Borrelia burgdorferi Escape Mutants That Survive in the Presence of Antiserum to the OspA Vaccine Are Killed When Complement Is Also Present

Transcription

1 INFECTION AND IMMUNITY, June 1998, p Vol. 66, No /98/$ Copyright 1998, American Society for Microbiology Borrelia burgdorferi Escape Mutants That Survive in the Presence of Antiserum to the OspA Vaccine Are Killed When Complement Is Also Present MÓNICA SOLÉ, 1 CARLOS BANTAR, 1 KARL INDEST, 1 YAN GU, 1 RAMESH RAMAMOORTHY, 1 RICHARD COUGHLIN, 2 AND MARIO T. PHILIPP 1 * Department of Parasitology, Tulane Regional Primate Research Center, Tulane University Medical Center, Covington, Louisiana 70433, 1 and Aquila Biopharmaceuticals, Inc., Worcester, Massachusetts Received 26 September 1997/Returned for modification 30 December 1997/Accepted 24 March 1998 As an initial attempt to investigate the possible role of outer surface protein A (OspA) escape mutants of Borrelia burgdorferi in decreasing the efficacy of the OspA vaccine, mutants of the HB19 strain of B. burgdorferi sensu stricto were selected in vitro from an uncloned, low-passage-number isolate. The antiserum used for selection was obtained from rhesus monkeys that had been given a vaccine of the same formulation and dose, and by the same route of administration, as that given to humans in several trials. All of the mutants selected in liquid medium and subsequently cloned twice in solid medium expressed a single abundant protein of 28 to 34 kda instead of both OspA and OspB. Depending on the mutant, this protein reacted strongly, weakly, or not detectably with the anti-ospa antibody used for selection. Analysis of the ospab locus of each of four representatives from these three groups of mutants by PCR with oligonucleotide primers that hybridize to flanking regions of the ospab operon, and of the corresponding phenotype with monoclonal antibodies that bind to the amino or carboxyl terminus of the OspA or OspB polypeptide, indicated that in all cases a deletion within the operon had occurred. Spirochetes from the four mutant strains chosen for further analysis could be killed in antibody-dependent, complement-mediated killing assays with the selecting anti-ospa antibody, despite their resistance to killing with this antibody in the absence of complement. Complement-mediated killing occurred at an antibody concentration higher than that required to kill wild-type spirochetes. If anti-ospa antibody acts only within the tick, where complement is probably ineffective due to tick-derived decomplementing factors, then OspA escape mutants, if infectious, could seriously diminish the efficacy of OspA vaccines. On the other hand, if the killing of B. burgdorferi with anti-ospa antibody also takes place within the human host, then our results indicate that chimeric/deletion escape mutants will be killed as well. * Corresponding author. Mailing address: TRPRC, Three Rivers Rd., Covington, LA Phone: (504) , ext Fax: (504) philipp@tpc.tulane.edu. Present address: Departamento de Biología Molecular, Facultad de Ciencias, Universidad Autónoma de Madrid, Madrid, Spain. Present address: CEMIC, Laboratorio de Microbiología, Buenos Aires, Argentina. A vaccine to prevent infection with Borrelia burgdorferi, the spirochete that causes Lyme disease, has recently undergone phase III trials in humans (36). The vaccine s key component is recombinant lipidated outer surface protein A (OspA) of B. burgdorferi. Proof that OspA is a protective antigen has been achieved in numerous animal experiments using diverse antigen-adjuvant combinations (7, 21; reviewed in reference 23), and vaccine formulations that are compatible with human use also have been shown to be efficacious in mice (13, 14, 34). Moreover, the vaccine appears to be safe, as indicated by safety evaluations performed in monkeys (24), dogs (21), and humans (16, 30, 35). The mechanism whereby B. burgdorferi spirochetes succumb to the immunity elicited by OspA has been under intense scrutiny. While it is clear that antibody is the key mediator of spirochetal death (reviewed in reference 2), there is still some uncertainty as to where and when during the infection s natural history this antibody-mediated killing may occur. A two-tiered mode of protection by anti-ospa antibody, one acting on the spirochetes within the tick prior to transmission and the second acting on delivery within the host skin, had initially been postulated (11). Evidence gathered subsequently indicated that the expression of OspA is interrupted or down-regulated when spirochetes reach the tick s salivary gland, en route to the vertebrate host (8). A corollary of this finding is that anti-ospa antibody can kill only within the tick, while OspA is still expressed, but not thereafter (8). Indeed, OspA-vaccinated mice that are able to fend off spirochetes delivered by injection are not protected against a challenge infection when the latter is conveyed by implantation of skin from B. burgdorferi-infected syngeneic donor mice (3). On the other hand, in humans infected with B. burgdorferi (by tick bite), antibody to OspA has been detected both during the early (15, 31) and chronic (15) phases of Lyme disease. These observations imply that the expression of OspA is not completely arrested in humans and thus spirochetes would be killed by anti-ospa antibody not only within the tick but also in human tissues or fluids. As we demonstrate in this report, an effort to ascertain the sites of spirochetal demise by anti-ospa antibody is pertinent to analyses of whether OspA escape mutants can affect the efficacy of OspA vaccines. Escape mutants, i.e., mutants of B. burgdorferi that survive and multiply in the presence of an antibody that kills the wild-type cells, have been obtained by cultivating spirochetes in the presence of both monoclonal and polyclonal antibodies (4, 5, 28). Typical OspA escape mutant phenotypes include those that express neither OspA nor OspB and, frequently, expressors of a chimeric molecule composed of an amino (N)-terminal portion of OspA and a carboxyl (C)-terminal portion of OspB. These deletion mutants have been found in multiple strains of B. burgdorferi (26) and in 2540

