INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Lyme disease (41) or Lyme borreliosis, is caused by a group of related tick-borne spirochetes classified as Borrelia burgdorferi sensu lato (including B. burgdorferi sensu stricto, B. afzelii, and B. garinii). Recent clinical trials have shown that monovalent recombinant subunit vaccines composed of the Borrelia outer surface protein A (OspA) lipoprotein were efficacious through two Lyme disease transmission seasons (40, 42). The mechanism of this protective effect differs from that of other vaccines. The OspA protein is expressed by spirochetes in the tick midgut, but this protein is down regulated during tick engorgement (13) and in the mammalian host (7). Protection by immunization with OspA therefore involves prevention of transmission of the spirochetes from the tick to the mammalian host and is dependent on having a critical threshold level of antibodies at the time of the tick bite (12). The addition of mammalian host stage antigens to the OspA vaccines may extend the duration or enhance the level of protective efficacy of such transmission-blocking vaccines (28). Alternatively, vaccines composed of one or more mammalian-stage antigens may be effective without OspA.

Several B. burgdorferi proteins expressed in the mammalian stage have been shown to be effective vaccines for preventing infection in laboratory animals challenged by experimental or natural routes. These protective antigens include OspC, P35/BBK32, P66/Oms66, and decorin binding protein (14, 17, 19, 20, 25, 27, 34). Decorin binding proteins A and B (DbpA and DbpB) are B. burgdorferi lipoproteins (23, 27) that are surface exposed and may act as spirochetal adhesins (24). We have demonstrated that immunization of mice with DbpA protected them from challenge with cultured spirochetes (27), and others (17, 25) have confirmed this protection. DbpA is expressed in vivo during spirochetemia in the mouse model (7) and is recognized by human Lyme disease patient sera (8, 29). These data suggest a potential role for DbpA in an improved Lyme vaccine.

Studies of DbpA vaccine effectiveness in other laboratories have relied on Escherichia coli vectors expressing cytosolic products as fusions to affinity tag sequences, a commonly used strategy for generating recombinant immunogens. However, recombinant cytosolic DbpA expressed as amino-terminal fusions to either polyhistidine (25) or glutathione S-transferase (GST) (17) was less than completely effective in these other studies. In the present study we sought to determine which form of the DbpA antigen would be most effective as a vaccine antigen and whether the protective efficacy of the protein was conformationally dependent. Due to the ability to measure ligand binding activity of the DbpA forms, we took advantage of the opportunity of being able to correlate function and therefore correct folding with the ability to elicit a protective response.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

B. burgdorferi and culture conditions. Cloned strains of B. burgdorferi sensu stricto isolate N40 (3) and mouse-adapted B. afzelii isolate PKo (2) were donated by S. Barthold. Isolate B. garinii VSBP was donated by R. Johnson (44). Spirochetes were propagated in tightly closed containers at 33 or 37°C in modified Barbour-Stoenner-Kelly (BSKII) medium (1). The cell densities of these cultures were determined by dark-field microscopy at ×400.

Expression and purification of recombinant proteins. Expression in E. coli BL21(DE3)pLysS and purification of chimeric lipoprotein Lpp:DbpA_N40 have been described previously (7). Lpp:DbpA_N40H₆ was expressed from plasmid pWCR129 (27) in E. coli BL21(DE3)pLysS. Membrane-associated proteins were solubilized in the detergent 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS) as described previously (7), and LppDbpA_N40H₆ was purified with minor modifications as follows. The CHAPS soluble fraction was loaded onto Ni-nitrilotriacetic acid Sepharose (Qiagen, Valencia, Calif.) equilibrated in 20 mM NaPO₄ (Na₂HPO₄ and NaH₂PO₄ mixed to attain pH 8.0)-300 mM NaCl-10 mM CHAPS. The protein was eluted using a linear gradient of the starting buffer with 250 mM imidazole. The eluted protein was dialyzed against phosphate-buffered saline (PBS) (pH 7.4) (HyClone, Logan, Utah)-8 mM CHAPS

