INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Lyme disease, or Lyme borreliosis, is an illness causing manifestations in humans including rash, fever, and malaise; if left untreated, the infection can cause arthritis and cardiac and neurological damage (22). The disease is caused by an infection with the bacterial pathogen Borrelia burgdorferi, which is transmitted to humans by way of the bite of infected ticks from the Ixodes ricinus complex (2). If the infection is treated early, antibiotics are effective in controlling it (23). Clinical trials of a prophylactic vaccine have recently been completed and have shown that the vaccine has promise in preventing cases of Lyme disease (20, 24). The vaccine is based upon immunization with the outer surface protein A (OspA) antigen. Its effectiveness requires the presence of neutralizing OspA antibodies in the host, which eradicate potential infecting borreliae within a feeding tick, thus preventing transmission of the organisms (3, 5).

Other B. burgdorferi proteins have been shown to elicit some protective immunity against borrelia infection in laboratory animals. Among these are OspB (4, 19), decorin binding protein A (9, 10), and OspC (17, 19). A previous study in this laboratory demonstrated that active immunization with a recombinant form of OspC protected mice against a challenge infection administered by tick bite (7). It was also observed in that study that other mice remained unprotected from the challenge infection even though they harbored OspC antibodies. The difference in this group, however, was that they had been immunized with OspC from B. burgdorferi B31 cells purified under denaturing conditions. This observation suggested that the OspC protective epitope was shaped by protein folding and secondary structure and was sensitive to denaturing conditions. This report describes results of tick bite challenges to groups of mice actively immunized with strain B31-derived recombinant OspC that had been treated by various denaturation procedures. In addition, a protective anti-OspC monoclonal antibody (MAb) was used to localize the regions of the molecule essential for the protective activity.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Borrelia strains and growth conditions. B. burgdorferi sensu stricto strain B31 (low passage number [<10 passages]) was originally provided by A. Barbour (University of California, Irvine) and maintained by the Molecular Bacteriology Section (Division of Vector-Borne Infectious Diseases [DVBID], Centers for Disease Control and Prevention, Fort Collins, Colo.). Borreliae were grown in Barbour-Stoenner-Kelley modified medium (Sigma Chemical Co., St. Louis, Mo.) supplemented with 6% rabbit serum (Pel-Freez, Rogers, Ark.) at 34°C until cell growth reached approximately 10⁷ to 10⁸ organisms/ml, after which the cell pellet was collected, washed, and frozen at 20°C until needed. Tick colonies of B31-infected Ixodes scapularis used for challenges were developed (16), maintained, and provided by J. Piesman (Centers for Disease Control and Prevention, Fort Collins, Colo.).

ospC gene cloning and expression. Construction of a B. burgdorferi genomic DNA library and isolation of the ospC gene have been described elsewhere (7). Following isolation of the ospC gene, it was subcloned from the LambdaZapII vector (Stratagene, La Jolla, Calif.) to the plasmid vector pBluescript II SK (Stratagene) by the in vivo excision method, according to the manufacturer's directions.

Western blot and spot blot analysis. E. coli clones containing the OspC antigen gene were grown in Luria-Bertani broth culture to late-log or stationary phase with the addition of 0.5 mM isopropyl--D-thiogalactopyranoside during mid-log phase. Cells were pelleted, and an aliquot was resuspended in 2× sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample loading buffer (containing 5% 2-mercaptoethanol and 4% SDS), boiled for 5 min, and loaded onto an 11.75% polyacrylamide gel for SDS-PAGE protein analysis. Fractionated proteins were blotted electrophoretically onto Immobilon-P (Millipore Corp., Bedford, Mass.) polyvinylidene difluoride membranes, and Western blot analyses were performed according to standard procedures (26).

