Decorin-binding adhesins from Borrelia burgdorferi

Abstract

Lyme disease is a tick-transmitted infection caused by the spirochete Borrelia burgdorferi. Ticks deposit B. burgdorferi into the dermis of the host, where they eventually become associated with collagen fibres. We demonstrated previously that B. burgdorferi is unable to bind collagen, but can bind the collagen-associated proteoglycan decorin and expresses decorin-binding proteins (Dbps). We have now cloned and sequenced two genes encoding the proteins, DbpA and DbpB, which have a similar structure, as revealed by circular dichroism (CD) spectroscopy of recombinant proteins. Competition experiments revealed a difference in binding specificity between DbpA and DbpB. Western blot analysis of proteinase K-treated intact B. burgdorferi and transmission electron microscopy studies using antibodies raised against recombinant Dbps demonstrated that these proteins are surface exposed. DbpA effectively inhibits the attachment of B. burgdorferi to a decorin substrate, whereas DbpB had a marginal effect, suggesting a difference in substrate specificity between the two Dbps. Polystyrene beads coated with DbpA adhered to a decorin-containing extracellular matrix produced by cultured skin fibroblasts, whereas beads coated with OspC did not. Taken together, these data suggest that Dbps are adhesins of the MSCRAMM (microbial surface component-recognizing adhesive matrix molecule) family, which mediate B. burgdorferi attachment to the extracellular matrix of the host.

Introduction

Lyme disease can affect multiple tissues including the skin, joints, heart and central nervous system (Steere, 1989; Barbour and Fish, 1993). The disease is caused by the spirochete Borrelia burgdorferi, an organism transmitted to mammalian hosts through a tick bite (Lane et al., 1991). During a bloodmeal, an infected tick of the Ixodes ricinus complex can deposit B. burgdorferi into the dermis of the host animal. Once in the skin, the spirochetes are thought to replicate and then disseminate to other organs. In several tissues, B. burgdorferi can be found in the extracellular matrix, often in close association with collagen fibres (Duray, 1992). We have shown previously that, although B. burgdorferi cannot attach directly to collagen fibres, the organisms can adhere to decorin, a collagen-associated proteoglycan, via two decorin-binding proteins (DbpA and DbpB) in the 20 kDa molecular weight range (Guo et al., 1995). Furthermore, we have found that the attachment of B. burgdorferi to a decorin substrate was time dependent and specific, in that soluble decorin, but not other tested matrix molecules, effectively inhibited bacterial adherence. In addition, B. burgdorferi also bound radiolabelled, soluble decorin in a specific, saturable process. A K_D of approximately 10⁻⁷ M was estimated for this interaction.

Decorin is a small proteoglycan that ‘decorates’ collagen fibres and is present in almost all tissues (Scott and Orford, 1981; Poole et al., 1986; Fisher et al., 1989). The core protein contains 10 leucine-rich repeats, a protein-folding motif found in a variety of proteins. Its single glycosaminoglycan chain is of the dermatan sulphate/chondroitin sulphate type (Yanagishita, 1993). Decorin has many activities including regulation of collagen fibril formation (Danielson et al., 1997), neutralization of transforming growth factor-β (Yamaguchi et al., 1990), suppression of cell growth (De Luca et al., 1996) and inactivation of the complement component C1q (Krumdieck et al., 1992). Decorin is therefore believed to play a role in tissue formation, inflammation and tumorigenesis. Dbps are the first bacterial proteins reported to have decorin-binding activity.

We demonstrate here that Dbps are surface exposed and can mediate attachment to a fibroblast matrix using decorin as the ligand. We propose that Dbps are adhesins of B. burgdorferi belonging to the MSCRAMM family (Patti et al., 1994).

Results

Cloning and nucleotide sequence of dbpA and dbpB

We have reported previously that B. burgdorferi can adhere to a decorin substrate and bind soluble decorin. Furthermore, B. burgdorferi strain N40 expresses two Dbps, as revealed by Western ligand blots, although it was unclear whether these proteins represented two distinct gene products or were expressed from the same gene (Guo et al., 1995). To isolate the Dbp gene(s), we used a λZAPII expression library constructed from B. burgdorferi strain 297 genomic DNA. Using digoxigenin-labelled decorin, we screened approximately 600 000 plaques and isolated three clones expressing functionally active Dbps. DNA sequencing revealed that two of the clones were identical, and the third contained a partially overlapping sequence. Further analysis of the nucleotide sequences revealed two 561 bp open reading frames (ORFs) separated by 117 bp of DNA within the overlapping region. These ORFs were subsequently shown to be genes encoding Dbps, and therefore designated dbpA and dbpB.

Analysis of the deduced amino acid sequences revealed signal peptidase II recognition sequences (Wu and Tokunaga, 1986), LISC for DbpA and LVAC for DbpB, within 30 residues of the N-terminus for both proteins. The protein sequences share 40% identity and 51% similarity (Fig. 1). Analysis of the deduced amino acid sequences suggested that the processed proteins have molecular weights of 20 kDa (DbpA) and 19 kDa (DbpB) (minus possible attached lipids) and isoelectric points of 8.8 and 9.3 respectively. The nucleotide sequences encoding these proteins were reported to GenBank by our laboratory previously (U75866 and U75867 for the dbpA and dbpB genes respectively) and confirmed recently by others (Feng et al., 1998; Hagman et al., 1998).

Figure 1. . PILEUP alignment of the deduced amino acid sequences of dbpA and dbpB. Black boxes indicate identical residues; grey boxes indicate similarly charged residues. Two gaps, represented by dots, were introduced to maximize the alignment. Putative signal peptidase II consensus sequences are boxed.