Decorin binding assays. A solid-phase plate assay was developed for measuring binding to decorin. Recombinant human decorin (39) was obtained from David Mann. Wells of Maxisorb (Nunc) plates were coated with 100 µl of decorin solution at 0.5 mg/ml in PBS and incubated overnight at 4°C. After unbound decorin was decanted, the plates were blocked with 5% nonfat dried milk in PBS-0.1% Tween 20. Samples of DbpA were applied to the plate in blocking buffer at different concentrations and incubated for 1 h at room temperature. The plates were washed with PBS-Tween and incubated with a 1:2,500 dilution of rabbit polyclonal sera against Lpp:DbpA_N40. Rabbit hyperimmune antiserum was raised as described previously (27). The plates were washed and incubated with a 1:2,500 dilution of goat anti-rabbit-horseradish peroxidase conjugate antibody (Kirkegaard & Perry Laboratories, Gaithersburg, Md.) for 1 h at room temperature. The binding of DbpA-antibody complex to the decorin-coated plate was visualized by addition of 2,2'-aminobis(3-ethylbenzthiazolinasulfonic acid) (ABTS) substrate (Kirkegaard & Perry Laboratories). For assays of the heat-denatured protein, the DbpA was incubated at 65°C for 1 h prior to assay. Urea denaturation was performed by addition of urea to 8 M and incubation of the sample for 30 min at 37°C. The renaturation was performed by dilution 1:100 into PBS-0.1% Tween 20 or by dialysis into PBS-10 mM CHAPS.

Immunization and challenge. Pathogen-free female C3H/HeJ mice were purchased from The Jackson Laboratory (Bar Harbor, Maine) and used at 6 to 8 weeks of age. Prior to immunization, DbpA or control antigens were formulated in one of five ways. Proteins were emulsified 1:1 (vol:vol) with Freund's adjuvant (Difco) or adsorbed with the aluminum adjuvant Alhydrogel (Superfos Biosector, Kvistgård, Denmark) and administered either intraperitoneally or subcutaneously in a volume of 100 µl. The primary immunizations with Freund's adjuvant were performed in complete Freund's adjuvant and were followed by a second immunization 4 weeks later in incomplete Freund's adjuvant. Some mice were immunized a third time after another 4-week interval. Protein was diluted in PBS and injected into vials containing trehalose dimycolate plus monophosphoryl lipid A, as supplied by the manufacturer (RIBI Immunochem Research, Inc., Hamilton, Mont.). The vials were vortexed until an emulsion was formed, and the material was injected into mice subcutaneously at a volume of 200 µl. Quil A (Sigma, St. Louis, Mo.) was added to protein to a final concentration of 250 µg/ml. For Alhydrogel + Quil A, protein was allowed to bind to Alhydrogel first and then Quil A was added to a final concentration of 250 µg/ml. Immunogen was injected subcutaneously at a volume of 100 µl. The mice were challenged at 2 weeks after the final immunization with 10⁴ B. burgdorferi N40. Infection status was determined from BSKII cultures of five tissues as described previously (27).

In vitro growth inhibition assay and ELISA. Microwell titer determinations for determination of growth-inhibitory titers and immunoglobulin G (IgG) enzyme-linked immunosorbent assay (ELISA) titers of DbpA antisera were performed as described previously (27). Spirochetes in BSKII growing for 3 to 5 days at 33°C, or 37°C for B. afzelii and B. garinii, in the presence of serially diluted antiserum were enumerated by dark-field microscopy at a magnification of ×400. The inhibition end-point titer was determined as the dilution of antiserum promoting a >85 to 90% reduction in the number of motile spirochetes compared to control samples, which typically yielded 50 to 75 spirochetes per microscopic field.