Antigen preparations. Recombinant OspC antigen generated by pSCREEN (OspC-pSCREEN antigen) was prepared for mouse immunizations as follows. Following growth of the culture, the E. coli cells were harvested by centrifugation and frozen at 20°C. The cells were resuspended in 0.1 culture volume of 50 mM NaH₂PO₄ (pH 8.0)-300 mM NaCl-20 mM imidazole-10 µg of lysozyme/ml. The mixture was incubated on ice for 30 min; then the cell suspension was sonicated with a Branson (Danbury, Conn.) Sonifier 450 on ice to prevent heat denaturation. The suspension was centrifuged to pellet and separate the insoluble cellular debris from the soluble fraction. The insoluble fraction containing recombinant OspC inclusion bodies was washed 3 times in 100 mM Tris-HCl (pH 7.5)-2 M urea-2% Triton X-100. The washed pellet was solubilized in 6 M guanidine-HCl-100 mM NaH₂PO₄-10 mM Tris (pH 8.0). After an incubation at room temperature for 30 min, any insoluble material left was removed by centrifugation, and the guanidine-solubilized material was collected. An aliquot of the guanidine-solubilized antigen, which carried the six-histidine (His) tag, was passed through a Qiagen (Santa Clarita, Calif.) Ni-nitrilotriacetic acid Spin Kit according to the manufacturer's directions. The eluate contained purified recombinant OspC antigen with the guanidine-HCl buffer removed and replaced with 50 mM NaH₂PO₄ (pH 8.0)-300 mM NaCl-250 mM imidazole. Heat-denatured antigen was prepared by boiling an aliquot of the original insoluble pellet fraction for 10 min.

Mouse immunizations and tick challenge. Specific-pathogen-free outbred mice from a breeding colony of the Institute for Cancer Research (Philadelphia, Pa.) that were maintained at the DVBID were used for vaccinations. Antigen preparations were generated as described above. Approximately 50 to 100 µg of antigen was mixed with an equal volume of adjuvant (TiterMax, Norcross, Ga.), and about 50 to 100 µl was administered subcutaneously to each mouse. All mice were boosted at 2 and 4 weeks after the initial inoculation and were bled 1 week following each boost. A final boost, if needed, was administered 1 to 2 weeks prior to tick infestation. Mice were bled serially to track seroconversion of individual mice to the immunogen. Seroconversion was assessed by Western immunoblotting against B. burgdorferi B31 low-passage-number antigen with Marblot strips (MarDx Diagnostics, Inc, Carlsbad, Calif.). Once seroconversion was observed, the mice were challenged with B. burgdorferi B31 by infestation of nymphal ticks harboring the spirochete. Ten ticks were placed on each mouse and allowed to feed to repletion. Three to four days later, engorged ticks were collected and stored at 21°C under a 97% relative humidity atmosphere. An ear biopsy culture was performed on each mouse approximately 30 days after tick feeding to test for the presence of borreliae in the tissue as previously described (21). Cultures were observed microscopically for borrelia growth 1 to 2 weeks later. Cultures were deemed positive if borrelia cells were observed in any field and negative if no borrelia cells were observed in 20 fields.

Construction of truncated recombinant OspC for epitope localization studies. Subclones containing portions of the ospC gene were constructed as follows. A BglII restriction site occurs at nucleotide position 430 and cuts the gene into two fragments of 330 and 200 bp. The two fragments were fractionated on an agarose gel and purified from the agarose with a Qiagen Gel Extraction Kit. The fragments were ligated into pBluescript at the BamHI site, in frame with the vector's lacZ fusion product to ensure proper expression.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Synthesis of recombinant OspC. B. burgdorferi B31 recombinant OspC protein expression in this laboratory was originally generated from the pBluescript vector, as an extension from the in vivo subcloning feature of the lambda phage clone LambdaZapII. An earlier study in this laboratory used crude sonicated E. coli lysate expressing this form of recombinant OspC as the immunogen, which proved to be 100% protective against a homologous challenge (7). To improve the yield of expressed recombinant OspC, the same coding sequence was subcloned into the vector pSCREEN-1b, a pET vector derivative. Following induction of the gene in this vector, expression levels rose to account for approximately one-half of the total protein expressed by the E. coli host, thereby providing substantial material for further immunizations and subsequent studies. Unlike expression in pBluescript, where a substantial amount of the expressed OspC remained soluble (Fig. 1A), the recombinant OspC generated by the pSCREEN system accumulated almost exclusively as insoluble inclusion bodies, as seen following cell disruption and sonication (Fig. 1B). This material was used in the immunization of mouse groups following the denaturation treatments described in Materials and Methods.