2 VOL. 66, 1998 OspA ESCAPE MUTANTS KILLED BY ANTIBODY AND COMPLEMENT 2541 several tick isolates from California (32). We investigated the type of escape mutants that could be obtained from a nonclonal population of B. burgdorferi in the presence of antibody elicited in rhesus monkeys by an OspA vaccine formulation identical to that being used in human trials (24, 25, 35). The OspA and OspB phenotypes of clonal populations of these mutants were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting with monoclonal antibodies (MAbs) that bind to either the N- or C-terminal portion of OspA or OspB. The size of the ospab locus of the mutants was determined by PCR with primers flanking the operon, and the operon was sequenced when this information was deemed relevant. Finally, the ability of each of these mutants to escape the combined effect of antibody and complement was quantified to ascertain if survival of these mutants within the vertebrate host was possible. To our surprise, mutants that were able to resist anti-ospa antibody-dependent killing were nonetheless killed in the presence of this antibody and complement. The antibody concentration at which 50% of spirochetal killing occurred depended on the mutant s phenotype. If anti-ospa antibody acts only within the tick, where complement is probably ineffective due to tick-derived decomplementing factors (22), then OspA escape mutants, if infectious, could seriously diminish the efficacy of OspA vaccines. On the other hand, if the killing of B. burgdorferi with anti-ospa antibody also takes place within the human host, then our results indicate that chimeric/deletion escape mutants will be killed as well. The data also permitted some insights into the hitherto unknown mechanism of antibody-dependent killing of spirochetes in the absence of complement. MATERIALS AND METHODS B. burgdorferi strains, in vitro culture, and spirochete cloning procedures. B. burgdorferi HB19, low passage number (passage 6), was obtained from the Centers for Disease Control and Prevention, Fort Collins, Colo. (CDC accession no. CT87005). The strain was chosen for the study because it belonged to the B. burgdorferi sensu stricto species, like the strain whose (recombinant) OspA was used to vaccinate the animals, and because in our hands it had a relatively high plating efficiency ( 50%). The organisms were cultured in BSK-H medium (Sigma Chemical Co., St. Louis, Mo.) supplemented with 10% heat-inactivated (56 C for 30 min) rabbit serum (Sigma) and antimicrobial agents (rifampin, 50 g/ml; phosphomycin, 200 g/ml; amphotericin B, 8 g/ml) at 34 C in a humidified atmosphere of 5% CO 2 3% O 2 92% N 2. Spirochetes were quantified under a dark-field microscope by counting spirochetes per calibrated field until a total count of 100 to 200 organisms was collected. The mean number of spirochetes per field was then calculated, and this value was expressed as spirochetes per milliliter of culture medium. Spirochetes were cloned by cultivation on solid medium as follows (28). Cells were added to 50% BSK-H medium containing 1% low-melting-temperature agarose (SeaPlaque; FMC Corporation, Rockland, Maine) that was kept liquid at 37 C. The cell suspension was then poured as a thin layer over solid 80% BSK-H medium containing 1% SeaKem LE agarose (FMC) in small petri dishes (6 cm in diameter; Falcon; Becton Dickinson, Franklin Lakes, N.J.). The plates were incubated at 34 C in a candle jar for 7 to 10 days. Clones were expanded by inoculating a small piece of agarose with a single colony into 1 ml of BSK-H medium. This culture was allowed to reach log phase, at which point cells were cryopreserved at 70 C until further use. Antibodies. The antiserum pool used for selection of escape mutants was obtained from four rhesus macaques that were immunized with the lipidated form of recombinant OspA as described previously (24, 25). The recombinant lipidated OspA used as immunogen had been derived from the B. burgdorferi sensu stricto strain ZS7. The amino acid sequence of OspA from ZS7 (GenBank/ EMBL accession no. X16467) is identical to that of OspA from HB19 (as determined by us [GenBank accession no. AF026059]), save for the substitution of a glutamic acid residue in position 2 of the HB19 OspA leader peptide by a lysine residue in the ZS7 OspA. Hence, the two mature proteins are identical. Complement in the antiserum pool used for escape mutant selection was heat inactivated at 56 C for 30 min. The phenotype of the escape mutants was characterized on Western blots with the aid of two MAbs that bind to the N terminus of OspB (82C and 84B) (6, 33), two MAbs that bind to the C terminus of OspB (84C and 3A5) (6, 33), one MAb that binds to the N terminus of OspA (T2H12.4) (20), and two MAbs that bind to the C terminus of OspA (L3B5 and H3G4) (20). The anti-ospb MAbs were a gift from Denee Thomas, Department of Periodontics, University of Texas Health Science Center, San Antonio. Source of complement. Normal rhesus monkey serum was used as a source of complement. It was collected and stored as described previously (1). A single pool of serum was used throughout. Selection of escape mutants. Escape mutants were selected either in the presence of a high concentration of selecting anti-ospa antibody that remained unchanged throughout the selection process (protocol I) or in the presence of increasing concentrations of the selecting antibody, starting at a relatively low concentration (protocol II). In protocol I, cultures were started at cells in a 1-ml volume of BSK-H medium with a 1/10 dilution of anti-ospa antiserum. Cells were counted every day for 3 days, and cell viability (as judged by motility) was monitored. Every 3 or 4 days, 100 l of cell suspension was transferred to a fresh tube that contained antibody (1:10) and