In vitro proteolysis. Samples of H₁₀-DbpA_N40, Lpp:DbpA_N40, Lpp:DbpA_N40-H₆, and the corresponding cysteine mutant form were incubated at either room temperature or 37°C in the presence of 10% (wt/wt) clostripain (Worthington Biochemical Corp., Lakewood, N.J.)-10 mM CHAPS-1 mM CaCl₂. Clostripain was activated in 2.5 mM dithiothreitol-1 mM CaCl₂ prior to use in the assay. The proteolysis was terminated at different times by addition of reaction aliquots into EDTA to a concentration of 25 mM. Samples were boiled in SDS sample buffer and run on a NuPage (Invitrogen, Carlsbad, Calif.) 10% acrylamide gel in morpholineethanesulfonic acid (MES) buffer provided by the vendor. Proteins were visualized on a Molecular Dynamics Storm (Amersham-Pharmacia Biotech, Piscataway, N.J.) imager by fluorescence staining with Sypro Red.

Circular dichroism (CD). All spectra were obtained using a Jasco 810 spectrapolarimeter (Jasco, Easton, Md.) equipped with a Peltier temperature controller in a 1-mm pathlength cell. Thermal denaturations were performed while collecting data at 220 nm using a scan rate of 1°C/min from 20 to 80°C. Samples for denaturation were used at a concentration of 0.1 mg/ml in PBS with 10% (wt/vol) CHAPS detergent. For full spectral analysis, samples were dialyzed into 20 mM NaPO₄-10 mM Zwittergent 3-12 (pH 8.0) prior to analysis to remove CHAPS interference of the spectra. Full-scan spectra were obtained by scanning from 260 to 180 nm at a rate of 100 nm/min and averaging two spectra.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Protection of mice from Borrelia infection by immunization with different recombinant forms of DbpA. We have shown previously that both passive and active immunization can protect mice from infection with both cultured (27) and in vivo-derived (7) borreliae. Our previous work suggested that different forms of the antigen may vary in their vaccine efficacy. Accordingly, we set out to determine the form of the antigen giving the most protective immune response. We expressed DbpA from B. burgdorferi strain N40 either as translocated/lipidated forms or as a cytoplasmically expressed form for the present study. The salient amino terminal and carboxy-terminal sequences of these different constructs are shown in Fig. 1.

View larger version (12K): [in this window] [in a new window]   FIG. 1.   Sequence comparison of five recombinant forms of DbpAN40 and the natural sequence. The amino-terminal and carboxy-terminal residues of the mature lipoprotein forms of DbpAN40, both natural (top line) and chimeric recombinant lipoprotein (Lpp) forms, and the nonlipoprotein forms (H10 or H6) are shown. (PAM)3- indicates the presumed tripalmitoyl posttranslational modification to the amino-terminal cysteine of the lipoprotein forms. Nonnatural vector-encoded or mutated residues are underlined and lowercase. The central 135 amino acids are omitted from the diagrams for clarity.

Recombinant chimeric DbpAs differ in their functional activity. In vitro studies have shown that DbpA binds with high specificity to decorin, an extracellular matrix dermatan sulfate proteoglycan (23, 24). At this time, DbpA is one of the few B. burgdorferi surface proteins with a biochemically defined and quantifiable function. It seemed likely that this activity was dependent on proper DbpA conformation and, if so, we could potentially exploit this structure-function relationship as a tool for defining conformational serological epitopes. We developed a solid-phase equilibrium binding assay to assess the ability to discriminate among the different forms of DbpA in their binding to decorin. DbpA bound to decorin-coated microwell plates in a concentration-dependent and saturable manner. Decorin bound similarly to immobilized DbpA (data not shown). To minimize variables in plate-coating conditions among different DbpA forms, we used the immobilized decorin format. As shown in Fig. 2A, the secreted and lipidated forms of DbpA bind to the decorin on the plate while the cytoplasmically produced amino-terminal histidine-tagged DbpA does not bind substantially at the same concentration. For this comparison, we substituted H₁₀DbpA_N40, another cytosolic form of the protein, for H₆DbpA_N40 since they were essentially equivalent in decorin binding activity and CD spectra (data not shown). Additionally, H₁₀DbpA_N40 was expressed and purified in higher yield, and use of this form allowed confirmation that some unidentified property of the amino-terminal sequence of H₆DbpA_N40 did not influence the structure and function of DbpA.

