- Email: [email protected]
Identification of a 36-kDa fibronectin-binding protein expressed by a virulent variant of Leptospira interrogans serovar icterohaemorrhagiae
FEMS Microbiology Letters 185 (2000) 17^22
www.fems-microbiology.org
Identi¢cation of a 36-kDa ¢bronectin-binding protein expressed by a virulent variant of Leptospira interrogans serovar icterohaemorrhagiae Fabrice Merien a , Johann Truccolo a , Guy Baranton b , Philippe Perolat a
a;
*
Leptospira Laboratory, Institut Pasteur de Nouvelle-Cale¨donie, P.O. Box 61, 98845 Noume¨a Cedex, New Caledonia b Unite¨ de Bacte¨riologie Mole¨culaire et Me¨dicale, Institut Pasteur, 75724 Paris Cedex 15, France Received 21 January 2000; accepted 24 January 2000
Abstract We investigated the ability of a virulent strain of Leptospira interrogans serovar icterohaemorrhagiae, its isogenic avirulent variant and a saprophytic strain to bind fibronectin using alkaline phosphatase-labelled fibronectin. A single 36-kDa fibronectin-binding protein was expressed only by the virulent strain and was located in the outer sheath according to proteinase K treatment results. The interaction of this protein with fibronectin was specific and the region of fibronectin bound to this potential adhesin overlapped the gelatin-binding domain. The inability of a RGDS synthetic peptide to inhibit the binding of fibronectin indicated that the cell-binding domain was not involved in this interaction. Considering the wide distribution of fibronectin within a host and the diversity of mammals involved in the epidemiology of leptospirosis, its implication in the cell attachment process of virulent leptospires is coherent with the multiplicity of target cells. ß 2000 Published by Elsevier Science B.V. All rights reserved. Keywords : Leptospira; Virulence; Fibronectin
1. Introduction Leptospirosis is a zooanthroponosis widespread throughout the world caused by bacteria belonging to the genus Leptospira which includes species pathogenic for animals and humans, and saprophytic ones found in surface waters and soils [1]. The course of human leptospirosis varies from mild to severe fatal forms, the most severe forms of human leptospirosis being principally caused by Leptospira interrogans serovar icterohaemorrhagiae [1]. A bacteraemic stage is necessary for the virulent leptospires to reach and colonise the target tissues of the host organism [1]. Virulent leptospires can survive by evading phagocytosis and are isolated from the blood during the ¢rst week of the disease. Conversely, avirulent leptospires are rapidly cleared [2]. In addition, the ability of a virulent strain of Leptospira to induce programmed cell death in macrophage-like cells in an in vitro system [3] and in hepatocytes in an animal model [4] has been shown
* Corresponding author. Tel.: +687 (27) 26 66; Fax: +687 (27) 33 90; E-mail : [email protected]
and could contribute to the survival of leptospires in the host. Several attachment and invasion in vitro assays have been developed to study the interaction of leptospires with eukaryotic cells including ¢broblastic cells [5], renal epithelial cells [6] or human endothelial cells [7]. However, these studies lacked comparisons between the respective actions of virulent and avirulent variants of a single clone. In a previous study [3], we demonstrated that the early active invasion of Vero cells and the induction of apoptosis in a murine monocyte-macrophage-like cell line (J774A.1) were correlated with the virulence of a strain of pathogenic L. interrogans serovar icterohaemorrhagiae. The ¢rst step in adhesion of bacteria to the target cell membrane involves molecular recognition between surface components of both cell types [8]. In acute leptospirosis in animal models, leptospires have been demonstrated in the interstitium, penetrating between hepatocytes and tubular epithelial cells, suggesting an attachment to extracellular matrix components [4,9], but few studies have analysed the surface components of Leptospira involved in cell attachment. Cell surface-associated extracellular matrix proteins were implicated in the attachment of pathogenic lepto-
0378-1097 / 00 / $20.00 ß 2000 Published by Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 0 9 7 ( 0 0 ) 0 0 0 4 4 - 6
FEMSLE 9284 15-3-00
18
F. Merien et al. / FEMS Microbiology Letters 185 (2000) 17^22
spires to ¢broblasts but the nature of the ligand^receptor interaction was not investigated [10]. As leptospires may act intracellularly and have a potential invasive capacity for eukaryotic cells, the identi¢cation of molecules involved in host cell adherence is critical to understand the ¢rst steps in pathogenesis of leptospirosis. As several studies have documented the binding of ¢bronectin to other spirochetes [11,12], the aim of this study was to investigate a set of Leptospira strains (virulent, avirulent, saprophytic) for the expression of ¢bronectinbinding proteins. 2. Materials and methods 2.1. Bacterial strains and immunological reagents The virulent isolate Verdun [13], serovar icterohaemorrhagiae (Reference Collection of the Institute Pasteur, Paris, France), was ¢rst isolated from a human clinical case in 1917. Subsequently, two isogenic clones have been derived. The virulence of one was maintained by liquid nitrogen cryopreservation and iterative passages in guinea pigs [1]. The second clone was regularly passaged in vitro to obtain a high-passage avirulent variant. Both strains were grown in liquid EMJH medium [1] at 30³C under aerobic conditions until the stationary phase was reached (about 108 bacteria ml31 ). The saprophytic strain Patoc I (L. bi£exa serovar patoc) was grown under the same conditions. The leptospiral rabbit antisera had agglutination titres of at least 1:12 800 and sera were aliquoted and stored at 320³C until use. 2.2. Cell lines and culture conditions Vero cells (African green monkey kidney ¢broblast) were maintained in modi¢ed Eagle's medium (Sigma Chemical Co., St. Louis, MO, USA) bu¡ered with sodium bicarbonate and supplemented with 10% foetal calf serum and 100 Wg ml31 of gentamicin (Sigma Chemical Co.) per ml. Cells were cultured at 37³C in an humidi¢ed atmosphere containing 5% CO2 . 2.3. Preparation of outer sheath Isolation of the outer sheath of Leptospira was performed using 1 M NaCl and 0.02% sodium dodecyl sulfate (SDS) [14]. The protein concentration was measured using the Bio-Rad protein assay according to the manufacturer's instructions (Bio-Rad Laboratories, Hercules, CA, USA). 2.4. Macromolecules Fibronectin, ¢bronectin proteolytic fragments (30 kDa and 45 kDa), and collagen type IV were of human origin. Laminin was prepared from basement membrane of
