via a fibronectin mediated mechanism. Adhesion blockingactivity in S. aureus antiserum has been described using bovine mammary epithelialcells as target.16 Adhesion blocking activity in the FnBP hyperimmune rabbit anti-serum, although not explicitly shown in this study, would be supported by the resultsfrom the histopathological examination of infected glands. Bacteria opsonized withFnBP antiserum caused significantly less severe mastitis compared to non-opsonizedbacteria (Table 2). The protective effect of the antiserum used in this study deservesrecognition. A recombinant protein representing a relatively small fragment of asingle S. aureus surface protein was used for immunization of rabbits which resultedin a high antibody titre when assayed against native S. aureus FnBP in ELISA.Moreover, the antiserum recognized the corresponding surface structure on a het-erologous strain. Considering the large number of proteins exposed on the sta-phylococcal surface it is remarkable that an antiserum recognizing a limited area ofa single surface protein provokes (as shown in this study) a significant protection inan in vivo model. Recent studies by other groups indicated that antibodies to capsularpolysaccharides of S. aureus were not protective when rats were challenged with ahomologous strain in an endocarditis model17 and likewise, S. aureus protein Aantiserum, passively administered to rats failed to modify any stage in the devel-opment of bacteremic infection following subcutaneous challenge with a homo-logous strain.18 Moreover, antibodies against a recombinant collagen binding proteinfrom S. aureus did not show protection in a mouse peritonitis model19 whereas, inthe same model antibodies against recombinant FnBPs were in agreement withresults presented herein.

Our finding further support the importance of Fn-binding to S. aureus as a virulencemechanism and indicate that FnBP may be a crucial target for antibodies to reduceS. aureus infections. Following the relative low immunoresponse against S. aureuswhole-cell vaccines it has been argued that important structures involved in S. aureusadhesion might be insufficiently available on such vaccine preparations.16 This prob-lem is likely to be overcome by molecular approaches making important epitopes ofS. aureus virulence factors available in form of recombinant proteins. Our resultssuggest that recombinant FnBPs may be promising candidates for the induction ofprotective antibodies against bovine intramammary infections caused by S. aureus.Antibodies raised against FnBPs may also be suitable for passive immunizationagainst certain hospital-acquired S. aureus infections in patients with high sus-ceptibilities to infections.

Materials and methods

Chemicals. Trypticase soy broth (TSB) was purchased from Difco Chemicals, Detroit, Michi-gan, b-galactosidase was purchased from Boehringer Mannheim, Mannheim, Germany. Allother chemicals were of analytical grade.

W. Mamo et al.54

Bacteria. Staphylococcus aureus strain SA113 used throughout the experiments had beenearlier characterized and extensively studied in a mouse mastitis model.20 Bacteria werecultured at 37°C for 18 h in TSB, then washed three times in phosphate buffered saline (PBS),pH 7.4 and adjusted to the desired number of bacteria (cfu/ml) in peptone water.

Antisera. Polyclonal antiserum was raised in rabbits against a recombinant fibronectinbinding protein. The antigen had been produced as earlier described21 and represented afusion protein consisting of the first 3 repeat units of the D-domain of FnBPA from S. aureus,strain 8325-422 as a C-terminal extension of a 120 kDa b-galactosidase part. The rabbits hadbeen immunised as recently described.19 Antisera against wild-type FnBPs was obtained fromrabbits immunized four times in intervals of 2 weeks with 50 mg of FnBPs (from S. aureusNewman), mixed with Freund’s incomplete adjuvant.

ELISA. Microtitre plates (maxisorp, transparent; Nunc, Denmark) were coated with nativeFnBP, isolated from the culture supernatant of a spontaneous mutant of S. aureus, strainNewman, as previously described by Jonsson et al.11 After coating overnight at 4°C the plateswere washed three times with PBS containing 0.1% Tween 20 and the wells were saturatedfor 1 h using the same buffer. Sera used in this study were from a rabbit immunized with therecombinant galFnBP as well as from two rabbits immunized with the native full length FnBPsand from a non-immunized rabbit (control). The sera were then two-fold serially diluted,starting with an initial dilution of 1:100 and of each dilution 100 ml were applied to the wells.After reaction for 2 h at room temperature, plates were washed with PBS-Tween 20 and 100ml/well of a horseradish peroxidase-protein A conjugate,23 diluted 1:1000 was added to thewells and allowed to react for 2 h. After final washings the plate was developed using thefluorogenic substrate HPPA and the plate was then measured in a fluoroscan II reader (Lab-systems, Finland).