Purification, isolation and characterization of recombinant Dbps

The presence of putative signal peptidase II recognition sites within the deduced amino acid sequences of dbpA and dbpB suggested that they encode lipoproteins. Studies of other bacterial lipoproteins have shown that signal peptidase II cleaves at the N-terminal side of cysteine after lipid attachment (Wu and Tokunaga, 1986). To simulate the native proteins in their mature form, we used polymerase chain reaction (PCR) to construct recombinant Dbps with the corresponding amino acid truncations. The proteins were expressed as N-terminal polyhistidine (His-tag) fusions and purified by nickel-chelating chromatography. In constructing DbpB, we deleted the entire leader sequence, so the recombinant protein is only preceded by the His-tag sequence. Attempts to produce a similar form of DbpA resulted in a protein that expressed poorly and degraded rapidly. We therefore produced a form of DbpA that contains the His-tag sequence fused to six amino acids of the leader sequence, which are presumably removed in the mature native protein. This protein thus has an intact cysteine residue and tends to form dimers (Fig. 2A). To eliminate dimerization, we constructed a recombinant DbpA in which cysteine at position 25 is replaced by an alanine residue (DbpA:C25A).

Figure 2. . Analysis of recombinant Dbps. Purified proteins, DbpB (lane 1), DbpA (lane 2) and DbpA:C25A (lane 3), were subjected to SDS–PAGE (5–20% gradient) under reducing conditions and stained with Coomassie brilliant blue (A) or transferred to a nitrocellulose membrane (B). After blocking additional protein-binding sites, proteins on the membrane were probed with digoxigenin-labelled decorin and visualized by alkaline phosphatase reactivity. Circular dichroism spectra (C) of DbpA, DbpB and DbpA:C25A.

DbpB appears as a single band at approximately 19 kDa when analysed by SDS–PAGE (Fig. 2A). Purified DbpA appears as two bands: a monomer migrating as a 20 kDa protein and a dimer with an apparent molecular mass of 40 kDa (Fig. 2A). DbpA:C25A migrates as a single band at approximately 20 kDa by SDS–PAGE (Fig. 2A). Despite a greater tendency of DbpA to dimerize when subjected to SDS–PAGE, subsequent studies showed that DbpA:C25A and DbpA behave identically in all assays.

CD spectroscopy revealed that the overall secondary structure of DbpA, DbpA:C25A and DbpB are strikingly similar (Fig. 2C). Deconvolution of the spectra indicated that Dbps consist of approximately 50–60% α-helix. This secondary structure composition differs markedly from that of the recently reported crystal structure of the major outer membrane protein OspA, which is more than 90% β-sheet (Li et al., 1997).

Recombinant Dbps bind decorin

The decorin-binding activity of the recombinant Dbps was analysed by Western ligand blot (Fig. 2B). Purified proteins were separated by SDS–PAGE, transferred to a nitrocellulose membrane and incubated with digoxigenin-labelled decorin followed by alkaline phosphatase-conjugated anti-digoxigenin F_ab fragments. Decorin-binding proteins were visualized by phosphatase substrate reactivity. Both forms of recombinant DbpA and DbpB bound decorin (Fig. 2B). Biotin-tagged, recombinant DbpA and DbpB bound in a concentration-dependent fashion to decorin coated on microtitre wells (Fig. 3A). When unlabelled recombinant Dbps were used in attempts to inhibit these interactions, we found that unlabelled DbpA effectively inhibited the binding of biotin-labelled DbpA and DbpB to the decorin substrate (Fig. 3B), whereas the binding of biotinylated DbpB was inhibited only by unlabelled DbpB (Fig. 3C). Unlabelled DbpB did not affect the interaction of DbpA with decorin. These results demonstrated that, although DbpA and DbpB both bind decorin and have similar structures, they exhibited different ligand specificities.

Figure 3. . Specificity of DbpA and DbpB binding to decorin. Microtitre wells were coated with decorin, and additional protein binding sites were blocked with BSA. Increasing concentrations of biotinylated purified recombinant DbpA (closed circles) and DbpB (open circles) were allowed to bind to the wells (A). To inhibit binding, increasing concentrations of unlabelled DbpA (B) or DbpB (C) were preincubated with the wells before the biotinylated Dbps were added. Binding was measured by an ELISA-like assay. All data points were performed in triplicate and expressed as the mean ± standard error for each data point.

DbpA can inhibit the attachment of B. burgdorferi to decorin

We used a microtitre well attachment assay to study the adherence of B. burgdorferi to decorin. Initial experiments showed that strain N40 (Guo et al., 1995) could adhere to decorin-coated microtitre wells. The effects of purified, recombinant Dbps or OspC on bacterial adherence were examined in experiments in which decorin-coated wells were preincubated with increasing concentrations of recombinant Borrelia proteins for 1 h before whole B. burgdorferi cells were added. DbpA and DbpA:C25A effectively inhibited the adherence of B. burgdorferi to decorin. DbpB affected adherence only marginally, while OspC had no inhibitory effect (Fig. 4). These results suggest that DbpA plays an active role in the observed adherence of strain N40 to decorin-containing substrata. Identical results were obtained when recombinant Dbps and OspC were used to inhibit the attachment of strain 297 to a decorin substrate (data not shown). DbpA was also capable of inhibiting the binding of live B. burgdorferi 297 to decorin-coated wells (data not shown).