mouse sarcoma. Aggrecan, decorin and gelatin were derived from bovine cartilage. Collagen types I and II were obtained from rat tail and chicken cartilage respectively. Heparin was derived from porcine intestinal mucosa, and chondroitin sulfate A was puri¢ed from bovine trachea. All products were from Sigma Chemical Co., including the RGDS synthetic peptide. Labeling of ¢bronectin with activated alkaline phosphatase (AP) was performed with an AP labeling kit according to the manufacturer's recommendations (Boehringer, Mannheim, Germany). 2.5. SDS-PAGE and detection of the ¢bronectin-binding activity Samples were subjected to SDS-PAGE (12.5%) on 8U9 cm gels (5 Wg protein per lane), according to the method of Laemmli [15]. After electrophoresis, proteins were transferred onto nitrocellulose membranes and ¢bronectinbinding activity was subsequently detected using the APconjugated ¢bronectin. Membranes were blocked for 1 h with phosphate-bu¡ered saline (PBS)^3% bovine serum albumin (BSA) and then probed with a 1:5000 dilution of AP-labelled ¢bronectin in PBS bu¡er. After overnight incubation at room temperature, membranes were washed three times with TTBS bu¡er (Tris 10 mM, 0.05% Tween 20) and bound ¢bronectin was detected by the addition of bromochloroindolyl phosphate and nitroblue tetrazolium (Boehringer). 2.6. Bacterial surface treatment with proteinase K An actively growing culture of leptospires in EMJH medium (20 ml, at least 108 leptospires ml31 ) was divided into two 10-ml aliquots and centrifuged for 10 min at 10 000Ug. The pellets were resuspended in 5 ml of PBSMg (PBS supplemented with 5 mM MgCl2 ) and centrifuged again for 10 min at 10 000Ug. The pellets were dislodged by the addition of 0.9 ml of PBS-Mg. Then, 100 Wl of proteinase K (Sigma) solution (4 mg ml31 ) was added to one tube, whereas the other tube received 100 Wl of PBS-Mg (untreated leptospires). The degradation of protein-containing moieties was performed for 30 min to 2 h at 37³C before adding phenylmethylsulfonyl £uoride (PMSF) to a ¢nal concentration of 50 Wg ml31 . The leptospires were washed twice in PBS-Mg and solubilised in sample bu¡er (62.5 mM Tris, 2% SDS, 10% glycerol, 5% 2-mercaptoethanol, pH 6.8). The protein content was estimated and lysates prepared for SDS-PAGE. 2.7. Mapping of the functional domain that interacts with leptospires In a ¢rst qualitative step, electrophoresis and transfer of proteins were performed as described above but the APlabelled ¢bronectin was ¢rst incubated for 1 h at room temperature with 5 Wg ml31 bu¡er, gelatin, heparin or
FEMSLE 9284 15-3-00
F. Merien et al. / FEMS Microbiology Letters 185 (2000) 17^22
chondroitin sulfate before probing the nitrocellulose membranes. In a second quantitative experiment, Immulon I microtitre plates (Dynatech Inc., Sully¢eld, VA, USA) were incubated overnight at 4³C with diluted outer sheath proteins (25 Wg ml31 in 0.1 M carbonate bu¡er pH 9.6). Plates were washed twice in TTBS and nonspeci¢c proteinbinding sites were blocked with PBS^BSA 3% for 2 h at room temperature. The AP-labelled ¢bronectin was incubated for 1 h at room temperature with the three previous ligands before being added to the plates (1:300 dilution) and incubated for 2 h at 37³C. Each molecule was tested in triplicate ; p-nitrophenyl phosphate (Sigma) was added to each well (200 Wl of a 1 mg ml31 solution) and the plates were incubated for 30 min at 37³C before measuring the absorbance at 405 nm. 2.8. Adherence assays Cells were harvested with 0.05% trypsin and 0.02% EDTA in a balanced salt solution (Boehringer), resuspended in prewarmed medium, counted and seeded in eight-chamber Lab-Tek tissue culture slides (Nunc Inc.) at a density of 2.5U104 per chamber. Cells were allowed to adhere for 24 h in a humidi¢ed atmosphere containing 5% CO2 at 37³C prior to use. Then, monolayers were washed twice with prewarmed medium without antibiotics, and 500 Wl (5U106 leptospires) of the virulent strain Verdun was added to each chamber. After 90 min incubation at 37³C, chambers were thoroughly washed twice with Hanks' balanced salt solution to remove unbound leptospires. Slides were air-dried and ¢xed for 2 min in pure methanol at 320³C. The leptospiral rabbit antiserum diluted 1 to 100 in PBS containing 0.5% BSA was applied to the slides for 30 min at room temperature before being washed three times in PBS. Then, the slides were incubated in the same conditions with goat anti-rabbit immunoglobulin G £uorescein F(abP)2 fragment (Boehringer) diluted 1:10 in PBS^0.5% BSA. After excess antibody was washed o¡ with PBS, slides were mounted in glycerol^PBS and examined with a Leitz DMRBE epi£uorescence microscope. Randomly selected cells in each chamber were examined to count the number of leptospires per cell. The mean was calculated by averaging the number of leptospires present in 100 cells in three independent assays, and the standard deviation was determined.
19
ml31 concentration whereas RGDS peptide was used at a 100 Wg ml31 concentration). Microplates were incubated for 2 h at 37³C and washed twice in TTBS. A 1:300 dilution of the AP-labelled ¢bronectin was added and plates were treated as described above. 3. Results 3.1. Evidence of a single ¢bronectin receptor As shown in Fig. 1, a ¢bronectin receptor was expressed only by the virulent variant. A single protein of 36 kDa that bound AP-labelled ¢bronectin was revealed after SDS-PAGE and transfer to nitrocellulose. Conversely, no ¢bronectin-binding activity was detected among the avirulent variant and the saprophytic strain. 3.2. Bacterial outer envelope treatment with proteinase K Surface-exposed proteins of live virulent leptospires were degraded with proteinase K. As shown in Fig. 2, this treatment resulted in the loss of the 36-kDa ¢bronectin-binding protein, con¢rming the localisation of this receptor to the outer envelope of virulent leptospires. 3.3. Attachment of leptospires to Vero cells Attachment of virulent leptospires to renal epithelial
2.9. Fibronectin-binding assays The capacity of various macromolecules to inhibit the interaction of outer sheath proteins of leptospires with AP-labelled ¢bronectin was evaluated in a microtitre plate assay. Proteins were coated and blocked as described above and unlabelled macromolecules were diluted in PBS^BSA 0.5%^Tween 20 0.05% and added to the wells (¢bronectin, 30-kDa and 45-kDa fragments, collagen types I, II and IV, aggrecan and decorin were all used at a 5 Wg
Fig. 1. Outer sheath proteins of leptospires were resolved by SDSPAGE (5 Wg protein per lane), transferred to nitrocellulose and probed with AP-labelled ¢bronectin. A: Virulent variant of strain Verdun (L. interrogans serovar icterohaemorrhagiae); B: avirulent variant of strain Verdun; C: saprophytic strain Patoc I (L. bi£exa serovar patoc). Molecular mass standards are given on the right in kDa. The virulent strain expresses a single protein of 36 kDa that binds AP-labelled ¢bronectin (black arrow).