medium (0.9 ml). The process was repeated until cells were found to be 100% motile 3 days after a passage. After three additional passages in the presence of antibody, a fraction of the cells in each tube was plated on solid medium and cloned twice, as described above, and the rest was cryopreserved for later use. In protocol II, cultures were started as in protocol I but in the presence of a 1:160 dilution of the same anti-ospa antiserum pool. Cells were passaged as indicated for protocol I, but the antibody concentration was increased twofold at each passage. After reaching a serum dilution of 1:10, cells were passaged three more times without a change of the antibody concentration, and survivors were cloned and cryopreserved as in protocol I. The dilution of antiserum used for selection (1/10) was twice the concentration that gave 94% of killing in an antibody-dependent, complementmediated killing (ADCK) assay (see below). Selection of mutants also was attempted in the presence of complement, by adding a volume of the normal monkey serum pool used as a source of complement equal to 25% of the final volume but starting at higher antiserum dilution (1:5,120). Mutants that were obtained with protocol I were designated I 1,2,...,n and those obtained with protocol (II) as II 1,2,...,n. PAGE and Western blotting. SDS-PAGE and Western blot analysis were performed as described previously (1). Sizing of the ospab operon by PCR. About 10 7 spirochetes of the appropriate clone were washed once with phosphate-buffered saline by centrifugation (10,000 g for 10 min), resuspended in 20 l of TE buffer (10 mm Tris-HCl, 1 mm EDTA [ph 8.0]), boiled for 10 min, and centrifuged as described above. A 10- l volume of the supernatant was used as source of template DNA. The PCR was performed in a 100- l reaction mixture in which the final concentration of deoxynucleoside triphosphates was 0.2 mm each, primers were used at 100 nm each, the Taq DNA polymerase (Promega Biotech, Madison, Wis.) concentration was 50 U/ml, and the buffer was supplied by the enzyme s manufacturer. Primer sequences were 5 ATTAAGTTATATTAATATAAAAGGAGAATAT ATTATG 3 (primer BC238), which hybridizes 34 bases upstream from the beginning of the OspA coding sequence, and 5 CTTGTAGGGGGGTTTACT TATATATTATTT 3 (primer BC214), which hybridizes 27 bases downstream from the end of the OspB coding region. Samples were denatured at 94 C, reannealed at 45 C, and extended at 72 C, all for 1-min periods. A total of 39 such cycles was run, plus one additional cycle in which the extension time was 10 min. A positive control tube that contained purified DNA from B. burgdorferi HB19 wild type and a negative control that contained all of the reaction components except template were included in all runs. Amplicons were sized by electrophoresis on 1% agarose gels, using a 1-kb DNA ladder as a standard (Life Technologies, GIBCO-BRL, Grand Island, N.Y.). ADCK of B. burgdorferi. B. burgdorferi HB19 mutants and wild type were cultured until they reached mid- to late-log phase ( to spirochetes per ml). The assay was carried out in 96-well tissue culture plates (Costar, Cambridge, Mass.). A total of spirochetes in 25 l of BSK-H were added to wells that contained 50 l of BSK-H medium with doubling serial dilutions of the monkey anti-ospa antiserum pool used for the mutant selection, at dilutions of between 1:10 and 1:10,240. The plates were incubated at 34 C for 20 min before the addition of 25 l of normal rhesus monkey serum (source of complement). After 24 h of incubation at 34 C, the total numbers of killed and live spirochetes were counted as described above. Spirochetes were considered dead when they were either nonmotile or were barely motile and with a clearly disrupted shape. In selected cases, surviving spirochetes were quantified by the ability to form colonies in solid medium as described previously (1). In all cases, the number of colonies obtained matched the number of living spirochetes that had been plated, after correction for the plating efficiency of our HB19 isolate, thus confirming our criterion for spirochete viability. The rate of killing was calculated by counting at least 50 spirochetes per slide (well) in 5 to 10 fields. The standard deviation was that of the mean of duplicate determinations. DNA sequencing. The ospab operon from HB19 wild type (uncloned) or I 1 mutant spirochetes was amplified by PCR with primers BH238 (5 GACGGGA TCCATTAATATAAAAGGAGAATATATTATG 3 ) and XB214 (5 TGCTC TAGAGGGGTTTACTTATATATTATTT 3 ). These primers hybridize to the same sites as primers BC238 and BC214 but contain in addition a BamHI site (BH238) and an XbaI site (XB214). The ospab sequence of both HB19 wild type and mutant I 1 spirochetes contains a single BamHI site that is located 460 bp downstream from the start of ospa. This BamHI site was useful in subcloning portions of the ospab operon. Two BamHI fragments, one from HB19 wild-type spirochetes and the other from I 1 mutants, which encompassed the N-terminal

3 2542 SOLÉ ET AL. INFECT. IMMUN. FIG. 1. (A) SDS-PAGE analysis of whole extracts from cloned HB19 mutants obtained with selection protocols I and II. Extracts were electrophoresed on a 15% acrylamide gel that was stained with Coomassie brilliant blue. Positions of the wild-type (wt) OspA and OspB bands are indicated. Reactivity of each of the intense bands that represent either OspA or OspAB chimeric molecules with the anti-ospa polyclonal antibody used for selection was assessed on a Western blot of a similar gel and is depicted at the bottom as strong ( ), weak ( ), or absent ( ). (B) Analysis of the size of the ospab locus of HB19 wild type and of mutant strains I 5,I 1,II 1, and II 2 by PCR. Amplicons were obtained as described in Materials and Methods and electrophoresed on a 1% agarose gel. A 1-kb ladder was used as size standard (st). The size of three of the components of the ladder are denoted. (C) Western blot analysis of the reactivity of whole-antigen extracts from