View larger version (14K): [in this window] [in a new window]   FIG. 2.   Binding of different forms of DbpAN40 to immobilized decorin. (A) Binding of Lpp:DbpAN40H6 (solid squares), Lpp:DbpAN40 (solid triangles) and H10-DbpAN40 (solid diamonds) to decorin. (B) Binding of LppDbpAN40H6 (solid squares) treated with urea (solid triangles) or after heat treatment at 65°C for 1 h (solid diamonds). A405nm absorbance at 405 nm.

Alterations to DbpA structure diminish protective immune responses. The above data suggested a correlation between the ability of the DbpA to bind to decorin and its ability to elicit a protective immune response in mice. If this is so, then perturbations of the protein structure that lead to a decrease in the ability to bind to decorin should also lead to a diminution in its potency as a protective immmunogen. In an attempt to generate such a functionally altered immunogen, we tested the ability of DbpA to bind to decorin following the unfolding of the protein by either heat denaturation or urea denaturation. As shown in Fig. 2B, the ability of the DbpA to bind to the decorin is retained in samples treated with 8 M urea but is lost on heat treatment at 65°C for 1 h.

DbpA antibodies that are cross-reactive and borreliacidal for different B. burgdorferi sensu lato species are directed against conformational epitopes. We showed previously that antiserum against a single DbpA immunogen could kill diverse isolates of B. burgdorferi, B. garinii, and B. afzelii (27). These diverse isolates had substantial heterogeneity in their DbpA sequences, which prompted speculation that the epitopes binding the cross-reactive borreliacidal antibodies were composed of discontiguous amino acids (38). The availability of antisera against forms of DbpA_N40 that are conformationally and antigenically distinct allowed us to test this hypothesis directly.

View larger version (67K): [in this window] [in a new window]   FIG. 3.   Decorin binding activity blot. Protein samples on a polyvinylidene difluoride membrane were probed with biotinylated decorin and were visualized using a streptavidin-alkaline phosphatase conjugate using ECF reagent. Lanes: 1 Lpp:DbpAN40H6; 2, Lpp:DbpAPko; 3, Lpp:DbpAVSBP-H6; 4, OspCN40.*, migration position of OspC.

Structural stability of different recombinant DbpA forms. The immunization experiments indicated a substantial difference in the epitope structures of Lpp:DbpA_N40H₆ and Lpp:DbpA_N40H₆(C176, 191S) that was not evident by CD or decorin binding assays. We next used the Arg-specific protease clostripain as a probe of the structural differences between Lpp:DbpA_N40H₆ and Lpp:DbpA_N40H₆(C176, 191S), reflected in a differential accessibility of their four Arg residues. Lpp:DbpA_N40 and H₁₀DbpA_N40 were also included in this comparison. At room temperature, incubation with 10% (wt/wt) clostripain resulted in a shift from the full-length Lpp:DbpA_N40H₆ molecule to a carboxy-terminally truncated form of the protein that was confirmed by Western blotting with a monoclonal antibody against the histidine tag. This single-cleavage product accounts for 50% of the remaining protein after 60 min of incubation (Fig. 4). In marked contrast, the (C176, 191S) mutant had lost 82% of the total protein after 15 min of incubation. The cysteine mutant DbpA protein was cleaved to a ~14-kDa limited digestion product which did not accumulate in the unaltered Lpp:DbpA_N40H₆ protein. At 37°C, the unaltered Lpp:DbpA_N40H₆ protein was nominally more sensitive to cleavage than at room temperature, whereas the cysteine mutant was almost completely degraded after 10 min of incubation. Lpp:DbpA_N40 and H₁₀DbpA_N40 showed only a limited sensitivity to clostripain digestion at either temperature, and a substantial amount of each full-length protein remained after the 60-min incubation. These data suggested that the cysteine mutant is less stable or structurally rigid than the unaltered Lpp:DbpA_N40H₆ protein and the other two forms of DbpA. The (C176, 191S) mutant protein was also the least effective vaccine antigen among these four.