For evaluation of antibodies against b-galactosidase in the antiserum raised against the gal-FnBP fusion protein, while fluoroplates (Maxisorp, Nunc, Denmark) were coated with 5 mg/mlof b-galactosidase and otherwise treated as described above.

Animals. Mice originally of the BSVS (Bacterial Sensitive Virus Sensitive)-strain,24 female,2–3 months old and weighing 25–30 g were used throughout the experiments. Intramammaryinoculation was done following their first pregnancy at day 10–15 of lactation as describedearlier.25

Opsonization of bacteria. One millilitre of polyclonal antiserum and 1 ml of bacteria of adesired concentration were incubated for 30 min at 37°C in a water bath. Following incubationthe suspension was washed twice, and resuspended to the initial concentration in peptonewater and subjected to the infection assay. Bacteria treated as above with pre-immune rabbitserum were used as controls and referred to as non-opsonized.

Virulence assay. Mice were infected via the teat duct by intramammary administration of106, 105, and 104 cfu/ml into the left 4th (L-4) and right 4th (R-4) mammary glands. Forty-eighth later, mice were killed, dissected and gross examination of pathological changes on L-4 andR-4 glands was recorded.

Recovery of bacteria in the mammary glands. The L-4 glands were dissected and homo-genized separately in 5 ml physiological saline to an even suspension. Tenfold serial dilutionsof homogenates were prepared and 25 ml of each dilution was plated on surface-dried bloodagar plates. Colonies were counted after 24 h incubation at 37°C and the calculated number ofbacteria (cfu/ml) in each gland infected with opsonized bacteria vs. glands inoculated withnon-opsonized bacteria was compared.

Gross and histopathological examination of the infected glands. The macroscopic appear-ance (gross examination) of inoculated glands was evaluated according to the type of lesionsobserved, as described by Haraldsson and Jonsson.25 For the histopathological examinationthe R-4 mammary glands were taken out and the type of lesions produced in the infectedglands tissue were evaluated and recorded.25

Statistical analysis. The results from intramammary growth were evaluated by the two-tailed Student’s t-test.

Opsonization of Staphylococcus aureus 55

We acknowledge Dr Barbara Rozalska, Dept. of Infectious Biology, University of Lodz, Polandfor supplying us with gal-FnBP antiserum and Dr Liisa Rantamaki, Dept. of Microbiology andEpizootology, College of Veterinary Medicine, Helsinki, Finland for supplying us with rabbitantisera against the wild type FnBPs.

References

1. Beachey EH. Bacterial adherence. Adhesion-receptor interactions mediating the attachment of bac-teria to mucosal surfaces. J Infect Dis 1981; 143: 325–45.

2. Christensen GD, Simpson WA, Beachey EH. Adhesion of bacteria to animal tissues. In: Savage DCand Fletcher M eds. Bacterial adhesion. Plenum Press, New York, 1985; 279–305.

3. Foster TJ. Potentials for vaccination against infections caused by Staphylococcus aureus. Vaccine1991; 16: 159–66.

4. Kuypers JM, Proctor RA. Reduced adherence to traumatized rat heart valves by a low-fibronectin-binding mutant of Staphylococcus aureus. Infect Immum 1989; 57: 2306–12.

5. Vercellotti GMN, Lussenhop D, Peterson PK et al. Bacterial adherence to fibronectin and endothelialcells: a possible mechanism for bacterial tissue tropism. J Lab Clin Med 1984; 103: 34–43.

6. Frost AJ. Selective adhesion of microorganisms to the ductular epithelium of bovine mammary gland.Infect Immun 1975; 12: 1154–6.