HB19 wild type and of mutant strains I 5 and I 1 with anti-ospb MAbs that react either with the N terminus of OspB (lanes 1 and 2; MAbs 82C and 84B, respectively) or with the C terminus of OspB (lanes 3 and 4; MAbs 84C and 3A5, respectively). Samples were electrophoresed on 15% acrylamide gels prior to blotting. Sizes of prestained molecular weight markers are indicated on the left. region of OspA, were subcloned into pbluescript. A 1,042-bp BamHI-XbaI fragment from the ospab operon of mutant I 1 that contained the C-terminal portion of OspA and a truncated OspB was cloned into puc18. The ospab operon from HB19 was also amplified by PCR with primers BC238 and BC214 and cloned into the AT cloning vector pgemt-easy (Promega). This construct was named pki-11. The C-terminus-encoding portion of the ospa gene from HB19 was sequenced from pki-11, using primers K104 (5 TAACAAGAGCA GACGGAAAA 3 ) and K105 (5 GTCAACGCTAAAGCAAATCC 3 ). Plasmid DNA for sequencing was prepared with a Qiagen Plasmid Midi kit as instructed by the manufacturer (Qiagen Inc., Chatsworth, Calif.). Nucleotide sequences were determined by sequencing both strands of all plasmid constructs by the dideoxy-chain termination method (29), using a modified T7 DNA polymerase (Sequenase version 2.0; U.S. Biochemical Corp., Cleveland, Ohio) and [ 35 S]dATP (NEN DuPont, Boston, Mass.). Nucleotide sequence accession number. The complete ospa coding region has been assigned GenBank accession no. AF RESULTS Selection of escape mutants. Two populations of uncloned escape mutants of B. burgdorferi HB19 were obtained by using the two protocols described in Materials and Methods, population I with protocol I and population II with protocol II. After the first round of cloning in solid medium, we arbitrarily selected 15 colonies from population I (I 1 to I 15 ) and 5 from population II (II 1 to II 5 ). Each of these mutant colonies was subcultured in liquid medium and cloned once again in solid medium. One of the 15 colonies from population I grew poorly in liquid medium and was not analyzed further. One colony was selected from each of the 19 plates. Colonies were expanded in liquid medium, aliquoted, and frozen for further characterization. Structural characterization of the mutant ospab locus and the corresponding proteins. Whole-cell extracts from each of the 19 colonies were analyzed by SDS-PAGE. All of the mutants, regardless of whether they had been isolated from population I or II, exhibited a single intense band in the 28- to 34-kDa range, of an apparent molecular mass that was either closely similar (mutants I 1 to I 4,II 2, and II 4 ) or smaller (I 5 to I 14,II 1,II 3, and II 5 ) than that of wild-type OspA (Fig. 1A). This result suggested that the mutants that had been selected could be deletion mutants of the type described previously by Rosa et al. (26). This type of mutant expresses a chimeric protein composed of an amino-terminal section of the OspA protein and a carboxyl-terminal section of the OspB protein. Both sections may vary in length. A Western blot of a similar gel developed with an aliquot of the same anti-ospa antiserum that was used to select the escape mutants showed that while the single intense band of mutants I 1 to I 4 reacted strongly with the antibody, the band of the remainder of the colonies reacted either weakly (II 1,II 3, and II 5 ) or not detectably (I 5 to I 14,II 2, and II 4 ) with this antiserum (Fig. 1A). We chose for further characterization one clone from the group that reacted strongly with the anti-ospa antibody (clone I 1 ), one from the group of weak reactors (II 1 ), and two from the group of nonreactors (I 5 and II 2 ). The selected clones were further characterized by sizing the mutant ospab operon by PCR. As estimated from the mobility of the molecular weight standards, mutant I 5 yielded an amplicon of 960 bp, mutant I 1 yielded one of 1,600 bp (in reality 1,550 bp, as indicated by its nucleotide sequence [see below]), and mutants II 1 and II 2 both yielded amplicons of 980 bp, all of which were smaller than the 1,780-bp-long wild-type amplicon (Fig. 1B). The result confirmed that the mutants had arisen through a deletion within the ospab operon. To further characterize the putatively chimeric proteins, Western blots of SDS extracts of mutants I 1,I 5,II 1, and II 2, were reacted with MAbs that bind to either the N or C terminus of OspB and MAbs that bind to either the N or C terminus of OspA (Table 1). Interestingly, while the protein expressed by mutants I 5,II 1, and II 2 reacted only with MAbs that bind to the C terminus of OspB (Table 1), that of mutant I 1 reacted only with MAbs that bind to the N terminus of this molecule (Table 1; Fig. 1C). These MAbs bound to a smaller (16.5-kDa) version of OspB (Fig. 1C). On the other hand, MAbs that bind to the C terminus of OspA bound to the protein expressed by mutant I 1 but not to that of mutant I 5, while a MAb that binds to the N-terminal portion of OspA bound to the proteins expressed by both mutants I 1 and I 5 (Table 1). It appeared,

4 VOL. 66, 1998 OspA ESCAPE MUTANTS KILLED BY ANTIBODY AND COMPLEMENT 2543 TABLE 1. Binding of MAbs that recognize the N- or C-terminal regions of OspA or OspB to Western blots of whole-cell extracts from HB19 mutant spirochetes a MAb Binding a WT OspA WT OspB I 1 I 5 II 1 II 2 82C (N) 84B (N) 84C (C) 3A5 (C) L3B5 (C) ND ND H3G4 (C) ND ND T2H12.4 (N) ND ND a WT, wild type; (C), carboxyl-terminal epitope; (N), amino-terminal epitope; ND, not determined. therefore, that mutant I 1 was able to express from the ospab operon not only a molecule that was indistinguishable in size from wild-type OspA and which reacted strongly with the anti- OspA polyclonal antibody used for mutant selection but also a smaller version of OspB. The other mutants, in contrast, appeared to express a single polypeptide which contained the C terminus of OspB and probably various lengths of the N terminus of OspA, as diagrammatically represented in Fig. 