FEMSLE 9284 15-3-00
20
F. Merien et al. / FEMS Microbiology Letters 185 (2000) 17^22
3.4. Mapping the functional domain that interacts with Leptospira
Fig. 2. Proteins from proteinase K-treated and untreated virulent leptospires (L. interrogans serovar icterohaemorrhagiae strain Verdun) were separated by SDS-PAGE, transferred to nitrocellulose and probed with AP-labelled ¢bronectin. C (control) : untreated virulent leptospires; P30: proteinase K-treated virulent leptospires (30 min, 37³C); P120: proteinase K-treated virulent leptospires (120 min, 37³C). The 36-kDa ¢bronectin-binding protein in shown on the left (black arrow).
cells (Vero cell line) was examined following preincubation of bacteria with soluble ¢bronectin. After 90 min at 37³C and indirect immuno£uorescence labelling, leptospires attached to Vero cells were counted by £uorescence microscopy. As shown in Fig. 3, a signi¢cant decrease of adhesion (up to 50% inhibition) was observed when the concentration of ¢bronectin increased (from 10 to 100 Wg ml31 ), compared to the control without ¢bronectin.
To ascertain the region of ¢bronectin that interacts with virulent leptospires, known ligands of ¢bronectin (heparin, gelatin and chondroitin sulfate) were used to block speci¢c binding sites. In a qualitative experiment, AP-labelled ¢bronectin was incubated with either heparin, gelatin or chondroitin sulfate for 1 h at 5 Wg ml31 and added to the nitrocellulose strips. The e¡ect of preincubation with the previous ligands on the ¢bronectin-binding activity was detected by a decrease of the 36-kDa band intensity compared to the untreated control. Preincubation of the probe with chondroitin sulfate or heparin produced bands of similar intensity. However, reduced binding of AP-labelled ¢bronectin to outer sheath proteins of virulent leptospires was noted after incubation of the probe with gelatin. Indeed, the intensity of the 36-kDa band was dramatically reduced compared to the control without ligand (Fig. 4). To quantify the inhibition, we tested the ability of heparin, gelatin or chondroitin sulfate to block the interaction of AP-labelled ¢bronectin with coated outer sheath proteins of virulent leptospires in a microtitre plate assay. As expected, prior incubation of the probe with gelatin reduced ¢bronectin binding to immobilised leptospire proteins by 50%. Chondroitin sulfate and heparin had no e¡ect on the interaction of ¢bronectin (Table 1). These results indicated that the site of the 36-kDa protein interaction overlapped with the gelatin-binding domain of ¢bronectin. 3.5. Inhibition experiments Inhibition experiments were conducted to evaluate the
Fig. 3. Adhesion of virulent leptospires (L. interrogans serovar icterohaemorrhagiae strain Verdun) to Vero cells in the presence of soluble ¢bronectin. Attached leptospires were counted by £uorescence microscopy and compared with the untreated control (no preincubation of leptospires with ¢bronectin). Data are the means þ S.D. of three experiments.
FEMSLE 9284 15-3-00
F. Merien et al. / FEMS Microbiology Letters 185 (2000) 17^22
21
Table 2 Ability of various molecules to inhibit the interaction of AP-labelled ¢bronectin with outer sheath proteins of a virulent strain of Leptospira Molecule Fibronectin Fibronectin 30-kDa fragment Fibronectin 45-kDa fragment Collagen type I Collagen type II Collagen type IV Laminin Aggrecan Decorin RGDS peptide
Inhibition (%)a 59 9 52 12 17 14 3 0 0 15
a (Average absorbance of wells receiving untreated AP-labelled ¢bronectin minus average absorbance of wells receiving molecules/average absorbance of wells receiving untreated AP-labelled ¢bronectin)U100. Results are representative of three experiments.
Fig. 4. Mapping of the ¢bronectin functional domain that interacts with the ¢bronectin-binding protein of virulent leptospires (L. interrogans serovar icterohaemorrhagiae strain Verdun). After SDS-PAGE and transfer to nitrocellulose, proteins (5 Wg per lane) were probed with preincubated AP-labelled ¢bronectin with either bu¡er (B), heparin (H), chondroitin sulfate (CS) or gelatin (G). The 36-kDa ¢bronectin-binding protein in shown on the left (black arrow).
speci¢city of the ¢bronectin-binding reaction. Microtitre plates were coated with outer sheath proteins of virulent leptospires (20 Wg ml31 ) and the speci¢city of ¢bronectin binding was tested using various unlabelled molecules to inhibit the interaction of AP-labelled ¢bronectin. Inhibition of AP-labelled ¢bronectin-binding activity was maximised at 59% with a concentration of unlabelled ¢bronectin of 10 Wg ml31 . Greater concentrations did not enhance the inhibition. Table 2 shows that no signi¢cant inhibition of AP-labelled ¢bronectin binding was observed with the ¢bronectin 30-kDa proteolytic fragment (heparin-binding fragment), collagen types I, II and IV, laminin, aggrecan, decorin and the synthetic RGDS peptide. However, a signi¢cant inhibition of the interaction of ¢bronectin with outer sheath proteins of virulent leptospires was seen with the ¢bronectin 45-kDa proteolytic fragment (Table 2) which corresponds to the gelatin-binding fragment. Table 1 E¡ect of preincubation of AP-labelled ¢bronectin with ligands on the ¢bronectin-binding activity of outer sheath proteins of virulent leptospires (L. interrogans serovar icterohaemorrhagiae strain Verdun) Ligand (50 Wg ml31 ) Heparin Chondroitin sulfate Gelatin a
Inhibition (%)a 0 0 51
(Average absorbance of wells receiving untreated AP-labelled ¢bronectin minus average absorbance of wells receiving preincubated AP-labelled ¢bronectin with ligands/average absorbance of wells receiving untreated AP-labelled ¢bronectin)U100. Results are representative of three experiments.
