Ixodid tick-associated spirochetes have been implicated as the etiological agents of Lyme disease. We raised a murine monoclonal antibody (H5332) against a spirochete, strain B31, isolated from Ixodes dammini ticks. In indirect immunofluorescence assays and western blot analyses, H5332 reacted with whole cells or isolated components of not only strain B31 but also spirochetes isolated from Ixodes ricinus ticks, a field mouse, a raccoon, and patients with Lyme disease. In contrast, H5332 did not bind to representative borreliae, treponemes, and leptospires. Using indirect immunofluorescence assays and immune electron microscopy, we found the H5332 determinant to be diffusely distributed over the surface of prefixed spirochetes but to be aggregated in patches when the organisms were incubated with H5332 and a second ligand before fixation. Radioimmunoprecipitation and western blot studies revealed the H5332 determinant to be either on or tightly associated with an abundant outer membrane protein with an apparent subunit molecular weight of 31,000.

A simple procedure for extraction of plasmid-enriched DNA from borreliae was used in a plasmid analysis of 13 strains of the Lyme disease agent, Borrelia burgdorferi. The extracted DNA was subjected to low-percentage agarose gel electrophoresis and examined either directly by ethidium bromide staining or after hybridization of the plasmids in situ with a DNA probe for the gene encoding the major outer membrane protein OspA. Each isolate had four to seven discernible plasmids of various sizes. Only 2 of the 13 strains had the same plasmid profile. The ospA gene probe hybridized to large plasmids to strains from both North America and Europe. A strain which had been passaged many times was found to have lost two of the six plasmids originally present. These findings indicate the potential usefulness of plasmid analysis as a strain-typing procedure and for identifying possible plasmid-conferred virulence factors.

Borrelia burgdorferi, the causative agent of Lyme disease, loses its ability to infect and cause disease in mammalian hosts after repeated in vitro passage. To identify proteins preferentially expressed by the low-passage strain and thus representing potential virulence factors, the polypeptide profiles of virulent, low-passage and nonvirulent, high-passage forms ofB. burgdorferi B31 were compared by nonequilibrium pH gradient two-dimensional gel electrophoresis. Four low-passage-associated proteins with relative molecular masses (Mrs) of 35,000, 28,000, 24,000, and 20,000 were identified. Of these, the 28and 35-kDa polypeptides were not expressed in detectable quantities in the high-passage B31 strain, whereas the 24-and 20-kDa proteins were present in reduced quantities. All four of these proteins were lipoproteins, as determined by labelling with[3H]palmitate. The abundant 28-kDa component, called outer surface protein D (OspD), is surface exposed on the basis of its proteolysis during treatment of intact organisms with proteinase K. The ospD gene is located on a 38-kb linear plasmid present in seven of nine low-passage strains of B. burgdorferi examined but absent in most high-passage, nonvirulent strains tested. Molecular cloning and sequence analysis of the ospD gene locus revealed an open reading frame encoding a 28,436-Da polypeptide with a putative signal peptidase II leader sequence. An unusual feature of the region upstream of the gene was the presence of seven contiguous, direct repeats of a 17-bp sequence that includes consensus -35 and -10 transcription initiation signals; however, only one transcription initiation site was active as determined by primer extension analysis. Further study of these and other polypeptides associated with low-passage strains may lead to identification of B. burgdorferi gene products required for infection and pathogenesis in mammalian hosts.