3. Functional antigenic characterization of escape mutants. Escape mutants were further characterized functionally by assessing if they were susceptible to killing in the presence of antibody plus complement. Several attempts had been made to select OspA escape mutants in the presence of complement (see Materials and Methods). No surviving spirochetes were obtained, and thus it seemed as though mutants that could be selected with antibody alone were sensitive to the combined action of antibody and complement. This turned out to be so. TABLE 2. ADCK 50 s of HB19 wild-type and mutant spirochetes Isolate comparison ADCK 50 WT a vs I 5...1/2,100 vs 1/120 WT vs I 1...1/2,000 vs 1/6,000 WT vs II 1...1/3,500 vs 1/38 WT vs II 2...1/3,000 vs 132 a WT, wild type. The four mutant spirochetes under study, I 1,I 5,II 1, and II 2, were subjected to ADCK titration using a serial dilution of the polyclonal anti-ospa antiserum pool used for the mutant selection plus normal rhesus monkey serum as a source of complement. The assay for each of the mutants was performed simultaneously with a killing assay of HB19 wild-type spirochetes in the same 96-well plate. All of the mutants were susceptible to ADCK but to different extents. Mutant I 1 was as susceptible to ADCK as the wild type (in fact, more so), with an ADCK 50 (the antibody dilution at which 50% of the spirochetes are killed in 24 h) of 1:6,000, compared to 1:2,000 for the wild-type HB19. The titration curves are shown in Fig. 2A. In contrast, mutant II 1 (as well as mutants I 5 and II 2 [not shown]) was markedly less susceptible to ADCK, with an ADCK 50 of 1:38, compared to 1:3,500 for the wild-type strain measured at the same time (Fig. 2B). The ADCK 50 values for all of the mutants and the corresponding wild-type values are shown in Table 2. Nucleotide sequence of the ospab operon of mutant I 1. The finding of comparable ADCK susceptibilities of mutant I 1 and wild-type HB19 was intriguing. On the one hand, I 1 was a mutant that could grow unimpeded at a 1:10 dilution of the anti-ospa antiserum used for selection; on the other hand, in the presence of complement it appeared as susceptible to kill- FIG. 2. ADCK assay of HB19 mutant I 1 (A) and II 1 (B) spirochetes compared to wild-type (WT) organisms. Killing assays for wild-type and mutant spirochetes were run simultaneously on the same microtiter plate. Error bars represent the standard deviation of the mean of two determinations.

5 2544 SOLÉ ET AL. INFECT. IMMUN. FIG. 3. Diagram of the ospab operon of HB19 wild type and of mutants I 1,I 5,II 1, and II 2. The filled portion of the ospa coding region represents the 69 bp missing from the ospa coding region that is available in GenBank (accession no. X16467); the complete version was filed under accession no. AF Nucleotides singled out represent substitutions encountered in the ospa nucleotide sequence of the HB19 isolate used in this study compared with that of the isolate under accession no. X The C in position 117 replaces an A, the A in position 446 replaces a G, and the C in position 465 replaces a T. The ospab operon of mutant I 1 spirochetes has a 220-nucleotide deletion flanked by nucleotides in positions 1256 and The ospab loci of mutants I 5,II 1, and II 2 have a deletion that encompasses sections from the 3 end of the ospa coding region and the 5 end of the coding region of ospb. The length of the deleted portion in each mutant is not known. ing by this antibody as the wild-type strain. The structural analysis indicated that this mutant expressed an OspA molecule that appeared to be indistinguishable in size from the wild-type OspA and which was able to bind to both N- and C-terminus-binding anti-ospa MAbs. An OspB protein significantly smaller than wild-type OspB was also expressed. The OspA and OspB expression pattern of I 1 was therefore different from that of other mutants; the latter likely expressed an OspAB chimera, as mentioned above. To sort out whether the ability of mutant I 1 to escape killing with the anti-ospa antibody in the absence of complement was due to minor changes in the ospa sequence possibly leading to an antigenically relevant conformational change in OspA or just to the deletion in the carboxyl terminus of OspB, the ospab operon of this mutant was sequenced. The ospab operon of the wild-type HB19 isolate from which the escape mutants had been selected was also sequenced, since the sequence available from Gen- Bank (accession no. L23136) was incomplete. The nucleotide sequence of the I 1 mutant ospa turned out to be identical to that of the wild-type ospa (Fig. 3). However, the wild-type HB19 ospab sequenced by us differed from the sequence published in GenBank in that the latter had an A instead of a C in position 117, a G instead of an A in position 446, and a T instead of acinposition 465 (Fig. 3). The C in position 465 (underlined) is part of the BamHI site (GGATCC) used by us for subcloning purposes, both for the wild-type operon and for the I 1 mutant operon (see Materials and Methods). Thus, this restriction site is not present in the HB19 ospa gene whose sequence is reported in GenBank. The 220 nucleotides comprised between bp 1256 and 1477 of the ospab operon were deleted in the I 1 mutant, thus resulting in the expression of a truncated OspB protein (Fig. 3). DISCUSSION As an initial attempt to investigate the possible role that OspA escape mutants of B. burgdorferi could play in decreasing the efficacy of the OspA vaccine, mutants of the HB19 strain of B. burgdorferi sensu stricto were selected in vitro from an uncloned low-passage-number isolate. The antiserum used for selection was obtained from rhesus monkeys (24, 25) that had been given a vaccine of the same formulation and dose, and by the same route of administration, as that given to humans in a phase II trial (35), namely, three 10- g doses of lipidated OspA with Al(OH) 3, administered intramuscularly at 4-week intervals (24, 25, 35). In previous experiments that were designed