View larger version (49K): [in this window] [in a new window]   FIG. 4.   Sensitivity of recombinant forms of DbpAN40 to proteolysis. Proteins were incubated with 10% (wt/wt) clostripain at either 22 or 37°C. Aliquots were taken at the times indicated, and the reactions were quenched with 50 mM EDTA. Samples were run on a 10% Bis-Tris polyacrylamide gel in MES buffer, and proteins were visualized by Sypro Red staining.

View larger version (20K): [in this window] [in a new window]   FIG. 5.   CD thermal denaturation scans of four recombinant forms of DbpAN40. Ellipticity was measured at 220 nm as a function of temperature for Lpp:DbpAN40H6 (), Lpp:DbpAN40 (-----), H10DbpAN40 (----), and Lpp:DbpAN40H6 (C176, 191S) (-----).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Conformational B-cell epitopes contribute to the effectiveness of many vaccine antigens. Biochemical confirmation of native folding is not possible for antigens lacking a quantifiable function. With a few possible exceptions (14, 33), this is true of experimental and approved vaccines for B. burgdorferi. The biochemical activity of decorin binding proteins is measurable by numerous assays (23, 24) and appears to promote a physiologically relevant function, extracellular matrix adherence, for the borreliae. Point mutants of DbpA₂₉₇ have been made and tested for their ability to bind to decorin in various assay formats. These studies indicate lysine residues that may be important to decorin binding either directly or as a consequence of their affects on the conformation of the DbpA (6). We have now demonstrated further evidence of the importance of conformation in the activity of DbpA, and we have shown that eliciting a potent, protective immune response to DbpA required a properly folded form of the immunogen. Denaturation or mutations in the DbpA that lead to decreased thermal stability and increased protease sensitivity lead to an immune response that was not protective, suggesting that linear epitopes are less important for protection than are conformational epitopes. This is also consistent with the cross-reactivity of borreliacidal DbpA antibodies among divergent B. burgdorferi sensu lato isolates that lack candidates for conserved linear DbpA epitopes.

Recombinant DbpA expressed in the cytosol of E. coli with either of two different polyhistidine tags instead of the signal peptide was less effective as a vaccine antigen. These proteins were different structurally and functionally from secreted and acylated DbpA, suggesting that the subcellular compartment in which the antigen folds also influenced its antigenicity. Although we have not excluded the formal possibility that both of these polyhistidine tags altered an important protective antigenic epitope near the amino terminus of recombinant DbpA, we consider this unlikely since that end of the natural protein is proximal to the tripalmitoyl cysteine membrane anchor, which would limit its accessibility to antibodies on intact spirochetes. The production of DbpA as a lipoprotein has multiple advantages from a vaccine efficacy standpoint. First and foremost is that the DbpA as expressed in Borrelia is a secreted lipoprotein, and the expression of the DbpA as a recombinant lipoprotein should best maintain its normal context and posttranslational processing. The folding of the DbpA in the periplasm is likely to be different from that in the cytoplasm, since the presence of chaperone proteins differs between these two compartments (31). Second, the lipid moiety serves as an integral adjuvant and as an activator of the immune response, which may improve the overall immunogenicity of the antigen. The effects of the cysteine-to-serine mutations on DbpA_N40 stability and antigenicity highlight the importance of the tertiary structure of the carboxy-terminal portion of the protein, presumably stabilized through a loop formed by a disulfide bond. In gram-negative bacteria and eukaryotes, disulfide bonds are catalyzed by extracytoplasmic oxidoreductases (37). Disulfide bonds in recombinant H₆DbpA_N40 and H₁₀DbpA_N40 may form spontaneously, albeit inefficiently, once they are released from the reducing environment of the E. coli cytosol, as has been shown to occur for the periplasmic enzyme alkaline phosphatase (11). Partial oxidation of cytosolic DbpA_N40 during purification, leading to the formation of the putative disulfide bond, would explain the intermediate vaccine efficacy of H₆DbpA_N40 and H₁₀DbpA_N40. Most DbpAs lack this carboxy-terminal cysteine pair and are presumably stabilized by other intrachain interactions.