These results indicate that the interaction of ¢bronectin with the ¢bronectin-binding protein of virulent leptospires is speci¢c. This interaction implicates the gelatin-binding domain of ¢bronectin. However, the integrin/cell-binding domain of ¢bronectin is not involved in the attachment process as previously reported [10]. 4. Discussion Although in vitro interactions of leptospires with a wide variety of cell types and structures have been documented, molecular mechanisms by which virulent leptospires spread and colonise target tissues of the human or animal hosts have been poorly investigated. In biochemical terms, this interaction involves complementarities between surface structures (adhesins) on the bacteria and membrane components on the host cell surface [8]. A number of molecules, including ¢bronectin, laminin, collagen and hyaluronic acid, have been used to study the interactions of pathogenic leptospires [10], but the bacterial surface structures potentially implicated have not been determined. In our study, we have identi¢ed a 36-kDa ¢bronectin-binding protein speci¢cally expressed by a virulent variant of pathogenic leptospires. Conversely, this protein was not found among the corresponding avirulent variant and a saprophytic strain (Patoc I). After treatment of live virulent leptospires with proteinase K, the ¢bronectin-binding activity was lost, con¢rming that this receptor was located on the outer surface of the bacteria. The interaction of this protein which should act as a receptor for ¢bronectin was found to be speci¢c. Indeed, unlabelled ¢bronectin inhibited, in a dose-dependent manner, the adherence of virulent leptospires with Vero cells. We also mapped the functional domain of ¢bronectin that interacts with this receptor; gelatin inhibits the interaction of our leptospiral 36-kDa protein with ¢bronectin, suggesting that the ¢bronectin domain that binds gelatin was involved. Interestingly, similar results have been reported for the P47 ¢bro-
FEMSLE 9284 15-3-00
22
F. Merien et al. / FEMS Microbiology Letters 185 (2000) 17^22
nectin-binding protein of Borrelia burgdorferi [12]. In our study, the inability of a RGDS synthetic peptide to inhibit the binding of ¢bronectin indicated that the integrin/cellbinding domain was not involved in this interaction. Other spirochetes, such as Treponema pallidum [11] and B. burgdorferi [12], express ¢bronectin-binding proteins but only those of Treponema interact with the cell-binding domain of ¢bronectin. Our data could suggest that the putative role of this protein is an adhesin. Indeed, ¢bronectin is a dimeric glycoprotein with a molecular mass V440 kDa which is present in a soluble form in the plasma and various body £uids and, in an insoluble form, on cell surfaces and in the intracellular matrix [16]. In our leptospiral model, the ¢bronectin-binding protein was only expressed by the virulent variant. Interestingly, in a previous study, the extent of adhesion to ¢bronectin-coated coverslips was correlated with the virulence of a Leptospira strain de¢ned using LD50 [10]; furthermore, `highly' virulent leptospires attached more e¤ciently than the `intermediately' virulent and avirulent variants of the same strain [10]. Many pathogenic bacteria express ¢bronectin-binding proteins which may be important virulence factors. As an example, ¢bronectin binding has been studied extensively in Staphylococcus aureus and is associated with invasiveness of the organism [17]. Fibronectin binding may lead to the colonisation of the mammalian host by attachment of pathogens to a variety of cells and tissues. Leptospires pass through intact skin and abrasions in the skin or mucosal surfaces of the host [1]. The wide distribution of ¢bronectin in the host may help adherence of virulent leptospires at the cutaneous or mucosal site of entry, and then in their dissemination within target organs after a bacteraemic stage. In experimental infection of Syrian hamsters [9] or guinea pigs [4] with a virulent strain of Leptospira, leptospires were observed in the liver in large number in the intercellular spaces where the extracellular matrix is rich in ¢bronectin. In conclusion, we have identi¢ed a ¢bronectin-binding protein exclusively expressed by a virulent variant of L. interrogans serovar icterohaemorrhagiae. Our results suggest that this 36-kDa protein localised on the cell surface may play a role in the attachment of virulent leptospires to host tissues. Indeed, the presence of a ¢bronectin-binding protein may be an essential prerequisite for the initial steps leading to infection. Virulent leptospires which lack expression of a ¢bronectin-binding protein would be unable to multiply within a host and be cleared rapidly. Considering the wide distribution of ¢bronectin within a host and the diversity of mammalian hosts implicated in the epidemiology of leptospirosis, ¢bronectin may contribute to the e¤cient attachment of virulent leptospires in di¡erent animal species involved in the natural history of this zoonosis. If molecular mechanisms of the adhesion of leptospires to eukaryotic membranes are poorly documented, their study would be critical for the future development of
vaccine candidates against leptospirosis. Future experiments should concern the underlying molecular mechanisms determining the phenotype of adhesion of virulent leptospires. Acknowledgements F.M. and P.P. were supported by the Institute Pasteur, Paris (International Network of Institutes Pasteur). The authors are indebted to Dominique Triballi for technical assistance. References [1] Faine, S., Adler, B., Bolin, C. and Perolat, P. (1999) Leptospira and Leptospirosis, 2nd edn. MediSci, Armadale. [2] Faine, S. (1964) Reticuloendothelial phagocytosis of virulent leptospires. Am. J. Vet. Res. 25, 830^835. [3] Merien, F., Baranton, G. and Perolat, P. (1997) Invasion of Vero cells and induction of apoptosis in macrophages by pathogenic Leptospira interrogans are correlated with virulence. Infect. Immun. 65, 729^738. [4] Merien, F., Truccolo, J., Rougier, Y., Baranton, G. and Perolat, P. (1998) In vivo apoptosis of hepatocytes in guinea pigs infected with Leptospira interrogans serovar icterohaemorrhagiae. FEMS Microbiol. Lett. 169, 95^102. [5] Vinh, T., Faine, S. and Adler, B. (1984) Adhesion of leptospires to mouse ¢broblasts (L929) and its enhancement by speci¢c antibody. J. Med. Microbiol. 18, 73^85. [6] Ballard, S.A., Williamson, M., Adler, B., Vinh, T. and Faine, S. (1986) Interactions of virulent and avirulent leptospires with primary cultures of renal epithelial cells. J. Med. Microbiol. 21, 59^67. [7] Thomas, D.D. and Higbie, L.M. (1990) In vitro association of leptospires with host cells. Infect. Immun. 58, 581^585. [8] Roth, J.A., Bolin, C.A., Brogden, K.A., Minion, F.C. and Wannemuehler, M.J. (1995) Virulence Mechanisms of Bacterial Pathogens, 2nd edn. ASM Press, Washington, DC. [9] Van den Ingh, T.S.G.A.M. and Hartman, E.G. (1986) Pathology of acute Leptospira interrogans serotype icterohaemorrhagiae infection in the Syrian hamster. Vet. Microbiol. 12, 367^376. [10] Ito, T. and Yanagawa, R.. (1987) Leptospiral attachment to four structural components of extracellular matrix. Jpn. J. Vet. Sci. 49, 875^882. [11] Thomas, D.D., Baseman, J.B. and Alderete, J.F. (1985) Fibronectin tetrapeptide is target for syphillis spirochete cytadherence. J. Exp. Med. 162, 1715^1719. [12] Probert, W.S. and Johnson, B.J.B. (1998) Identi¢cation of a 47 kDa ¢bronectin-binding protein expressed by Borrelia burgdorferi isolate B31. Mol. Microbiol. 30, 1003^1015. [13] Pettit, A. (1928) Contribution a© l'e¨tude des Spiroche¨tide¨s. Editions Me¨dicales, Paris. [14] Auran, N.E., Johnson, R.C. and Ritzi, D.M. (1972) Isolation of the outer sheath of Leptospira and its immunogenic properties in hamsters. Infect. Immun. 5, 968^975. [15] Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680^685. [16] Proctor, R.A. (1987) Fibronectin : a brief overview of its structure, function and physiology. Rev. Infect. Dis. 9, S317^S321. [17] Proctor, R.A. (1987) The staphylococcal ¢bronectin receptor: evidence for its importance of invasive infections. Rev. Infect. Dis. 9, S335^S340.