to select escape mutants in vitro, the evidence gathered indicated that these mutants were generated in a manner uninfluenced by the presence of the selecting antibody. There appeared to be no adaptational process at play (28). The selecting antibodies that were used in these experiments were either MAbs directed to various epitopes of OspA or OspB or a polyclonal antibody obtained from rats immunized with whole (HB19) spirochetes and Freund s complete adjuvant. The result of our selection procedure also indicates that mutants did not appear as a consequence of an adaptational process, for no obvious differences were observed between phenotypes of the clonal populations of mutants obtained by the selection procedure in which a gradually increasing concentration of antibody was used (protocol II) and that in which a single (high) concentration of antibody was chosen for selection (protocol I). Admittedly, however, proof of such differences lies in numbers, and the number of clones that were derived from each selection protocol is small. It is interesting that in contrast with the result obtained by Sadziene et al. (28), who were able to select mutants lacking the 49-kb plasmid that encodes the ospab operon by using either the rat polyspecific polyclonal antiserum mentioned above or anti-ospa or OspB MAbs, we did not observe such mutants. The mutants which we analyzed were deletion mutants that lacked portions of the ospa and/or ospb genes (see below). It is possible that our B. burgdorferi growth conditions and/or selection procedure did not favor the proliferation of mutants lacking the 49-kb plasmid. One difference between our selection protocol and that of Sadziene et al. is that we passaged the selected population three additional times after the cultures had reached a motility comparable to that of wild-type spirochetes grown in the absence of selecting antibody. Sadziene et al. did not perform these additional passages. Hence, we may have further selected against slow-

6 VOL. 66, 1998 OspA ESCAPE MUTANTS KILLED BY ANTIBODY AND COMPLEMENT 2545 growing mutant phenotypes. Mutants lacking the 49-kb plasmid may have been such slow growers. SDS-PAGE and Western blot analysis of the 19 mutants that were cloned indicated that all of them expressed a single protein of an abundance similar to that of OspA in the 28- to 34-kDa range. When this protein had the same apparent molecular weight as wild-type OspA, it also was able to bind strongly to the selecting antibody, whereas when the molecular weight was lower, the binding was diminished or not detectable. These latter results were interpretable in terms of having selected mutants expressing chimeric OspAB proteins that lacked a carboxyl-terminal portion of OspA. If the selecting anti-ospa antiserum contained mainly anti-carboxyl-terminal specificities, then, depending on the magnitude of the carboxylterminal deletion in the mutant OspA genes, the chimeric protein would or would not bind the anti-ospa antibody. The more detailed PCR and Western blot analysis of mutants I 5 and II 2, which did not bind to the selecting antibody, and of mutant II 1, which bound to it poorly, confirmed that these three mutants all contained an ospab operon that was shorter than the wild-type operon and that the protein expressed therefrom bound to MAbs that were specific to either C-terminal epitopes of OspB or N-terminal epitopes of OspA, but not the converse. In contrast, the OspA-like protein expressed by mutant I 1 was able to bind MAbs that were specific to both N- and C-terminal epitopes of OspA. Mutant I 1 also expressed a smaller version of OspB that bound only to MAbs specific for N-terminal OspB epitopes. While the smaller OspB protein could have arisen from a nonsense mutation in the ospb coding region, the result of the PCR analysis of mutant I 1 indicated that its ospab operon was shorter than that of the wild type. This was evidence that a deletion within the ospb coding region had taken place. Nucleotide sequencing of the I 1 ospab operon confirmed that a deletion had occurred between nucleotides 424 and 645 within the ospb gene. Hence, all of the mutants analyzed and likely all of the mutants obtained were ospab deletion mutants. Our functional analysis of mutants yielded a surprising result. All of the deletion mutants that were chosen for further study were killed in the presence of the selecting antiserum and rhesus monkey complement. Did this mean that deletion mutants that could be selected with an antiserum elicited by the OspA vaccine would have no effect on vaccine efficacy? The answer to this question depended entirely on where during the natural history of a B. burgdorferi infection anti-ospa antibody-dependent killing took place. If all of the killing took place within the tick, as some authors maintain (8), then the type of mutant selected with the vaccine antiserum could easily circumvent this phase of killing. Recently, Mather and coworkers (22) have shown that saliva of Ixodes scapularis, the deer tick, but not saliva of the dog tick Dermacentor variabilis harbors a factor able to degrade complement. The presence of this factor in ixodid ticks determines that these ticks are able to serve as spirochetal vectors, whereas Dermacentor ticks are not (22). In clonal populations of B. burgdorferi which are allowed to grow in vitro, the prevalence of mutants that resist killing by anti-ospa antibody ranges between 10 5 and 10 2 (28). If such frequencies are reproduced in a feeding nymph, in which spirochete numbers may reach a mean of 7,848 within 15 h of attachment (9), then several mutants may be present in a single tick. Deletion mutants such as those selected by us in vitro have been found in multiple strains of B. burgdorferi (26) and in several tick isolates from California (32). Since Mather et al. (22) found that complement is not active within the tick, this type of mutant, if it is able to traverse the tick midgut wall, will reach the tick salivary