Figure 4. . Inhibition of the attachment of B. burgdorferi N40 to decorin substrata. Decorin-coated microtitre wells were preincubated with DbpA, DbpA:C25, DbpB or OspC before B. burgdorferi were allowed to attach. Borrelia attachment to the substrate was quantified by ELISA. This experiment was repeated three times with similar results.

B. burgdorferi expression of Dbps

To determine whether both Dbps are expressed by spirochetes grown in vitro, we analysed B. burgdorferi whole-cell lysates by Western blot using anti-Dbp antibodies that had been adsorbed to remove a low but detectable cross-reactivity (Fig. 5). Two proteins (approximately 20 kDa) expressed by B. burgdorferi 297 were recognized selectively by the adsorbed antibodies to DbpA and DbpB, respectively, demonstrating that both Dbps are expressed by this organism when grown in vitro (Fig. 5). No Dbps were present in the lysate from B. burgdorferi 297 G5 clone (Norton Hughes et al., 1993), a spontaneous laboratory mutant of 297 lacking the ospAB plasmid, which also encodes dbpA and dbpB (Sherwood Casjens, personal communication; Fraser et al., 1997; not shown).

Figure 5. . B. burgdorferi expression of Dbps and flagellin. Borrelia burgdorferi 297 were either treated with proteinase K for 10 min or left untreated before fractionation by SDS–PAGE (5–15% gradient) and transferred to nitrocellulose membranes. After blocking additional protein binding sites, proteins on the membrane were probed with adsorbed anti-DbpA antibodies, adsorbed anti-DbpB antibodies or anti-flagellin (Fla) (H9724) antibodies. Prebleed serum did not show any cross-reactivity (data not shown).

Dbps are surface exposed

To determine whether Dbps are surface exposed, we analysed the susceptibility of these proteins on intact borreliae to proteinase K digestion. As a control, we also analysed the Fla protein, a subunit of the periplasmic flagella, which should be protected against proteinase K when the outer membrane is intact. After proteinase K treatment, DbpA and DbpB could no longer be detected, whereas Fla appeared to be essentially unaffected (Fig. 5).

We used transmission electron microscopy (TEM) to confirm the localization of DbpA on the Borrelia membrane. B. burgdorferi 297 and strain 297 G5 clone were examined. The spirochetes were incubated with rabbit antisera raised against recombinant DbpA before allowing protein A-conjugated colloidal gold particles to bind. Anti-DbpA showed impressive labelling along the entire length of the membrane structure (Fig. 6A). In comparison, insignificant labelling was seen for the G5 clone (Fig. 6B) or preimmune IgG (Fig. 6C) controls. Similar experiments using anti-DbpB antibodies gave less clear results (data not shown).

Figure 6. . Transmission electron micrograph of B. burgdorferi 297 (A and C) and 297 G5 clone (B) incubated with anti-DbpA antibodies (A and B) or prebleed serum (C). Bound antibodies were visualized after incubating with protein A conjugated to colloidal gold.

DbpA binds to an extracellular matrix produced by cultured skin fibroblasts

Human skin fibroblasts grown for 48 h produce an extensive extracellular matrix. Immunofluorescence staining of these cultures revealed the presence of fibronectin in a characteristic fibrillar arrangement (Fig. 7A), as reported previously (Raghunath et al., 1993). The extracellular matrix produced by the cultured fibroblasts also contains decorin and an anti-decorin antibody localized to fibrillar structures in the matrix (Fig. 7C). When these cultures were incubated with recombinant DbpA, we demonstrated using appropriate antibodies that this Borrelia protein bound the produced extracellular matrix (Fig. 7B). A fibrillar staining pattern was observed similar to that seen with antibodies to fibronectin.

Figure 7. . Fluorescent visualization of a fibroblast cell matrix. Cell matrices were visualized with goat anti-rabbit rhodamine- conjugated secondary antibody after incubation with rabbit antibodies raised against (A) fibronectin and (C) decorin. Rabbit anti-DbpA was used to detect DbpA bound to the matrix (B). (D) is representative of negative controls, which included: DbpA + secondary antibody only (shown), secondary antibody alone or OspC + anti-DbpA + secondary antibody.

DbpA-coated beads attach to a fibroblast cell matrix

If DbpA is a B. burgdorferi adhesin, it should be able to mediate attachment to decorin present in an organized extracellular matrix. Other investigators have reported that B. burgdorferi can attach and invade various cultured cells, including fibroblast cells, using unidentified adhesins (Thomas and Comstock, 1989; Klempner et al., 1993) To avoid complications resulting from spirochete interactions with unidentified host cell ligands, we chose to develop a cell matrix attachment assay using purified recombinant proteins covalently coupled to fluorescent polystyrene beads.

DbpA, DbpB and OspC were coupled to fluoresbrite carboxylate polystyrene beads. Coupling efficiency was > 85%, as determined by quantifying protein in solution before and after coupling. The activity of the Dbps after coupling was analysed by incubating the beads with end-over-end rotation with approximately 50 000 counts per minute (cpm) of ¹²⁵I-labelled decorin (Table 1). Binding of DbpA beads to ¹²⁵I-labelled decorin was significantly higher than binding of beads coupled with DbpB (Table 1). The decorin binding activity of DbpB appears to have been lost during the coupling procedure. Therefore, we were only able to assess the attachment of DbpA-coated beads to the cell matrix.

Table 1.  . Binding of protein-coupled fluoresbrite beads. a. DCN, decorin; cpm, counts per minute.b. Mean and standard error of five fields (40 ×). NA, data not available.