FEMSLE 9284 15-3-00
www.fems-microbiology.org
Identi¢cation of a 36-kDa ¢bronectin-binding protein expressed by a virulent variant of Leptospira interrogans serovar icterohaemorrhagiae Fabrice Merien a , Johann Truccolo a , Guy Baranton b , Philippe Perolat a
a;
*
Leptospira Laboratory, Institut Pasteur de Nouvelle-Cale¨donie, P.O. Box 61, 98845 Noume¨a Cedex, New Caledonia b Unite¨ de Bacte¨riologie Mole¨culaire et Me¨dicale, Institut Pasteur, 75724 Paris Cedex 15, France Received 21 January 2000; accepted 24 January 2000
Abstract We investigated the ability of a virulent strain of Leptospira interrogans serovar icterohaemorrhagiae, its isogenic avirulent variant and a saprophytic strain to bind fibronectin using alkaline phosphatase-labelled fibronectin. A single 36-kDa fibronectin-binding protein was expressed only by the virulent strain and was located in the outer sheath according to proteinase K treatment results. The interaction of this protein with fibronectin was specific and the region of fibronectin bound to this potential adhesin overlapped the gelatin-binding domain. The inability of a RGDS synthetic peptide to inhibit the binding of fibronectin indicated that the cell-binding domain was not involved in this interaction. Considering the wide distribution of fibronectin within a host and the diversity of mammals involved in the epidemiology of leptospirosis, its implication in the cell attachment process of virulent leptospires is coherent with the multiplicity of target cells. ß 2000 Published by Elsevier Science B.V. All rights reserved. Keywords : Leptospira; Virulence; Fibronectin
1. Introduction Leptospirosis is a zooanthroponosis widespread throughout the world caused by bacteria belonging to the genus Leptospira which includes species pathogenic for animals and humans, and saprophytic ones found in surface waters and soils [1]. The course of human leptospirosis varies from mild to severe fatal forms, the most severe forms of human leptospirosis being principally caused by Leptospira interrogans serovar icterohaemorrhagiae [1]. A bacteraemic stage is necessary for the virulent leptospires to reach and colonise the target tissues of the host organism [1]. Virulent leptospires can survive by evading phagocytosis and are isolated from the blood during the ¢rst week of the disease. Conversely, avirulent leptospires are rapidly cleared [2]. In addition, the ability of a virulent strain of Leptospira to induce programmed cell death in macrophage-like cells in an in vitro system [3] and in hepatocytes in an animal model [4] has been shown
* Corresponding author. Tel.: +687 (27) 26 66; Fax: +687 (27) 33 90; E-mail : [email protected]
and could contribute to the survival of leptospires in the host. Several attachment and invasion in vitro assays have been developed to study the interaction of leptospires with eukaryotic cells including ¢broblastic cells [5], renal epithelial cells [6] or human endothelial cells [7]. However, these studies lacked comparisons between the respective actions of virulent and avirulent variants of a single clone. In a previous study [3], we demonstrated that the early active invasion of Vero cells and the induction of apoptosis in a murine monocyte-macrophage-like cell line (J774A.1) were correlated with the virulence of a strain of pathogenic L. interrogans serovar icterohaemorrhagiae. The ¢rst step in adhesion of bacteria to the target cell membrane involves molecular recognition between surface components of both cell types [8]. In acute leptospirosis in animal models, leptospires have been demonstrated in the interstitium, penetrating between hepatocytes and tubular epithelial cells, suggesting an attachment to extracellular matrix components [4,9], but few studies have analysed the surface components of Leptospira involved in cell attachment. Cell surface-associated extracellular matrix proteins were implicated in the attachment of pathogenic lepto-
0378-1097 / 00 / $20.00 ß 2000 Published by Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 0 9 7 ( 0 0 ) 0 0 0 4 4 - 6
FEMSLE 9284 15-3-00
18
F. Merien et al. / FEMS Microbiology Letters 185 (2000) 17^22
spires to ¢broblasts but the nature of the ligand^receptor interaction was not investigated [10]. As leptospires may act intracellularly and have a potential invasive capacity for eukaryotic cells, the identi¢cation of molecules involved in host cell adherence is critical to understand the ¢rst steps in pathogenesis of leptospirosis. As several studies have documented the binding of ¢bronectin to other spirochetes [11,12], the aim of this study was to investigate a set of Leptospira strains (virulent, avirulent, saprophytic) for the expression of ¢bronectinbinding proteins. 2. Materials and methods 2.1. Bacterial strains and immunological reagents The virulent isolate Verdun [13], serovar icterohaemorrhagiae (Reference Collection of the Institute Pasteur, Paris, France), was ¢rst isolated from a human clinical case in 1917. Subsequently, two isogenic clones have been derived. The virulence of one was maintained by liquid nitrogen cryopreservation and iterative passages in guinea pigs [1]. The second clone was regularly passaged in vitro to obtain a high-passage avirulent variant. Both strains were grown in liquid EMJH medium [1] at 30³C under aerobic conditions until the stationary phase was reached (about 108 bacteria ml31 ). The saprophytic strain Patoc I (L. bi£exa serovar patoc) was grown under the same conditions. The leptospiral rabbit antisera had agglutination titres of at least 1:12 800 and sera were aliquoted and stored at 320³C until use. 2.2. Cell lines and culture conditions Vero cells (African green monkey kidney ¢broblast) were maintained in modi¢ed Eagle's medium (Sigma Chemical Co., St. Louis, MO, USA) bu¡ered with sodium bicarbonate and supplemented with 10% foetal calf serum and 100 Wg ml31 of gentamicin (Sigma Chemical Co.) per ml. Cells were cultured at 37³C in an humidi¢ed atmosphere containing 5% CO2 . 2.3. Preparation of outer sheath Isolation of the outer sheath of Leptospira was performed using 1 M NaCl and 0.02% sodium dodecyl sulfate (SDS) [14]. The protein concentration was measured using the Bio-Rad protein assay according to the manufacturer's instructions (Bio-Rad Laboratories, Hercules, CA, USA). 2.4. Macromolecules Fibronectin, ¢bronectin proteolytic fragments (30 kDa and 45 kDa), and collagen type IV were of human origin. Laminin was prepared from basement membrane of