glands unscathed. On the other hand, if a significant fraction of spirochetal killing mediated by anti- OspA antibodies occurs after infection of the vertebrate host, i.e., a portion of the spirochetes continue to express surfaceexposed OspA, then deletion mutants will be killed by the anti-ospa antibody and complement. OspA vaccine efficacy would not be affected by such mutants. Several issues that are pertinent to this analysis need to be investigated further. The ability of ospab deletion mutants to penetrate the tick midgut, or, more generally, if such mutants are at all infectious via ticks, needs to be ascertained. An OspB escape mutant of B. burgdorferi N40 has been shown to be infectious to mice and has also been isolated from mice vaccinated with OspB (10, 12). This mutant has a nonsense mutation (TAA) at position 577, the C terminus of OspB (10), and thus expresses OspA and a smaller version of OspB. Such a mutant, which is similar to our I 1 mutant, is thus able to infect the vertebrate host, at least when it is delivered by needle inoculation. Another important issue is whether the escape mutants have retained the ability of the wild-type spirochetes to down-regulate the expression of the ospab operon. The result of our functional analysis indicates that if this downregulation is impaired as a side effect of the escape mutation, the deletion escape mutants will be killed upon infection of the vertebrate host, despite their ability to avoid antibody-dependent killing in the tick. We are currently analyzing these issues. Two final aspects merit discussion. First, deletion mutants that had retained the C terminus of OspB and the N terminus of OspA were killed by the anti-ospa antibody in the presence of complement, albeit at a 20- to 100-fold-higher antibody concentration. Binding of antibody to surface-exposed epitopes of B. burgdorferi antigens is required for the effective formation of a membrane attack complex via the classical pathway (17 19). Hence, we must assume that the anti-ospa polyclonal antibody contained specificities able to bind either to (newly) exposed N-terminal epitopes of OspA or to cross-reactive C-terminal epitopes of OspB. A comparison of the amino acid sequences of the 121 C-terminal amino acids of OspA and OspB, with both proteins aligned by the pam250 algorithm, yielded an identity of 52%. Such a level of homology could lead to antigenic cross-reactivity. Antibodies of such cross-reactive specificities may be present at low concentrations or be of low avidity. Either way, a higher concentration of the antiserum would have been required for effective killing to take place. Clearly, immunization protocols that lead to highaffinity antibodies and perhaps also the inclusion of OspB together with OspA in a cocktail vaccine may be useful to increase vaccine efficacy vis à vis ospab chimeric/deletion mutants. Second, a deletion mutant that retained and expressed a wild-type ospa gene was nonetheless able to resist killing with the anti-ospa antibody in the absence of complement. The mechanism of spirochetal killing mediated by antibody is not understood (27). It differs from ADCK, however, as indicated by the results presented in this report. In addition, it can be slower in acting: incubation of wild-type B. burgdorferi with the anti-ospa antiserum pool used herein in the absence of complement killed 15% of the spirochetes after 3.5 h, whereas 100% of the organisms were killed in this period when complement was present (22a). Thus, with this anti-ospa antiserum, antibody-dependent killing acts more slowly than ADCK. Our data suggest that epitopes present in the C terminus of OspB which are cross-reactive with anti-ospa antibody specificities are required for the antibody-mediated mechanism to be effective. It is possible that both homologous (A-A and B-B) and heterologous (A-B) cross-linking of Osps

7 2546 SOLÉ ET AL. INFECT. IMMUN. by antibody is necessary for antibody-mediated killing to take place. ACKNOWLEDGMENTS We gratefully acknowledge Denee Thomas, University of Texas Health Science Center, San Antonio, for her gift of MAbs to the N and C termini of OspB. The excellent secretarial help of Christie Trew as well as the photographic skill of Murphy Dowouis (TRPRC) and the skilled technical assistance of Cindy Gingrich-Baker (Aquila Pharmaceuticals) in making MAbs are acknowledged with thanks. This work was supported by grants U50/CCU from the Centers for Disease Control and Prevention and AI from the National Institutes of Health (both to M.T.P.) and by NCRR-NIH grant RR M.S. was supported in part by fellowship Bio 2-CT from the European Economic Community s project on the Biotechnology of Extremophiles. REFERENCES 1. Aydintug, M. K., Y. Gu, and M. T. Philipp Borrelia burgdorferi antigens that are targeted by antibody-dependent, complement-mediated killing in the rhesus monkey. Infect. Immun. 62: Barthold, S. W Lyme borreliosis in the laboratory mouse. J. Spirochetal Tick-Borne Dis. 3: Barthold, S. W., E. Fikrig, L. K. Bockenstedt, and D. H. Persing Circumvention of outer surface protein A immunity by host-adapted Borrelia burgdorferi. Infect. Immun. 63: Cinco, M Selection of a Borrelia burgdorferi antigenic variant by cultivation in the presence of increasing amounts of homologous immune serum. FEMS Microbiol. Lett. 92: Coleman, J. L., R. C. Rogers, and J. L. Benach Selection of an escape variant of Borrelia burgdorferi by use of bactericidal monoclonal antibodies to OspB. Infect. Immun. 60: Comstock, L. E., E. Fikrig, R. J. Shoberg, R. A. Flavell, and D. D. Thomas A monoclonal antibody to OspA inhibits association of Borrelia burgdorferi with human endothelial cells. Infect. Immun. 