Evidence for protective conformational epitopes has also been reported for other B. burgdorferi membrane proteins. Oms66 (or p66) is an outer membrane-spanning surface-exposed protein of Borrelia spp. that has recently been shown to be protective against infection by in vivo-adapted borreliae in a mouse model (14). This study showed that natural p66 antigen isolated from B. burgdorferi had membrane channel-forming (porin) activity and was protective against the homologous isolate, but recombinant p66 produced in the E. coli cytosol as a GST fusion lacked porin activity and was not protective (14). OspC lipoprotein has also been shown to be an in vivo-expressed protective antigen; however, this protein is a protective immunogen against homologous challenge but not against heterologous challenge (5, 35, 36). OspC immunogen was shown to be sensitive to SDS or thermal denaturation (20, 21). The denatured forms of the OspC protein were not protective and did not elicit borreliacidal antibodies.

B. burgdorferi is unusual in its abundance of membrane lipoproteins, and several of these have shown substantial effectiveness as vaccines. However, even more putative or biochemically confirmed lipoproteins have shown partial or no protection. Most of these partially effective or apparently ineffective recombinant vaccine antigens have been expressed in E. coli as fusions to the cytosolic protein GST, including OspE and the related p21, OspF and the related pG, BmpA(P39), p30, p37, p55, BBK50/P37, and lp6.6 (4, 9, 10, 15, 16, 19, 30, 32, 43). It is possible that this convenient expression and purification strategy may have, in some cases, altered the folding of these otherwise-secreted proteins and compromised the vaccine efficacy of the recombinant products, as was the case with Oms66/p66 (14). Several B. burgdorferi lipoproteins expressed as GST fusions, including OspA, OspB, OspC, and p35/BBK32, have shown vaccine efficacy, however (18, 19, 45).

We have previously reported that Freund's adjuvant formulations of cytosolically expressed recombinant DbpA protected mice from B. burgdorferi challenge (27), and this was confirmed by others (17, 25). In vitro growth inhibition assays with antiserum against a cytosolic polyhistidine-tagged form DbpA failed to yield reproducible results in the study by Hagman et al. (25). The results of our present study suggest that the poor biological activity of this antiserum may have been due to their use of a suboptimal DbpA immunogen. We have now shown, by direct comparison, that secreted forms of DbpA are more effective than nonsecreted DbpA when formulated with Freund's or Ribi adjuvants. Secreted lipoprotein DbpA was also effective when formulated with Alhydrogel (this study), a clinically relevant adjuvant, or injected without adjuvant (28). Hagman et al. recently reported that a Freund's adjuvant formulation of cytosolic polyhistidine-tagged DbpA failed to elicit sterilizing immunity in mice against B. burgdorferi challenge by multiple tick bite (26). This study also reported evidence that DbpA targets for protective immunity are expressed by tick-borne B. burgdorferi only in the mammalian stage, unlike OspA and OspC, which are expressed by spirochetes within ticks and are targets for transmission-blocking immunity (13, 22). Previously we have shown that immunization with secreted lipoprotein DbpA elicits antibodies that can kill mammalian-stage spirochetes that are resistant to OspA antibody (7). A more conformationally correct form of the immunogen may achieve effective DbpA immunity against infection or disease caused by tick transmitted B. burgdorferi. Experiments are in progress to address this possibility. Recent observations by Ohnishi et al. (32a) reveal a previously unappreciated heterogeneity in the antigenic profile of the population of tick-borne B. burgdorferi. DbpA was not examined in that study. Even if DbpA is not sufficient as a monovalent subunit vaccine to prevent tick-borne infection, it may serve as a component of an effective multisubunit vaccine combined with other in vivo-expressed antigens or in combination with OspA (28).