View larger version (58K): [in this window] [in a new window]   FIG. 1.   (A) Western blot of E. coli lysates expressing recombinant OspC in the pBluescript vector. Lanes: 1, soluble sonicated fraction; 2, insoluble sonicated fraction. Approximately 10 times less protein was loaded in lane 1 than in lane 2, indicating the predominance of the recombinant OspC in the soluble fraction. Due to the small amount of OspC protein expressed by pBluescript, the antigen is observed better by immunoblotting than by Coomassie brilliant blue staining as in panel B. (B) Coomassie brilliant blue-stained SDS-PAGE gel of E. coli lysates expressing recombinant OspC in the (pET) pSCREEN-1b vector. Lanes: 1, soluble sonicated fraction; 2, insoluble sonicated fraction containing recombinant OspC inclusion bodies. Arrows indicate the recombinant OspC bands.

B. burgdorferi mouse challenges by tick bite. Following booster immunizations with the various immunogens, the mice were individually assayed for seroconversion to OspC by immunoblotting prior to tick challenge (Fig. 2A). Two and four weeks after tick feeding, protection from infectivity was assessed by serological profiles and culture of ear tissue, respectively. Mice immunized with OspC denatured either by heat or by guanidine treatment were not protected from infection (Fig. 2B; Table 1). Removal of guanidine by passage of OspC through a nickel cation column to purify the antigen by His tag affinity chromatography did not restore the OspC to its protective conformation, as evidenced by the lack of protection (Fig. 2B; Table 1).

View larger version (66K): [in this window] [in a new window]   FIG. 2.   Serological profiles of individual mice immunized with recombinant OspC expressed by the pSCREEN (pET) vector system as assayed by immunoblotting. (A) Mouse groups 1 to 5 prior to challenge; (B) mouse groups 1 to 5 at 2 weeks following challenge, with seroconversion, indicative of infection, beginning in most mice. Mice were immunized with E. coli lysate containing (pET) pSCREEN vector only with no insert (group 1), heat-denatured recombinant OspC (group 2), guanidine-HCl-denatured recombinant OspC (group 3), nondenatured, insoluble recombinant OspC (group 4), or recombinant OspC, denatured with guanidine, with removal of guanidine following antigen purification on a Ni2+ cation column. Sizes of antigen bands (in kilodaltons) are given on the right. (C) Serological profile of mouse group 4, at 4 weeks following challenge, with three of five mice seroconverting, indicative of infection. Sizes of antigen bands (in kilodaltons) are given on the right.

4

Localization of the OspC protective epitope. To identify the fragment of the recombinant OspC molecule containing the protective epitope, truncated forms were produced, expressed, and reacted by immunoblotting with an OspC MAb. This MAb was chosen to localize the protective epitope because of its demonstrated ability to passively immunize mice against a challenge infection (15). This murine anti-OspC antibody (termed B5) was one of a panel of MAbs that was generated by using B. burgdorferi-infected ticks to transmit antigen as the primary and booster inoculations (8, 14).

View larger version (17K): [in this window] [in a new window]   FIG. 3.   (A) Western blot demonstrating MAb B5 reactivity against B. burgdorferi B31 low-passage-number cells (B31 low) and recombinant (Rec.) OspC proteins generated in the expression vectors pBluescript and pSCREEN. The recombinant proteins are larger than the B31 OspC due to the presence of fusion partners. Molecular masses (in kilodaltons) are given on the left. (B) Spot blot demonstrating MAb B5 reactivity against non-heat-denatured ("not boiled") and heat-denatured ("boiled") antigens. B31 high, B. burgdorferi B31 high-passage-number (>50 passages) cells, which produce small amounts of OspC; E. coli pSCREEN only, lysate of cells harboring the vector with no cloned insert. Recombinant (Rec.) OspC E. coli lysates are the same as in the Western blot in panel A.

View larger version (18K): [in this window] [in a new window]   FIG. 4.   Diagram showing truncated constructs of the ospC gene coding sequence. The long solid arrow at the top represents the coding sequence minus the signal peptide, beginning at amino acid (aa) 33. The dotted line at the left represents the missing signal peptide. The BglII restriction site is indicated. Arrows indicate the relative positions of the primers used to amplify and clone the constructs shown below. Lines bounded by dots represent the regions encoded by the truncated constructs, whose designations appear at the right. 5' Bgl II and 3' Bgl II were clones made by digestion of the gene with BglII. HK-12 and BS-32 were made by unidirectional deletions. 117-590, F1-590, 117-R2, and F1-R1 were made by PCR with the primers shown. (+) and (), reactivity and no reactivity, respectively, with MAb B5.