61: Coughlin R. T., D. Fish, T. N. Mather, J. Ma, C. Pavia, and P. Bulger Protection of dogs from Lyme disease with a vaccine containing outer surface protein (Osp) A, OspB, and the saponin adjuvant QS21. J. Infect. Dis. 171: de Silva, A. M., S. R. Telford, 3rd, L. R. Brunet, S. W. Barthold, and E. Fikrig Borrelia burgdorferi OspA is an arthropod-specific transmissionblocking Lyme disease vaccine. J. Exp. Med. 83: de Silva, A. M., and E. Fikrig Growth and migration of Borrelia burgdorferi in Ixodes ticks during blood feeding. Am. J. Trop. Med. Hyg. 53: Fikrig, E., H. Tao, F. S. Kantor, S. W. Barthold, and R. A. Flavell Evasion of protective immunity by Borrelia burgdorferi by truncation of outer surface protein B. Proc. Natl. Acad. Sci. USA 90: Fikrig, E., S. R. Telford III, S. W. Barthold, F. S. Kantor, A. Spielman, and R. A. Flavell Elimination of Borrelia burgdorferi from vector ticks feeding on OspA-immunized mice. Proc. Natl. Acad. Sci. USA 89: Fikrig, E., H. Tao, S. W. Barthold, and R. A. Flavell Selection of variant Borrelia burgdorferi isolates from mice immunized with outer surface protein A or B. Infect. Immun. 63: Gern, L., O. Rais, C. Capiau, P. Hauser, Y. Lobet, E. Simoen, P. Voet, and J. Pêtre Immunization of mice by recombinant OspA preparations and protection against Borrelia burgdorferi infection induced by Ixodes ricinus tick bites. Immunol. Lett. 39: Golde, W. T., T. R. Burkot, J. Piesman, M. C. Dolan, C. Capiau, P. Hauser, G. Dequesne, and Y. Lobet The Lyme disease vaccine candidate outer surface protein A (OspA) in a formulation compatible with human use protects mice against natural tick transmission with Borrelia burgdorferi. Vaccine 13: Kalish, R. A., A. C. Steere, and J. M. Leong Early and late antibody responses to full-length and truncated constructs of outer surface protein A of Borrelia burgdorferi in Lyme disease. Infect. Immun. 63: Keller, D., F. T. Koster, D. H. Marks, P. Hosbach, L. F. Erdile, and J. P. Mays Safety and immunogenicity of a recombinant outer surface protein A Lyme vaccine. JAMA 271: Kochi, S. M., R. C. Johnson, and A. P. Dalmasso Complementmediated killing of the Lyme disease spirochete Borrelia burgdorferi. Role of antibody in formation of an effective membrane attack complex. J. Immunol. 146: Kochi, S. M., R. C. Johnson, and A. P. Dalmasso Facilitation of complement-dependent killing of the Lyme disease spirochete, Borrelia burgdorferi, by specific immunoglobulin G Fab antibody fragments. Infect. Immun. 61: Kochi, S. M., and R. C. Johnson Role of immunoglobulin G in killing of Borrelia burgdorferi by the classical complement pathway. Infect. Immun. 56: Ma, J. P., R. T. Coughlin, P. Bulger, P. M. Franchi, and C. Gingrich-Baker Molecular analysis of neutralizing epitopes on outer surface proteins A andbofborrelia burgdorferi. Infect. Immun. 63: Ma, J. P., M. Hine, E. R. Clough, D. Fish, R. T. Coughlin, G. A. Beltz, and M. G. Shew Safety, efficacy, and immunogenicity of a recombinant Osp subunit canine Lyme disease vaccine. Vaccine 14: Mather, T. N., M. T. Yeh, J. M. C. Ribeiro, A. Scorpio, D. R. Nelson, and R. T. Coughlin Ixodes saliva: vector competence for Borrelia burgdorferi and potential vaccine strategies, abstr. A029. In Abstracts of the VII International Congress on Lyme Borreliosis. San Francisco, Calif. 22a.Nowlings, J., and M. Philipp. Unpublished data. 23. Philipp, M. T., and B. J. B. Johnson Animal models of Lyme disease: pathogenesis and immunoprophylaxis. Trends Microbiol. 2: Philipp, M. T., Y. Lobet, R. P. Bohm, Jr., M. D. Conway, V. A. Dennis, P. Desmons, Y. Gu, P. Hauser, R. C. Lowrie, Jr., and E. D. Roberts Safety and immunogenicity of recombinant outer surface protein A (OspA) vaccine formulations in the rhesus monkey. J. Spirochetal Tick-Borne Dis. 3: Philipp, M. T., Y. Lobet, R. P. Bohm, Jr., E. D. Roberts, V. A. Dennis, Y. Gu, R. C. Lowrie, Jr., P. Desmons, P. H. Duray, J. England, P. Hauser, J. Piesman, and K. Xu Efficacy of recombinant outer surface protein A (OspA) vaccine formulations in the rhesus monkey. Vaccine 15: Rosa, P. A., T. Schwan, and D. Hogan Recombination between genes encoding major outer surface proteins A and B of Borrelia burgdorferi. Mol. Microbiol. 6: Sadziene, A., and A. G. Barbour Growth inhibition of Borrelia burgdorferi sensu lato by antibodies. A contribution to understanding the pathogenesis and improving diagnosis of Lyme borreliosis. Wien. Med. Wochenschr. 145: Sadziene, A., P. A. Rosa, P. A. Thompson, D. M. Hogan, and A. G. Barbour Antibody-resistant mutants of Borrelia burgdorferi: in vitro selection and characterization. J. Exp. Med. 176: Sanger, F. S., S. Nicklen, and A. R. Coulson DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74: Schoen, R. T., F. Meurice, C. M. Brunet, S. Cretella, D. S. Krause, J. E. Craft, and E. Fikrig Safety and immunogenicity of an outer surface protein A vaccine in subjects with previous Lyme disease. J. Infect. Dis. 172: Schutzer, S. E., P. K. Coyle, and M. Brunner Identification of specific Borrelia burgdorferi components in circulating antigen-antibody complexes, p In S. E. Schutzer (ed.), Lyme disease: molecular and immunologic approaches. Current communications in cell and molecular biology. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. 32. Schwan, T. G., M. E. Schrumpf, R. H. Karstens, J. R. Clover, J. Wong, M. Daugherty, M. Struthers, and P. A. Rosa Distribution and molecular analysis of Lyme disease spirochetes, Borrelia burgdorferi, isolated from ticks throughout California. J. Clin. Microbiol. 31: Shoberg, R. J., M. Jonsson, A. Sadziene, S. Bergstrom, and D. D. Thomas Identification of a highly cross-reactive outer surface protein B epitope among diverse geographic isolates of Borrelia spp. causing Lyme disease. J. Clin. Microbiol. 32: Telford, S. R., III, F. S. Kantor, Y. Lobet, S. W. Barthold, A. Spielman, R. A. Flavell, and E. Fikrig Efficacy of human Lyme disease vaccine formulations in a mouse model. J. Infect. Dis. 171: van Hoecke C., M. Comberbach, D. de Grave, P. Desmons, D. Fu, P. Hauser, E. Lebacq, Y. Lobet, and P. Voet Evaluation of the safety, reactogenicity and immunogenicity of three recombinant outer surface protein (OspA) Lyme vaccines in healthy adults. Vaccine 14: Wormser, G. P Lyme disease vaccine. Infection 24: Editor: J. G. Cannon