Fibroblasts grown on glass coverslips were incubated with a suspension of protein-coated beads in 1% BSA in PBS (Fig. 8F). After extensive washing, the slides were mounted, and adherent beads were visualized by fluorescence microscopy (Fig. 8). The number of beads per field was counted for each treatment group and expressed as an average of five fields (Table 1). Beads coated with DbpA readily adhered to the fibroblast cultures (Fig. 8A), whereas beads coated with OspC did not (Fig. 8E). Furthermore, adherence of beads to the fibroblast culture was inhibited by preincubation of the fibroblasts with soluble recombinant DbpA (Fig. 8D) or preincubation of the beads with anti-DbpA antisera (Fig. 8B), but only marginally with preimmune sera (Fig. 8C). Taken together, these results indicate that DbpA is capable of mediating adherence to decorin when this molecule forms a substrate on its own or as part of an extracellular matrix assembled by cultured skin fibroblasts.

Figure 8. . Attachment of Dbp-coupled beads to fibroblast cell matrix. Human skin fibroblast cells were grown on slides and fixed with acetone. Cells were visualized by phase-contrast microscopy (F). Protein-coupled fluoresbrite polystyrene beads were allowed to attach. After extensive washing, the cells were mounted, and attached beads were visualized by fluorescence microscopy. A. DbpA-coated beads. B. DbpA-coated beads preincubated with anti-DbpA antisera. C. DbpA-coated beads preincubated with normal rabbit serum. D. Cell matrix preincubated with DbpA before the addition of DbpA-coated beads. E. OspC-coated beads.

Discussion

We have identified two genes from B. burgdorferi that encode decorin-binding proteins, DbpA and DbpB. dbpA and dbpB have been localized to a 54 kb linear plasmid, which also contains the ospAB operon (Sherwood Casjens, personal communication; Fraser et al., 1997). The G5 clone, a mutant of strain 297 that has lost this plasmid (Norton Hughes et al., 1993), does not express any detectable Dbps by Western blot analysis. Recent reports have indicated that, in strain B. burgdorferi N40, the dbpB/A genes are on a linear plasmid other than the one containing ospA/B (Feng et al., 1998).

The presence of putative peptidase II recognition sequences within both sequences indicated that the Dbps are lipoproteins. The deduced amino acid sequences of the Dpbs were compared with protein sequences in the National Center for Biotechnology Information databases using a gapped BLASTP search (Altschul et al., 1997). However, we found no striking similarities to any known proteins. The deduced amino acid sequences of DbpA and DbpB were aligned using the PILEUP program (Feng and Doolittle, 1987), (Fig. 1) and determined to share 40% identity. Their ‘normalized alignment score’ was calculated to be 350 (Doolittle, 1986), which indicates that the two proteins are certainly evolutionarily related.

Although the secondary structures of DbpA and DbpB are also very similar, as shown by CD analysis of the recombinant proteins (Fig. 2C), the proteins appeared to have different binding specificities. Studies using the recombinant proteins showed that DbpA could inhibit the binding of DbpB to decorin, whereas DbpB could not inhibit the binding of DbpA to decorin (Fig. 3). Thus, the ligand specificities of DbpA and DbpB differ. A K_D of 1–4 × 10⁻⁷ M was estimated from enzyme-linked immunosorbent assay (ELISA)-type binding studies in which DbpA was immobilized on microtitre wells and incubated with increasing concentrations of decorin. This fits well with the value estimated previously for the binding of decorin to intact bacteria. In a similar experiment, we could not saturate the DbpB-coated wells with decorin, suggesting that the K_D for this interaction is substantially higher than 10⁻⁷ M.

Our results demonstrated that DbpA and, most probably, also DbpB are present on the spirochete outer membrane. This result is consistent with the Dbps being lipoproteins. Furthermore, DbpA recognized decorin in an organized matrix produced by cultured skin fibroblasts and can mediate the adherence of protein-coated beads to fibroblast cultures. Taken together, these results suggest that at least DbpA acts as an MSCRAMM-type adhesin and mediates the adherence of B. burgdorferi to a decorin-containing substrate. We speculate that DbpA plays a critical role in the early stage of Lyme disease by mediating tissue adherence of B. burgdorferi, which enter the host during feeding of an infected tick. Adherence in the skin could facilitate survival of the small number of spirochetes (estimated to be between one and 10; Barthold, 1991) deposited by the tick. B. burgdorferi are thought to replicate in the skin before endovascular dissemination, and several lines of evidence suggest that survival is much more difficult once the spirochetes enter the bloodstream (Galbe et al., 1993; Sambri et al., 1996). If decorin binding is an early step in pathogenesis, Dbps would be excellent targets for vaccine development. Indeed, vaccine studies using Dbps demonstrated that mice passively or actively vaccinated with DbpA are protected when challenged with pathogenic B. burgdorferi (Hanson et al., 1998).

Spirochete adherence to decorin may also be involved in the later stages of Lyme disease. B. burgdorferi appear to associate with collagen fibres in the extracellular matrix of most infected tissues after disseminating from the skin, and decorin is found in most tissues. Additionally, decorin adherence could be a mechanism of cell invasion. B. burgdorferi has been shown to invade and survive in eukaryotic cells (Ma et al., 1991; Klempner et al., 1993; Brouqui et al., 1996). In cultured fibroblasts, proteoglycans are internalized and degraded in lysosomes by a secretion–recapture pathway (Schmidt et al., 1990). Decorin can bind to cell surface receptors and become internalized by receptor-mediated endocytosis. The decorin endocytosis receptor, a 51 kDa cell surface protein that binds decorin, was isolated recently (Hausser et al., 1996). Perhaps Dbps allow B. burgdorferi to use the decorin endocytosis receptor to invade cells by receptor-mediated endocytosis.