View larger version (40K): [in this window] [in a new window]   FIG. 5.   Western blot analysis of OspC expression in E. coli from truncated constructs. (A) Blot probed with the anti-OspC MAb B5. (B) The same blot stripped of antibody from panel A and reprobed with polyclonal anti-recombinant OspC derived from mice immunized with guanidine-denatured recombinant OspC. Lanes: F1-R1, pBAD construct of entire OspC coding sequence minus the leader peptide; 1, construct 117-590; 2, construct F1-590; 3, construct 117-R2; 4, pBAD vector only; 5, B. burgdorferi lysate. Arrows indicate the positions of the recombinant (rec.) OspCs (lanes 1 to 3) and the B. burgdorferi (Bb) OspC (lane 5). The recombinants are slightly larger than the native OspC due to the presence of a fusion partner from the expression vector. Construct F1-R1 expresses a doublet recombinant protein.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Recombinant forms of OspC have been demonstrated by various groups to confer protective immunity against a B. burgdorferi infection (7, 17, 19). The recombinant OspC preparations used in protection studies were purified without steps involving denaturation of the antigen. A study in this laboratory observed, however, that OspC purified from B. burgdorferi B31 cells that involved denaturation procedures did not elicit protective immunity despite high anti-OspC titers in the host (7).

To further explore this observation, this study was designed to denature recombinant OspC proven to be protective and assess its property as a protective immunogen. In our previous experiments, OspC was expressed in a soluble form by using the vector pBluescript. Because the yield of recombinant protein was low in this system, a pET vector system was alternately used to increase quantity. The fact that only two of five mice were protected by the undenatured pET vector-generated OspC conflicted with the previous observation of 100% protection by the pBluescript-generated OspC (Table 1).

The pET vector (pSCREEN)-OspC construct differed from its pBluescript counterpart in that it possessed a 38-kDa amino-terminal fusion product. The large fusion partner helped in stabilizing the expressed product, thus aiding in the synthesis of higher recombinant-protein yields. Further analysis showed that the pET vector-generated OspC was produced almost entirely as insoluble inclusion bodies aggregated within the E. coli cytoplasm. Therefore, it is probable that such an insoluble form results in folding of the protein so that the protective epitope is either physically changed or masked compared to the soluble form of the molecule. The fact that two of the mice were protected probably indicates that a minute portion of the OspC preparation was present as a soluble form, and these mice elicited a protective response against that form of the antigen. Denatured preparations of the pET-derived OspC resulted in no mice (of 15) being protected. Although this result is probably due to the antigen's being denatured, comparisons to the control group receiving undenatured pET-derived OspC could not be unequivocal, because only 40% of the latter were protected, versus the 100% in our previous study using soluble antigen.

Therefore, another challenge was performed to assess the effects of the denaturation on the protective pBluescript-OspC antigen preparation. This expression vector produces recombinant protein in a soluble form, albeit not in the quantities seen with the pSCREEN system. However, even with the crude E. coli lysate form, 100% protection had been observed (in 12 mice tested). Five mice were immunized with the undenatured preparation, and five mice were immunized with the same preparation that was heat denatured. Four of the five mice immunized with the heat-denatured antigen were not protected, whereas three of four mice immunized with the undenatured antigen were protected (one mouse died during the trial period). Thus, the total number of mice immunized with the pBluescript-OspC in two studies from this laboratory was 16, with 15 demonstrating protection (94%) (Table 1). When the antigen was boiled, the protection rate dropped to one mouse of five (20%). The results of these experiments demonstrate that the protective epitope on the OspC molecule of B. burgdorferi B31 is conformational in nature and is affected by physical properties such as solubility, heat, and chemical denaturation.

Localization of the protective epitope on the recombinant OspC was attempted by using an OspC-specific murine MAb generated in this laboratory. The antibody was made by a method in which the mouse was immunized with tick-transmitted B. burgdorferi B31 instead of culture-grown organisms (8, 14). It was determined that this MAb, B5, when passively administered to mice, provided protection against tick-transmitted B. burgdorferi B31 infection (15). Thus, this antibody was a powerful tool in which to map or localize the OspC protective epitope.