7. Mamo W, Froman G, Wadstrom T. Interaction of subepithelial connective tissue components withStaphylococcus aureus and coagulase-negative staphylococci from bovine mastitis. Vet Microbiol1988; 18: 163–76.

8. Froman G, Switalski L, Speziale P, Hook M. Isolation and characterization of a fibronectin receptorfrom Staphylococcus aureus. J Biol Chem 1987; 262: 6564–71.

9. Proctor RA, Prendergast E, Mosher DF. Fibronectin mediates attachment of Staphylococcus aureusto human neutrophils. Blood 1982; 59: 681–7.

10. Ryden C. Studies on interaction between staphylococcal cells and some connective tissuecomponents. PhD Thesis, Uppsala University, Almquist and Wiksell 1987; Stockholm, Sweden.

11. Jonsson K, Signas C, Muller, H-P, Lindberg M. Two different genes encode fibronectin binding proteinsin Staphylococcus aureus. The complete nucleotide sequence and characterization of the secondgene. Eur J Biochem 1991; 202: 1041–8.

12. Mamo W, Jonsson P, Flock JI et al. Vaccination against Staphylococcus aureus mastitis: Immu-nological response of mice vaccinated with fibronectin-binding proteins (FnBps) to challenge with S.aureus. Vaccine 1994; 12: 988–92.

13. Luk JM, Flock JI, Wadstrom T. Detection of rabbit sera of blocking antibodies against staphylococcalfibronectin-binding proteins by enzyme-linked immunosorbent assay. FEMS Microbiol Immunol 1989;47: 505–10.

14. Raja RH, Raucci G, Hook M. Peptide analogue to a fibronectin receptor inhibit attachment of Staphy-lococcus aureus to fibronectin-containing substrates. Infect Immun 1990; 58: 2595–8.

15. Mamo W, Boden M, Flock JI. Vaccination with Staphylococcus aureus fibrinogen-binding proteins(FbBPs) reduces colonization of S. aureus in a mouse mastitis model. FEMS Immunol Med Microbiol1994; 10: 47–54.

16. Olmsted SB, Norcross NL. Effect of specific antibody on adherence of Staphylococcus aureus tobovine mammary epithelial cells. Infect Immun 1992; 60: 249–56.

17. Nemeth J, Lee JC. Antibodies to capsular polysaccharides are not protective against experimentalStaphylococcus aureus endocarditis. Infect Immun 1995; 63: 375–80.

18. Greenberg DP, Bayer AS, Cheung AL, Ward JI. Protective efficacy of protein A-specific antibodyagainst bacteremic infection due to Staphylococcus aureus in an infant rat model. Infect Immun 1989;57: 1113–8.

19. Rozalska B, Wadstrom T. Protective opsonic activity of antibodies against fibronectin binding proteins(FnBPs) of Staphylococcus aureus. Scand J Immunol 1993; 37: 575–80.

20. Jonsson P, Lindberg M, Haraldsson I, Wadstrom T. Virulence of Staphylococcus aureus in a mousemastitis model: Studies of alpha haemolysin, coagulase and protein A as possible virulence deter-minants with protoplast fusion and gene cloning. Infect Immun 1985; 49: 765–9.

21. Ciborowski P, Flock JI, Wadstrom T. Immunological response of a Staphylococcus aureus fibronectinbinding protein. J Med Microbiol 1992; 37: 376–81.

22. Signas C, Raucci G, Jonsson K et al. Nucleotide sequence of the gene for fibronectin binding proteinfrom Staphylococcus aureus and its use in the synthesis of biologically active peptides. Proc NatlAcad Sci USA, 1989; 86: 699–703.

23. Rantamaki L, Muller H-P. Isolation and characterization of a2-macroglobulin from mastitis milk. JDiary Research 1992; 59: 273–85.

24. Chandler RL. Experimental bacterial mastitis in the mouse. J Med Microbiol 1970; 3: 273–82.25. Haraldsson I, Jonsson P. Histopathology and pathogenesis of mouse mastitis induced with Staphy-

lococcus aureus mutants. J Comp Path 1984; 94: 183–96.