Bacterial binding to decorin may interfere with the proteoglycan's interactions with C1q and TGF-β and affect decorin's ability to regulate inflammation. Unregulated inflammation could lead to some of the clinical symptoms seen in Lyme disease. Thus, the Dbps of Borrelia burgdorferi may also contribute to the molecular pathogenesis of chronic Lyme disease.

Experimental procedures

Bacterial strains, culture and materials

Low-passage B. burgdorferi strain 297 (passage 6), strain 297 G5 clone (passage 3) and strain N40 (passage 2) were obtained and passaged a maximum of three times during the course of this study. B. burgdorferi was cultured in BSK II (Barbour–Stoenner–Kelly) medium at 34°C (Barbour, 1984). Cultures were incubated in a GasPak chamber (BBL) with 3–6% oxygen until the cells reached log phase. Cells were harvested by centrifugation at 14 500 × g for 30 min and gently washed in sterile, filtered phosphate-buffered saline (PBS; 137 mM NaCl, 3 mM KCl, 4 mM Na₂HPO₄, 1 mM KH₂PO₄, pH 7.4) three times. The spirochetes were resuspended in sterile PBS, and the cell density was adjusted to 1 × 10⁹ organisms ml⁻¹ by using a reference standard curve relating the A₆₀₀ to organism number as determined by darkfield microscopy. The spirochetes were stored at 4°C and retained decorin-binding activity for at least 1 month.

Escherichia coli strain M15 (Qiagen), XL-1 Blue (Stratagene) and SOLR (Stratagene) cells were grown at 37°C in Lennox broth (Difco) or Superbroth (2X-YT) (Maniatis et al., 1989), a medium recommended by Qiagen for the growth of E. coli M15, containing the appropriate antibiotics.

Human fibroblast skin cells (ATCC CRL-1475) were cultured on 16-well chamber slides (Nunc) in Dulbecco's modified essential media containing 10% fetal bovine serum (DMEM) at 37°C.

Labelling of decorin

Bovine decorin from fetal skin was purified as described previously (Choi et al., 1989). Purified decorin was stored in 4 M guanidine hydrochloride at −80°C and dialysed extensively against PBS before use. Decorin was labelled with digoxigenin-3-O-methylcarbonyl-e-aminocaproic acid-N-hydroxy-succinimide ester (digoxigenin) (Boehringer Mannheim) according to the manufacturer's instructions and stored at −20°C. To label decorin with biotin, 7.5 mg of NHS-LC-biotin [sulphosuccinimidyl-6-(biotinamido) hexanoate; Pierce] was dissolved in 100 μl of DMSO and combined with 0.5 mg of decorin and 0.5 ml of 0.2 M sodium borate (pH 8.0) in a total reaction volume of 1 ml. The mixture was incubated on an end-over-end rotator at room temperature (RT) for 2 h or at 4°C overnight, then dialysed against PBS and stored at −20°C. Decorin was radiolabelled by the chloramine T method. Ten microlitres (1 mCi) of [¹²⁵I]-Na (Amersham Life Science) was used to label 100 μg of decorin in 1 ml of PBS.

Cloning of dbpA and dbpB from B. burgdorferi 297

Digoxigenin-labelled decorin was used to screen a B. burgdorferi strain 297 λZAPII expression library (a gift from Dr Robin Isaacs, University of Mississippi Medical Center, Jackson, MS, USA). The library was screened using standard methods (Maniatis et al., 1989) with the following modifications. After blocking additional protein-binding sites on the filter lifts with a solution containing 3% bovine serum albumin (BSA), digoxigenin-labelled decorin was allowed to bind to proteins on the filter, followed by binding of anti-digoxigenin F_ab peroxidase (Boehringer Mannheim) to the digoxigenin-labelled decorin. Clones expressing Dbps were identified by developing the filters with 50% (w/v) 4-chloro-1-naphthol (Bio-Rad) in Tris-buffered saline (TBS; 20 mM Tris-HCl, 150 mM NaCl, pH 7.5) containing 17% methanol and 0.02% H₂O₂. The DNA inserts in pBluescript SK− (pBluescript) were excised from the λZapII vector with helper phage and transformed into SOLR cells according to the instructions from Stratagene.

Nucleotide sequencing

Nucleotide sequencing was performed with the Sequenase version 2.0 sequencing kit (US Biochemicals) according to the manufacturer's instructions and by automated sequencing (University of Texas at Houston Medical School, Molecular Genetics Core Facility, Houston, TX, USA). Oligonucleotides corresponding to sequences from the pBluescript clones were used as DNA sequencing primers. The final sequence was determined from both strands and confirmed by automated sequencing.