mouse sarcoma. Aggrecan, decorin and gelatin were derived from bovine cartilage. Collagen types I and II were obtained from rat tail and chicken cartilage respectively. Heparin was derived from porcine intestinal mucosa, and chondroitin sulfate A was puri¢ed from bovine trachea. All products were from Sigma Chemical Co., including the RGDS synthetic peptide. Labeling of ¢bronectin with activated alkaline phosphatase (AP) was performed with an AP labeling kit according to the manufacturer's recommendations (Boehringer, Mannheim, Germany). 2.5. SDS-PAGE and detection of the ¢bronectin-binding activity Samples were subjected to SDS-PAGE (12.5%) on 8U9 cm gels (5 Wg protein per lane), according to the method of Laemmli [15]. After electrophoresis, proteins were transferred onto nitrocellulose membranes and ¢bronectinbinding activity was subsequently detected using the APconjugated ¢bronectin. Membranes were blocked for 1 h with phosphate-bu¡ered saline (PBS)^3% bovine serum albumin (BSA) and then probed with a 1:5000 dilution of AP-labelled ¢bronectin in PBS bu¡er. After overnight incubation at room temperature, membranes were washed three times with TTBS bu¡er (Tris 10 mM, 0.05% Tween 20) and bound ¢bronectin was detected by the addition of bromochloroindolyl phosphate and nitroblue tetrazolium (Boehringer). 2.6. Bacterial surface treatment with proteinase K An actively growing culture of leptospires in EMJH medium (20 ml, at least 108 leptospires ml31 ) was divided into two 10-ml aliquots and centrifuged for 10 min at 10 000Ug. The pellets were resuspended in 5 ml of PBSMg (PBS supplemented with 5 mM MgCl2 ) and centrifuged again for 10 min at 10 000Ug. The pellets were dislodged by the addition of 0.9 ml of PBS-Mg. Then, 100 Wl of proteinase K (Sigma) solution (4 mg ml31 ) was added to one tube, whereas the other tube received 100 Wl of PBS-Mg (untreated leptospires). The degradation of protein-containing moieties was performed for 30 min to 2 h at 37³C before adding phenylmethylsulfonyl £uoride (PMSF) to a ¢nal concentration of 50 Wg ml31 . The leptospires were washed twice in PBS-Mg and solubilised in sample bu¡er (62.5 mM Tris, 2% SDS, 10% glycerol, 5% 2-mercaptoethanol, pH 6.8). The protein content was estimated and lysates prepared for SDS-PAGE. 2.7. Mapping of the functional domain that interacts with leptospires In a ¢rst qualitative step, electrophoresis and transfer of proteins were performed as described above but the APlabelled ¢bronectin was ¢rst incubated for 1 h at room temperature with 5 Wg ml31 bu¡er, gelatin, heparin or
FEMSLE 9284 15-3-00
F. Merien et al. / FEMS Microbiology Letters 185 (2000) 17^22
chondroitin sulfate before probing the nitrocellulose membranes. In a second quantitative experiment, Immulon I microtitre plates (Dynatech Inc., Sully¢eld, VA, USA) were incubated overnight at 4³C with diluted outer sheath proteins (25 Wg ml31 in 0.1 M carbonate bu¡er pH 9.6). Plates were washed twice in TTBS and nonspeci¢c proteinbinding sites were blocked with PBS^BSA 3% for 2 h at room temperature. The AP-labelled ¢bronectin was incubated for 1 h at room temperature with the three previous ligands before being added to the plates (1:300 dilution) and incubated for 2 h at 37³C. Each molecule was tested in triplicate ; p-nitrophenyl phosphate (Sigma) was added to each well (200 Wl of a 1 mg ml31 solution) and the plates were incubated for 30 min at 37³C before measuring the absorbance at 405 nm. 2.8. Adherence assays Cells were harvested with 0.05% trypsin and 0.02% EDTA in a balanced salt solution (Boehringer), resuspended in prewarmed medium, counted and seeded in eight-chamber Lab-Tek tissue culture slides (Nunc Inc.) at a density of 2.5U104 per chamber. Cells were allowed to adhere for 24 h in a humidi¢ed atmosphere containing 5% CO2 at 37³C prior to use. Then, monolayers were washed twice with prewarmed medium without antibiotics, and 500 Wl (5U106 leptospires) of the virulent strain Verdun was added to each chamber. After 90 min incubation at 37³C, chambers were thoroughly washed twice with Hanks' balanced salt solution to remove unbound leptospires. Slides were air-dried and ¢xed for 2 min in pure methanol at 320³C. The leptospiral rabbit antiserum diluted 1 to 100 in PBS containing 0.5% BSA was applied to the slides for 30 min at room temperature before being washed three times in PBS. Then, the slides were incubated in the same conditions with goat anti-rabbit immunoglobulin G £uorescein F(abP)2 fragment (Boehringer) diluted 1:10 in PBS^0.5% BSA. After excess antibody was washed o¡ with PBS, slides were mounted in glycerol^PBS and examined with a Leitz DMRBE epi£uorescence microscope. Randomly selected cells in each chamber were examined to count the number of leptospires per cell. The mean was calculated by averaging the number of leptospires present in 100 cells in three independent assays, and the standard deviation was determined.
19
ml31 concentration whereas RGDS peptide was used at a 100 Wg ml31 concentration). Microplates were incubated for 2 h at 37³C and washed twice in TTBS. A 1:300 dilution of the AP-labelled ¢bronectin was added and plates were treated as described above. 3. Results 3.1. Evidence of a single ¢bronectin receptor As shown in Fig. 1, a ¢bronectin receptor was expressed only by the virulent variant. A single protein of 36 kDa that bound AP-labelled ¢bronectin was revealed after SDS-PAGE and transfer to nitrocellulose. Conversely, no ¢bronectin-binding activity was detected among the avirulent variant and the saprophytic strain. 3.2. Bacterial outer envelope treatment with proteinase K Surface-exposed proteins of live virulent leptospires were degraded with proteinase K. As shown in Fig. 2, this treatment resulted in the loss of the 36-kDa ¢bronectin-binding protein, con¢rming the localisation of this receptor to the outer envelope of virulent leptospires. 3.3. Attachment of leptospires to Vero cells Attachment of virulent leptospires to renal epithelial
2.9. Fibronectin-binding assays The capacity of various macromolecules to inhibit the interaction of outer sheath proteins of leptospires with AP-labelled ¢bronectin was evaluated in a microtitre plate assay. Proteins were coated and blocked as described above and unlabelled macromolecules were diluted in PBS^BSA 0.5%^Tween 20 0.05% and added to the wells (¢bronectin, 30-kDa and 45-kDa fragments, collagen types I, II and IV, aggrecan and decorin were all used at a 5 Wg
Fig. 1. Outer sheath proteins of leptospires were resolved by SDSPAGE (5 Wg protein per lane), transferred to nitrocellulose and probed with AP-labelled ¢bronectin. A: Virulent variant of strain Verdun (L. interrogans serovar icterohaemorrhagiae); B: avirulent variant of strain Verdun; C: saprophytic strain Patoc I (L. bi£exa serovar patoc). Molecular mass standards are given on the right in kDa. The virulent strain expresses a single protein of 36 kDa that binds AP-labelled ¢bronectin (black arrow).