We were unable to deMonstrate the presence of the classic enterobacterium-type lipopolysaccharide in the cells of the Lyme disease spirochete, Borrelia burgdorferi B31. This finding was primarily based on chemical analysis and the absence of free lipid A upon mild acid hydrolysis of the appropriate cell extracts. These re4ults do not preclude the possible existence of an unusual lipopolysaccharide-like compound(s) in B. burgdorferi. Beck et al. (4) recently reported the presence of lipopolysaccharide (LPS) irq the Lyme disease spirochete (Borrelia burgdorferi). We wanted to confirm their results and then examine the nature of the lipid A moiety of LPS from this source. We extracted the cells of B. burgdorferi B31 by two well-known methods used in the isolation of LPS from gram-negative bacteria. The appropriate fractions obtained by these two methods were hydrolyzed under mild acid conditions and examined by thin-layer chromatography (TLC) for the presence of free lipid A. Chemical analyses for the presence of 2-keto-3-deoxyoctonate, glucosamine, and hydroxy fatty acids were performed on the extracts. We found no evidence for the presence of lipid A in any of the fractions tested. Thus, we must conclude that the classic LPS associated with gram-negative bacteria is absent in cells of B. burgdorferi.Cells of B. burgdorferi B31 (ATCC 35210) were grown in BSK II medium (1), washed three times with 50 mM Tris (pH 7.4)-150 mM NaCI-5 mM MgCI2 (2), and lyophilized. In one experiment, the extraction method of Galanos et al. (7) with modifications was used (12,14). The results of this extraction are summarized in Table 1. Almost all of the extractable material was found in the 90% aqueous ethanol fraction, whereas very little material appeared in the phenol-waterchloroform-petroleum ether fraction, which should contain the rough-type LPS (R-LPS). This latter fraction was hydrolyzed in 0.1 N HCI as previously described for the preparation of monophosphoryl lipid A (14), and 30 ,ug of the organic extract was analyzed by TLC for the presence of lipid A (12,14). A mixture of hexa-, penta-, and tetraacyl monopliosphoryl lipid A (from LPS of Salmonella typhimurium) was used as a chromatographic standard (13). There was no evidence of the presence of free lipid A in this sample, indicating the absence of R-LPS in the original extract (phenol-water-chloroform-petroleum ether).Previous studies of lipids X and Y (monosaccharide precursors of LPS) showed that much of X and Y was extracted into the aqueous ethanol fraction (15,16). Thus, it is conceivable that the 113.2 mg of the 90% ethanol extract (Table 1) might contain the R-LPS. This material was initially fractionated by solvent precipitation to yield 79.9 mg of an acetone-insoluble glycolipid-phospholipid fraction. This * Corresponding author. fraction was further purified by preparative TLC on Silica Gel G (Merck & Co., Inc.) by a solvent system of chloroform-methanol-water (65:25:4). Five major fractions (designated A to E) were obtained. Each of these fractions was acid hy...

Relapsing fever and Lyme disease spirochetes of the genus Borrelia display at their surfaces abundant lipoproteins: Vmp proteins in Borrelia hermsii and Osp proteins in Borrelia burgdorferi. Vmp and Osp proteins largely determine serotype specificity, and neutralizing antibodies of infected or immunized animals are directed at them. For the present study, we examined B. hermsii serotype 33, which is unique among strain HS1 serotypes in the low frequency of switches to other serotypes during infections and in vitro cultivation. Failing to clone the complete vmp33 gene, we accomplished its further characterization by (i) determining three partial amino acid sequences, (ii) designing oligonucleotide primers based on these amino acid sequences, (iii) cloning and sequencing the central portion of vmp33, and (iv) using outwardly directed primers and the inverse PCR to clone the 5' and 3' ends of the gene and flanking regions. The transcriptional start site was identified by primer extension analysis. Vmp33 was a polypeptide of 211 amino acids; the three partial amino acid sequences were identified in the open reading frame. Vmp33 was found to be more similar to other 20-kDa Vmp proteins of B. hermsii and to OspC proteins of B. burgdorferi than it was to 35to 39-kDa Vmp proteins of the same strain. Moreover, OspC proteins were more similar to Vmp33 than they were to OspA,-B, or-D proteins of B. burgdorferi. These sequence similarities were consistent with Western blot (immunoblot) findings of crossreactions between Vmp33 and OspC with anti-Vmp33 and anti-OspC sera. The promoter for the expressed vmp33 gene was found to be different from the expression site for other active vmp genes characterized to date. These results indicate that Vmp33 and other small Vmp's belong with OspC to a genus-wide family of 20-kDa proteins and that expression of these proteins may be coordinated with expression of other Vmp and Osp proteins in Borrelia spp.