Construction of expression plasmids

DbpA, DbpB, OspC (outer surface protein C from strain N40) and a site-directed mutant of DbpA, DbpA:C25A, were constructed using PCR. The mutant was constructed using extension overlap PCR (Ho et al., 1989). Oligonucleotide primers used for PCR were 5′-CGCGGATCCACCAATCTTCTTAAACTA-3′ (B-N10F), 5′-GCGCTGCAGTTACGATTTAGCAGTGCT-3′ (EndP) for DbpA; 5′-CGCGGATCCTGTAGTATTG-

GATTAGTA-3′ (B64F), 5′-GCGCTGCAGTTATTTCTTTTTTTTGCT-3′ (564PR) for DbpB; 5′-CGCGCTGCAGGTAATAATTCAGGAAAAGATGG-3′ (no. 13), 5′-CGCAAGCTTTAA-

GGTTTTTTTGGACTTTCTGC-3′ (no. 12) for OspC; B-N10F, End P, 5′-GTTAGTCCTGCTGATATAA-3′ (C25A-R), 5′-TTATATCAGCAGGACTAAC-3′ (C25A-F) for DbpA:C25A. The bases incorporated for C25→A mutation are underlined. The primer B-N10F contains four base changes from the DNA sequence that do not affect the amino acid sequence of the protein. Oligonucleotides were purchased from Genosys Biotechnologies or Life Technologies. The reaction mixture contained at least 10 ng of template DNA, 1.5 μM forward and reverse primers, 2.5 mM MgCl₂, 0.2 mM each dNTP, 20 mM Tris-HCl (pH 8.4), 50 mM KCl and 2.5 units of Taq DNA polymerase in a 100 μl reaction volume. The mixture was overlaid with 100 μl of mineral oil and amplified for 30 cycles consisting of incubations at 94°C (denaturation) for 1 min, 50°C (annealing) for 2 min and 72°C (extension) for 3 min. After amplification, the PCR product was subjected to electrophoresis through a 1% SeaKem GTG agarose gel (FMC BioProducts), and the appropriate bands were excised and purified using the GeneClean kit (BIO101; Vista) according to the manufacturer's instructions. The PCR product was digested with PstI and BamHI and ligated with plasmid pQE-30 (Qiagen). The resulting plasmids were transformed by heat shock into competent E. coli M15 cells.

Expression and purification of recombinant proteins

A culture of E. coli harbouring the appropriate plasmid was grown in Superbroth until it reached an A₆₀₀ of 0.6. Isopropyl-β-D-thiogalactopyranoside (Life Technologies) was added to a final concentration of 0.2 mM, and the cells were incubated at 37°C for an additional 4 h. One litre of cells was harvested, resuspended in 10 ml of binding buffer (BB; 20 mM Tris-HCl, 0.5 M NaCl, 15 mM imidazole, pH 8.0) and lysed by a French pressure cell (11 000 pounds inch⁻²). The lysate was centrifuged at 40 000 × g for 15 min, and the supernatant fraction was filtered through a 0.45 μm filter.

A 1 ml iminodiacetic acid Sepharose column (Sigma) was charged with 75 mM NiCl₂·6H₂O and equilibrated with BB. The filtered supernatant was applied to the column and washed with 10 volumes of BB, then 10 volumes of BB containing 60 mM imidazole. Bound proteins were eluted with BB containing 200 mM imidazole, dialysed against PBS containing 10 mM EDTA, then dialysed against PBS. Protein concentrations were determined by the bicinchoninic acid protein assay (Pierce).

Circular dichroism spectroscopy

Purified recombinant proteins were dialysed into 10 mM Na₂HPO₄ (pH 7.4) at a concentration of 50 μg ml⁻¹. Circular dichroism (CD) spectroscopy measurements were taken using a Jasco J720 spectropolarimeter calibrated with a 0.06% (w/v) 10-d-camphorsulphonic acid ammonium salt solution. Measurements were taken at room temperature in a 0.2 mm path length quartz cell. All far-UV (250–190 nm) spectra were acquired with a time constant of 1 s, a scan rate of 20 nm min⁻¹, and four scans accumulated and then averaged. The molar ellipticity (theta) is expressed in degcm dmol⁻¹.

Antibodies

Rabbit antisera 34289 (HTI Bio-Products) and R625 (BioDesign International) were raised against DbpB and DbpA respectively. Both proteins were purified by nickel-chelating chromatography and cation exchange chromatography. Antisera from the second bleeds were used for the assays in this work. Both antisera were titred to more than a 1:50 000 dilution and showed a small but detectable cross-reactivity.

To adsorb antibodies to remove Dbp cross-reactivity completely, Dbps were coupled to cyanogen bromide-activated Sepharose 6B (Pharmacia) according to the manufacturer's instructions. The appropriate antisera were diluted 1:100 and incubated with 50 μl of protein-coupled Sepharose beads on an end-over-end rotator for 1 h in a final volume of 1.1 ml. The beads were centrifuged, and the adsorbed antisera was collected.

Anti-OspA rabbit serum was generated against purified recombinant OspA lipoprotein (Erdile et al., 1993) derived from B. burgdorferi strain B31 (antiserum was a gift from Mark Hanson, MedImmune, Gaithersburg, MD, USA). Rabbit antisera raised against decorin core protein was a gift from Hans Kresse (University of Münster, Germany) (Raghunath et al., 1993). Monoclonal antibody H9724 raised against B. burgdorferi flagellin was a gift from Steven Norris (University of Texas Medical School, Houston, TX, USA) (Norris et al., 1992). Anti-human fibronectin rabbit serum was generated against human fibronectin isolated from freshly frozen plasma as described previously (Froman et al., 1984).

SDS–PAGE and Western ligand blot

Proteins were fractionated by SDS–PAGE (Laemmli, 1970) and probed using a Western ligand blot assay. For SDS–PAGE, 2 × 10⁷B. burgdorferi cells or 4 μg of purified Dbp or OspC (not shown) were boiled in sodium dodecyl sulphate (SDS) for 3–5 min under reducing conditions and subjected to electrophoresis through a 5–15% or 5–20% gradient acrylamide slab gel at 150 V for 45 min. The separated proteins in the gel were stained with Coomassie brilliant blue.