MAb B5 was reactive in Western blots against recombinant OspC generated from all of the vector expression systems used in this study (Fig. 3A and 5A), as well as against B. burgdorferi OspC (Fig. 3A). If MAb B5 recognized a conformational epitope, then one would not expect these OspC proteins to be immunoreactive on blots with this antibody, since the preparations had been denatured by treatment with SDS, 2-mercaptoethanol, and boiling in the Laemmli SDS-PAGE system. This discrepancy may be explained by antigen renaturation following protein transfer to the blotting membrane. SDS is not present in the transfer buffer and may be washed from the gel and protein during blotting by methanol present in the transfer buffer. Partial or substantial removal of SDS may then allow refolding of the OspC to occur upon transfer to the membrane. A similar situation has been observed with disulfide bridge-associated conformational epitopes in the tick-borne encephalitis virus E glycoprotein (28). A simple experiment to support this notion was performed whereby undenatured and boiled lysates, instead of being subjected to Western blotting, were "spot" blotted onto nitrocellulose and reacted with MAb B5. Figure 3B demonstrates clearly that OspC immunoreactivity to MAb B5 was lost when the antigen was heat denatured.

MAb B5 reactivity against the recombinant OspC was abolished when only 7 amino acids were deleted from the NH₂ terminus, or 13 amino acids were deleted from the COOH terminus, of the molecule. Therefore, either the ends of the molecule harbor the epitope, or more likely, the entire molecule is required for proper folding to form the epitope. There is only one cysteine in the recombinant OspC (there is a second cysteine in the OspC coding sequence, but it is located in the signal peptide, which is not present in the recombinant protein), so cysteine-cysteine disulfide linkages are not possible. A recent study by Mathiesen et al. has shown that a dominant epitope of the Borrelia garinii OspC consists of the 10 C-terminal amino acids (13). The epitope was mapped by using sera from patients with neuroborreliosis from a B. garinii infection, but it was not within the scope of that study to determine whether the antisera were protective in passive immunization experiments. Nevertheless, the C-terminal location of that epitope is somewhat consistent with that seen in this study, whereby the deletion of the B. burgdorferi OspC C terminus abolished MAb B5 reactivity. However, when the C terminus was left intact and the N terminus was deleted (construct 117-R2) (Fig. 3), antibody reactivity was also lost. Therefore, the conformationally dependent protective epitope is probably different from that seen with B. garinii-infected individuals.

It remains to be determined if OspC molecules from other strains of B. burgdorferi have conformational protective epitope properties similar to those observed for strain B31. Sequence analysis of OspCs from various strains shows regions of heterogeneity that may reflect differences in structural properties and antigenicity (11, 12, 25, 27). OspC cross-protective immunization studies are not extensive, but one study has shown that OspC antigen from strain SON188 does not cross-protect against a needle challenge with in vitro-cultured strain CA4 or 297 (18). Additionally, in a study by Bockenstedt et al., immunization with recombinant OspC derived from strain N40 (and also noted with strain 297) failed to protect mice against a tick-borne challenge infection, but recombinant OspC immunogen from strain PKo was protective (1). Those results of strain-specific immunity were attributed to differences in OspC surface expression on the borreliae, although OspC antigens for both strains were demonstrated to be localized beneath the outer membrane. Therefore, the protective capabilities of OspC antigens from different strains appear to be unique, perhaps dependent on their physical properties.

Active or passive OspC vaccines may present an alternative to the OspA-based vaccine which has recently received federal Food and Drug Administration approval. Passive immunization with polyclonal antiserum generated from recombinant OspC from B. burgdorferi ZS7 has demonstrated therapeutic capabilities against chronic infections in mice (29). These data, plus the fact that infected hosts and Lyme disease patients produce an immunodominant humoral response to OspC (6), are evidence that borreliae constitutively express OspC during the course of infection. Therefore, in studies where OspC is used either in an active immunization or to generate protective antibody for passive immunization, it becomes critical that recombinant OspC be expressed or produced in such a manner as to preserve the conformation of the protective epitope that may be particular for that strain.