Spirochetes of the genusSpirochetes are a distinct phylum of eubacteria (41, 55). Known pathogenic members of this division are in the genera Treponema, Borrelia, and Leptospira. The genus Borrelia comprises microorganisms that are transmitted to vertebrates by hematophagous arthropods. Included in this genus are B. hermsii and B. turicatae, agents of relapsing fever in North America; B. burgdorferi sensu lato, the cause of Lyme disease; B. anserina, the agent of avian spirochetosis; and B. coriaceae, the putative agent of epidemic bovine abortion (5).The genomes of Borrelia spp., like those of many other bacteria, are composed of a chromosome and plasmids. What distinguishes the Borrelia genome from that of most other bacteria is its largely linear structure. The chromosome of B. burgdorferi is a linear duplex DNA molecule about 1 Mb in length (6,13,14,16,17). Linear plasmids ranging in size from 5 to 55 kb are present in B. burgdorferi and B. hermsii (2, 32, 43, 54); These linear plasmids carry genes for outer membrane lipoproteins, including OspA, OspB, and OspD in B. burgdorferi (3, 4) and the serotype-specifying Vmp proteins of B. hermsii (31, 43).Borrelia spp. also have circular DNA molecules, the form of extrachromosomal DNA in most bacteria (30, 51). Although these constitute only a small fraction of the Borrelia genome, circular plasmids carry what appear to be critical genes. In B. burgdorferi, these include ospC, which codes for an outer surface protein (35,46), and homologs of guaA and guaB, which code for enzymes involved in de novo purine biosynthesis (36).The structure of linear plasmids in B. burgdorferi has been the subject of detailed examination. Electron microscopic examination of the 49-kb linear plasmid of this species showed that the two strands of the duplex DNA are covalently closed at each end (4). The telomeres of the 16-and 49-kb plasmids of B. burgdorferi have been cloned, and sequence analysis of the telomeres revealed that there are short inverted repeats at their termini and that they have terminal hairpin loops (25, 28). A similar telomeric sequence was found on linear plasmids of B. hermsii, suggesting that other linear plasmids of this genus have similar structure (31). The telomeres of linear plasmids in Borrelia spp. have structural and sequence homology to those of some linear eukaryotic DNA viruses, specifically poxviruses and African swine fever virus (25).B. hermsii and B. burgdorferi have several genome equivalents arrayed along the lengths of the wavy filamentous cells (32). Linear and circular plasmids have a copy number of approximately one to two per chromosome in both B. burgdorferi and B. hermsii (14,26,32). This arrangement suggests a tight coupling of plasmid and chromosome replication and segregation in borrelias.To further understand the genetic organization of Borrelia spp., we examined the genomes of B. hermsii, B. burgdorferi, B. turicatae, and B. anserina in more detail. In the course of this characterization, we found linear double-stranded DNA molecules w...

Borrelias that cause Lyme disease lose the ability to infect and cause disease in laboratory animals following 10 to 16 passages of in vitro culture. In this study, clonal populations of the Sh-2-82 (Sh2) and B31 strains of Borrelia burgdorferi were isolated by subsurface plating on BSK-II agar plates and examined for infectivity in the C3H/HeN mouse model. Mice were injected intradermally with 10 5 B. burgdorferi organisms, and the tibiotarsal joint, heart, and bladder were cultured 2 to 4 weeks postinfection to determine whether viable organisms were present. Clones exhibited either a high-infectivity phenotype, in which cultures were consistently positive at all organ sites, or a low-infectivity phenotype, in which a low proportion of cultures were positive (5 of 40 in a representative experiment). In an Sh2 population that had undergone five in vitro passages, 7 of 10 clones were of the high-infectivity phenotype, and the remaining clones were of the lowinfectivity phenotype. The proportion of high-infectivity clones decreased with continued in vitro passage, with only 1 of 10 clones exhibiting the high-infectivity phenotype after 10 passages and 0 of 10 clones yielding positive cultures after 20 passages. Representative high-and low-infectivity clones from passage 5 Sh2 cultures had 50% infectious doses of 1.8 ؋ 10 2 and 1 ؋ 10 5 , respectively. Subclones consistently reflected the same infectivity phenotypes as those of the parent clones. The protein profiles and plasmid contents of the high-and low-infectivity clones were compared and exhibited few discernible differences. On the basis of these results, the loss of infectivity during in vitro culture results from the outgrowth of low-infectivity clones and begins to occur within the first five in vitro passages. Further examination of clonal populations may lead to the identification of genetic and protein factors important in the virulence and pathogenicity of Lyme disease borrelias.

Immunodeficient mice infected with Borrelia turicatae, a relapsing fever agent, have a disorder that resembles disseminated Lyme disease. Two serotypes, A and B, differed in their arthritogenicity in both CB-17 SCID and C3H SCID mice. In CB-17 SCID mice infected with serotype A or B, arthritis was assessed by measurement of tibiotarsal diameter, functional ability on a beam walk test, and microscopic assessment of joint inflammation. Serotype B-infected mice had greater joint swelling, functional disability, and leukocytic infiltration in the joints than serotype A-infected mice. Joint swelling and disability peaked at 2 weeks of infection and then decreased, while leukocyte infiltration in the joints persisted. To investigate the basis for the differences in arthritogenicity of serotypes A and B, spirochete burdens in infected mice were measured by quantitative PCR of spirochete DNA in joints, direct immunofluorescence of spirochetes in joints, and counts of spirochetes in the blood. At 2 weeks of infection there were seven times more spirochetes in the joints of serotype B-infected mice than in those of serotype A-infected mice, measured by both quantitative PCR and direct enumeration. Although serotypes A and B had the same infectivity and growth rate in vivo, serotype B spirochetes were eightfold more abundant in the blood than serotype A spirochetes and produced greater fatality in newborn mice. These findings indicate that differences in disease severity in mice infected with serotype A or B are attributable to differences in the spirochete burden in the joints and blood.