FEMSLE 9284 15-3-00
20
F. Merien et al. / FEMS Microbiology Letters 185 (2000) 17^22
3.4. Mapping the functional domain that interacts with Leptospira
Fig. 2. Proteins from proteinase K-treated and untreated virulent leptospires (L. interrogans serovar icterohaemorrhagiae strain Verdun) were separated by SDS-PAGE, transferred to nitrocellulose and probed with AP-labelled ¢bronectin. C (control) : untreated virulent leptospires; P30: proteinase K-treated virulent leptospires (30 min, 37³C); P120: proteinase K-treated virulent leptospires (120 min, 37³C). The 36-kDa ¢bronectin-binding protein in shown on the left (black arrow).
cells (Vero cell line) was examined following preincubation of bacteria with soluble ¢bronectin. After 90 min at 37³C and indirect immuno£uorescence labelling, leptospires attached to Vero cells were counted by £uorescence microscopy. As shown in Fig. 3, a signi¢cant decrease of adhesion (up to 50% inhibition) was observed when the concentration of ¢bronectin increased (from 10 to 100 Wg ml31 ), compared to the control without ¢bronectin.
To ascertain the region of ¢bronectin that interacts with virulent leptospires, known ligands of ¢bronectin (heparin, gelatin and chondroitin sulfate) were used to block speci¢c binding sites. In a qualitative experiment, AP-labelled ¢bronectin was incubated with either heparin, gelatin or chondroitin sulfate for 1 h at 5 Wg ml31 and added to the nitrocellulose strips. The e¡ect of preincubation with the previous ligands on the ¢bronectin-binding activity was detected by a decrease of the 36-kDa band intensity compared to the untreated control. Preincubation of the probe with chondroitin sulfate or heparin produced bands of similar intensity. However, reduced binding of AP-labelled ¢bronectin to outer sheath proteins of virulent leptospires was noted after incubation of the probe with gelatin. Indeed, the intensity of the 36-kDa band was dramatically reduced compared to the control without ligand (Fig. 4). To quantify the inhibition, we tested the ability of heparin, gelatin or chondroitin sulfate to block the interaction of AP-labelled ¢bronectin with coated outer sheath proteins of virulent leptospires in a microtitre plate assay. As expected, prior incubation of the probe with gelatin reduced ¢bronectin binding to immobilised leptospire proteins by 50%. Chondroitin sulfate and heparin had no e¡ect on the interaction of ¢bronectin (Table 1). These results indicated that the site of the 36-kDa protein interaction overlapped with the gelatin-binding domain of ¢bronectin. 3.5. Inhibition experiments Inhibition experiments were conducted to evaluate the
Fig. 3. Adhesion of virulent leptospires (L. interrogans serovar icterohaemorrhagiae strain Verdun) to Vero cells in the presence of soluble ¢bronectin. Attached leptospires were counted by £uorescence microscopy and compared with the untreated control (no preincubation of leptospires with ¢bronectin). Data are the means þ S.D. of three experiments.
FEMSLE 9284 15-3-00
F. Merien et al. / FEMS Microbiology Letters 185 (2000) 17^22
21
Table 2 Ability of various molecules to inhibit the interaction of AP-labelled ¢bronectin with outer sheath proteins of a virulent strain of Leptospira Molecule Fibronectin Fibronectin 30-kDa fragment Fibronectin 45-kDa fragment Collagen type I Collagen type II Collagen type IV Laminin Aggrecan Decorin RGDS peptide
Inhibition (%)a 59 9 52 12 17 14 3 0 0 15
a (Average absorbance of wells receiving untreated AP-labelled ¢bronectin minus average absorbance of wells receiving molecules/average absorbance of wells receiving untreated AP-labelled ¢bronectin)U100. Results are representative of three experiments.
Fig. 4. Mapping of the ¢bronectin functional domain that interacts with the ¢bronectin-binding protein of virulent leptospires (L. interrogans serovar icterohaemorrhagiae strain Verdun). After SDS-PAGE and transfer to nitrocellulose, proteins (5 Wg per lane) were probed with preincubated AP-labelled ¢bronectin with either bu¡er (B), heparin (H), chondroitin sulfate (CS) or gelatin (G). The 36-kDa ¢bronectin-binding protein in shown on the left (black arrow).
speci¢city of the ¢bronectin-binding reaction. Microtitre plates were coated with outer sheath proteins of virulent leptospires (20 Wg ml31 ) and the speci¢city of ¢bronectin binding was tested using various unlabelled molecules to inhibit the interaction of AP-labelled ¢bronectin. Inhibition of AP-labelled ¢bronectin-binding activity was maximised at 59% with a concentration of unlabelled ¢bronectin of 10 Wg ml31 . Greater concentrations did not enhance the inhibition. Table 2 shows that no signi¢cant inhibition of AP-labelled ¢bronectin binding was observed with the ¢bronectin 30-kDa proteolytic fragment (heparin-binding fragment), collagen types I, II and IV, laminin, aggrecan, decorin and the synthetic RGDS peptide. However, a signi¢cant inhibition of the interaction of ¢bronectin with outer sheath proteins of virulent leptospires was seen with the ¢bronectin 45-kDa proteolytic fragment (Table 2) which corresponds to the gelatin-binding fragment. Table 1 E¡ect of preincubation of AP-labelled ¢bronectin with ligands on the ¢bronectin-binding activity of outer sheath proteins of virulent leptospires (L. interrogans serovar icterohaemorrhagiae strain Verdun) Ligand (50 Wg ml31 ) Heparin Chondroitin sulfate Gelatin a
Inhibition (%)a 0 0 51
(Average absorbance of wells receiving untreated AP-labelled ¢bronectin minus average absorbance of wells receiving preincubated AP-labelled ¢bronectin with ligands/average absorbance of wells receiving untreated AP-labelled ¢bronectin)U100. Results are representative of three experiments.