For Western ligand blot assays, the proteins were transferred from the polyacrylamide gel to a nitrocellulose membrane (Schleicher & Schuell) by electroblot for 1.5 h at 4°C. Additional protein binding sites on the membrane were blocked by incubating in 5% non-fat dry milk in TBST (0.15 M NaCl, 20 mM Tris-HCl, 0.05% Tween 20, pH 7.4) for 2 h at room temperature or overnight at 4°C, followed by three 5 min washes in TBST. The membrane was then incubated at room temperature with 0.5 μg of digoxigenin-labelled decorin ml⁻¹ TBST for 1 h, washed and incubated with 1:3000 anti-digoxigenin F_ab alkaline phosphatase conjugate (Boehringer Mannheim) in TBST for 1 h. The membrane was washed, and decorin-binding proteins were visualized with 5-bromo-4-chloro-3-indoyl phosphate p-toluidine salt (BCIP) and p-nitroblue tetrazolium chloride (NBT) (Bio-Rad). No cross-reactivity of digoxigenin-labelled decorin with OspC or any other Borrelia proteins was detected (not shown). Stock solutions of NBT and BCIP were made according to the manufacturer's instructions and diluted in carbonate bicarbonate buffer (14 mM Na₂CO₃, 36 mM NaHCO₃, 5 mM MgCl₂·6H₂O, pH 9.8) to a final concentration of 300 μg NBT and 150 μg BCIP ml⁻¹ buffer.

Western blot assays and proteinase K treatment of B. burgdorferi

For Western blot assays, protein from 1 × 10⁹B. burgdorferi or 15 μg of purified recombinant Dbp was fractionated and transferred to membranes as described above for the Western ligand blot assay. The membrane was blocked and washed as described above. All incubation steps were performed using TBST–1% milk. For probing proteinase K-treated B. burgdorferi (Norris et al., 1992), adsorbed anti-DbpA and anti-DbpB were used at dilutions of 1:50 000 and 1:25 000, respectively, and anti-flagellin (Fla) at 1:2000. For probing whole-cell lysates, adsorbed anti-DbpA was used at a dilution of 1:25 000 and adsorbed anti-DbpB at 1:4000. For probing purified recombinant protein Dbps, anti-DbpA was used at a dilution of 1:100 000 and anti-DbpB at 1:25 000. After incubating in primary antibody for 1 h and washing as described above, the membranes were incubated in a 1:3000 dilution of goat anti-rabbit alkaline phosphatase conjugate or goat anti-mouse alkaline phosphatase conjugate in TBST–1% milk for 1 h. The membrane was washed, and decorin-binding proteins were visualized as described above.

Dbp decorin binding assay

Purified recombinant Dbps were labelled with EZ-Link p-aminobenzoyl biocytin (Pierce), resulting in biotinylation of tyrosine and histidine residues. p-Diazobenzoyl biocytin (DBB) was prepared from its stable precursor, p-aminobenzoyl biocytin, for reactivity according to the manufacturer's instructions. Protein (0.5 ml; 0.5–0.8 mg ml⁻¹) in PBS was incubated with 0.475 ml of 0.2 M boric acid (pH 8.5) and 25 ml of DBB. The reaction was allowed to incubate at room temperature on an end-over-end rotator for 1 h, then immediately dialysed against PBS at 4°C. The proteins were stored at 4°C for up to 2 weeks. Immulon-1 microtitre plate wells (Dynatech Labs) were coated with 1 μg of decorin in 50 μl of PBS by incubating at 4°C overnight. Remaining protein binding sites in the wells were blocked by incubating with 1% BSA (w/v) in PBS for 1 h. The wells were washed three times for 5 min each with 200 μl of PBS containing 0.1% BSA (PBSB) and incubated with 100 μl of the indicated concentrations of biotinylated Dbp diluted in PBSB for 1 h. After decanting and washing, the wells were incubated for 1 h with 100 μl of alkaline phosphatase streptavidin (US Biochemicals) diluted 1:3000 in PBSB. After decanting and washing, the wells were incubated for 30 min at 37°C with 100 μl of 1 mg Sigma 104 phosphatase substrate ml⁻¹ DEA (1 M diethanolamine, pH 9.8, 0.5 mM MgCl₂).

To inhibit Dbp attachment to decorin, unlabelled Dbps were preincubated with decorin-coated wells before 0.3 μg of biotinylated Dbp was allowed to attach. Microtitre wells were coated, blocked and washed as described above, then incubated for 1 h with 100 μl of the indicated concentration of unlabelled Dbp in PBSB. The wells were washed, and the assay was continued as described above. Protein concentrations were assayed using the bicinchoninic acid (BCA) protein determination kit (Pierce).

B. burgdorferi decorin attachment assay

Microtitre plate wells were coated with decorin, blocked and washed as described for the Dbp decorin binding assay. The wells were then incubated with the inhibitor at the indicated concentration in a total volume of 100 μl for 1 h. After decanting, the wells were incubated for 1 h with 25 μl of a suspension containing 1 × 10⁹ organisms ml⁻¹ PBS containing 0.1% BSA. To remove unattached bacteria, the wells were then washed three times for 5 min with 200 μl of PBSB, then incubated for 1 h with 100 μl of a 1:5000 dilution of anti-OspA serum. The wells were washed and incubated with 100 μl of 1:1000 goat anti-rabbit alkaline phosphatase conjugate diluted in PBS, 0.1% BSA for 1 h, then washed and incubated with 100 μl of 1 mg ml⁻¹ Sigma 104 phosphatase substrate dissolved in 1 M diethanolamine, 0.5 mM MgCl₂, pH 9.8, at 37°C for 30 min. The absorbance at 405 nm was determined in a microplate reader (Molecular Devices).

Transmission electron microscopy

Mid-log phase cultures (2 ml) were harvested by centrifugation at 600 × g for 5 min. The resulting pellets were gently resuspended into 0.2 ml of Hanks' buffered salt solution (Life Technologies). Droplets containing 0.05 ml of cell suspension were placed on the freshly uncovered surface of Parafilm. Copper/paladium electron microscopy grids were coated on the paladium surface with Parlodion (Ted Pella) and floated, coated side down, on the droplets. After 2 min, the grids and adherent spirochetes were transferred onto droplets of blocking solution containing 5% (m/v) immunoglobulin-free BSA (Sigma) in Hanks' buffer and incubated for 20 min. Blocked grids were incubated for 20 min on 1:100 dilutions of antisera in blocking solution, then washed twice for 10 min in blocking solution. The samples were labelled by incubation on droplets containing a 1:50 dilution of protein A, conjugated to 10 nm colloidal gold (Ted Pella), in blocking solution. After labelling, the grids were washed briefly in three changes of deionized water and negatively stained in 0.5% ammonium molybdate. The samples were examined on a Hitachi HU-11 E transmission electron microscope operated at 75 kV (Hanson et al., 1998).

Cell matrix attachment assay

Proteins were coupled to 1-μm-diameter fluoresbrite carboxylated polystyrene microparticles or beads (Polysciences) using the carbodiimide method according to the manufacturer's instructions. A 2.5% (w/v) suspension of fluoresbrite carboxylated polystyrene beads was incubated with 680–800 μg of DbpA, DbpB or OspC in a final volume of 1 ml. To assess the decorin binding ability of the protein-coated beads, 20 μl of beads in a 2.5% suspension was added to 3 ml of PBS, 0.1% BSA, 0.001% Tween 20 and approximately 50 000 cpm of ¹²⁵I-labelled decorin. The mix was allowed to incubate for 1 h, and the amount of labelled decorin bound to the beads was quantified after washing.

Human skin fibroblasts were cultured and fixed as described previously (Godfrey et al., 1990). Briefly, cells were plated onto 16-well chamber slides at a density of 2.5 × 10⁵ cells ml⁻¹, grown for 2 or 3 days then fixed with −20°C acetone. Unoccupied protein binding sites on the slides were blocked with 1% BSA in PBS for 1 h. The cells were incubated for 1 h with a 100 μl suspension of fluoresbrite beads (2.5%) coupled to the appropriate protein in a solution of PBS containing 1% BSA. The cells were washed three times for 5 min each with PBS, briefly flushed with excess PBS and then mounted with Universal mount (Research Genetics).

To inhibit bead attachment with purified protein, cells were preincubated for 1 h with 100 μg ml⁻¹ DbpA or OspC in PBS containing 1% BSA before allowing protein-coupled fluoresbrite beads to attach. To inhibit bead attachment with antisera, protein-coupled beads were preincubated for 1 h with a 1:10 dilution of anti-DbpA or prebleed serum in PBS containing 1% BSA. The entire mixture was then added to the cells to allow attachment.

Immunofluorescence

Fibroblast cells were grown and fixed as described above. After fixing, the cells were washed and blocked with 60 μl of PBS containing 5% horse serum and 1% goat serum. All washes were performed by immersing the slides in PBS for 10 min twice. The blocking solution was carefully aspirated after a 1 h incubation at room temperature, and 0.1 μg of DbpA (549) or OspC was added in a 60 μl volume for 20 min at room temperature. After washing, 60 μl of a 1:500 dilution of anti-DbpA or anti-OspC was added for 1 h at room temperature. After washing, a final incubation with 60 μl of rhodamine-conjugated goat anti-rabbit IgG (secondary antibody; Cappel) was added for 30 min at room temperature in the dark. After washing, the slide was visualized using fluorescence microscopy. Photographs were taken at 40 × magnification using Kodak 1000 speed film. All antibody and protein dilutions were made using blocking buffer. Control wells included a secondary antibody-only well, anti-DbpA + secondary antibody, and DbpA or OspC with secondary antibody only.

Direct staining of the matrix was performed using polyclonal rabbit anti-human fibronectin (HFN) and decorin antibodies. Experiments using these antibodies were carried out as described above. Anti-HFN and anti-decorin were diluted 1:500 in 60 μl of blocking buffer and incubated for 30 min at room temperature. Subsequent steps were performed as described above.

Nucleotide sequence accession number

The GenBank accession numbers for the complete dbpA and dbpB sequences presented in this article are U75866 and U75867 respectively.

Footnotes

1. Present address: Harvard Medical School, Department of Microbiology and Molecular Genetics, Boston, MA 02115, USA

Acknowledgements

This work was supported by NIH grant AR 44415, Medimmune and a Grant-in-Aid of research grant from Sigma Xi, the Scientific Research Society (B.P.G.) and the Vessa Notchev Fellowship (B.P.G.). We thank Robin Isaacs for the λZAPII B. burgdorferi 297 genomic library and strain 297, Steven Norris for B. burgdorferi N40, Carrie Norton Hughes for B. burgdorferi 297 G5 clone, and Hans Kresse for the anti-decorin antibody. We thank Colleen Jennings and Young-Ah Kim for technical assistance, Steve LaBrenz for assistance with CD experiments, and Mark Hansen, MedImmune, for helpful discussions.