These results indicate that the interaction of ¢bronectin with the ¢bronectin-binding protein of virulent leptospires is speci¢c. This interaction implicates the gelatin-binding domain of ¢bronectin. However, the integrin/cell-binding domain of ¢bronectin is not involved in the attachment process as previously reported [10]. 4. Discussion Although in vitro interactions of leptospires with a wide variety of cell types and structures have been documented, molecular mechanisms by which virulent leptospires spread and colonise target tissues of the human or animal hosts have been poorly investigated. In biochemical terms, this interaction involves complementarities between surface structures (adhesins) on the bacteria and membrane components on the host cell surface [8]. A number of molecules, including ¢bronectin, laminin, collagen and hyaluronic acid, have been used to study the interactions of pathogenic leptospires [10], but the bacterial surface structures potentially implicated have not been determined. In our study, we have identi¢ed a 36-kDa ¢bronectin-binding protein speci¢cally expressed by a virulent variant of pathogenic leptospires. Conversely, this protein was not found among the corresponding avirulent variant and a saprophytic strain (Patoc I). After treatment of live virulent leptospires with proteinase K, the ¢bronectin-binding activity was lost, con¢rming that this receptor was located on the outer surface of the bacteria. The interaction of this protein which should act as a receptor for ¢bronectin was found to be speci¢c. Indeed, unlabelled ¢bronectin inhibited, in a dose-dependent manner, the adherence of virulent leptospires with Vero cells. We also mapped the functional domain of ¢bronectin that interacts with this receptor; gelatin inhibits the interaction of our leptospiral 36-kDa protein with ¢bronectin, suggesting that the ¢bronectin domain that binds gelatin was involved. Interestingly, similar results have been reported for the P47 ¢bro-
FEMSLE 9284 15-3-00
22
F. Merien et al. / FEMS Microbiology Letters 185 (2000) 17^22
nectin-binding protein of Borrelia burgdorferi [12]. In our study, the inability of a RGDS synthetic peptide to inhibit the binding of ¢bronectin indicated that the integrin/cellbinding domain was not involved in this interaction. Other spirochetes, such as Treponema pallidum [11] and B. burgdorferi [12], express ¢bronectin-binding proteins but only those of Treponema interact with the cell-binding domain of ¢bronectin. Our data could suggest that the putative role of this protein is an adhesin. Indeed, ¢bronectin is a dimeric glycoprotein with a molecular mass V440 kDa which is present in a soluble form in the plasma and various body £uids and, in an insoluble form, on cell surfaces and in the intracellular matrix [16]. In our leptospiral model, the ¢bronectin-binding protein was only expressed by the virulent variant. Interestingly, in a previous study, the extent of adhesion to ¢bronectin-coated coverslips was correlated with the virulence of a Leptospira strain de¢ned using LD50 [10]; furthermore, `highly' virulent leptospires attached more e¤ciently than the `intermediately' virulent and avirulent variants of the same strain [10]. Many pathogenic bacteria express ¢bronectin-binding proteins which may be important virulence factors. As an example, ¢bronectin binding has been studied extensively in Staphylococcus aureus and is associated with invasiveness of the organism [17]. Fibronectin binding may lead to the colonisation of the mammalian host by attachment of pathogens to a variety of cells and tissues. Leptospires pass through intact skin and abrasions in the skin or mucosal surfaces of the host [1]. The wide distribution of ¢bronectin in the host may help adherence of virulent leptospires at the cutaneous or mucosal site of entry, and then in their dissemination within target organs after a bacteraemic stage. In experimental infection of Syrian hamsters [9] or guinea pigs [4] with a virulent strain of Leptospira, leptospires were observed in the liver in large number in the intercellular spaces where the extracellular matrix is rich in ¢bronectin. In conclusion, we have identi¢ed a ¢bronectin-binding protein exclusively expressed by a virulent variant of L. interrogans serovar icterohaemorrhagiae. Our results suggest that this 36-kDa protein localised on the cell surface may play a role in the attachment of virulent leptospires to host tissues. Indeed, the presence of a ¢bronectin-binding protein may be an essential prerequisite for the initial steps leading to infection. Virulent leptospires which lack expression of a ¢bronectin-binding protein would be unable to multiply within a host and be cleared rapidly. Considering the wide distribution of ¢bronectin within a host and the diversity of mammalian hosts implicated in the epidemiology of leptospirosis, ¢bronectin may contribute to the e¤cient attachment of virulent leptospires in di¡erent animal species involved in the natural history of this zoonosis. If molecular mechanisms of the adhesion of leptospires to eukaryotic membranes are poorly documented, their study would be critical for the future development of
vaccine candidates against leptospirosis. Future experiments should concern the underlying molecular mechanisms determining the phenotype of adhesion of virulent leptospires. Acknowledgements F.M. and P.P. were supported by the Institute Pasteur, Paris (International Network of Institutes Pasteur). The authors are indebted to Dominique Triballi for technical assistance. References [1] Faine, S., Adler, B., Bolin, C. and Perolat, P. (1999) Leptospira and Leptospirosis, 2nd edn. MediSci, Armadale. [2] Faine, S. (1964) Reticuloendothelial phagocytosis of virulent leptospires. Am. J. Vet. Res. 25, 830^835. [3] Merien, F., Baranton, G. and Perolat, P. (1997) Invasion of Vero cells and induction of apoptosis in macrophages by pathogenic Leptospira interrogans are correlated with virulence. Infect. Immun. 65, 729^738. [4] Merien, F., Truccolo, J., Rougier, Y., Baranton, G. and Perolat, P. (1998) In vivo apoptosis of hepatocytes in guinea pigs infected with Leptospira interrogans serovar icterohaemorrhagiae. FEMS Microbiol. Lett. 169, 95^102. [5] Vinh, T., Faine, S. and Adler, B. (1984) Adhesion of leptospires to mouse ¢broblasts (L929) and its enhancement by speci¢c antibody. J. Med. Microbiol. 18, 73^85. [6] Ballard, S.A., Williamson, M., Adler, B., Vinh, T. and Faine, S. (1986) Interactions of virulent and avirulent leptospires with primary cultures of renal epithelial cells. J. Med. Microbiol. 21, 59^67. [7] Thomas, D.D. and Higbie, L.M. (1990) In vitro association of leptospires with host cells. Infect. Immun. 58, 581^585. [8] Roth, J.A., Bolin, C.A., Brogden, K.A., Minion, F.C. and Wannemuehler, M.J. (1995) Virulence Mechanisms of Bacterial Pathogens, 2nd edn. ASM Press, Washington, DC. [9] Van den Ingh, T.S.G.A.M. and Hartman, E.G. (1986) Pathology of acute Leptospira interrogans serotype icterohaemorrhagiae infection in the Syrian hamster. Vet. Microbiol. 12, 367^376. [10] Ito, T. and Yanagawa, R.. (1987) Leptospiral attachment to four structural components of extracellular matrix. Jpn. J. Vet. Sci. 49, 875^882. [11] Thomas, D.D., Baseman, J.B. and Alderete, J.F. (1985) Fibronectin tetrapeptide is target for syphillis spirochete cytadherence. J. Exp. Med. 162, 1715^1719. [12] Probert, W.S. and Johnson, B.J.B. (1998) Identi¢cation of a 47 kDa ¢bronectin-binding protein expressed by Borrelia burgdorferi isolate B31. Mol. Microbiol. 30, 1003^1015. [13] Pettit, A. (1928) Contribution a© l'e¨tude des Spiroche¨tide¨s. Editions Me¨dicales, Paris. [14] Auran, N.E., Johnson, R.C. and Ritzi, D.M. (1972) Isolation of the outer sheath of Leptospira and its immunogenic properties in hamsters. Infect. Immun. 5, 968^975. [15] Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680^685. [16] Proctor, R.A. (1987) Fibronectin : a brief overview of its structure, function and physiology. Rev. Infect. Dis. 9, S317^S321. [17] Proctor, R.A. (1987) The staphylococcal ¢bronectin receptor: evidence for its importance of invasive infections. Rev. Infect. Dis. 9, S335^S340.
FEMSLE 9284 15-3-00