The medical and scientific communities are still searching for an explanation for the Lyme disease (LD) cases reported in the United States each year with no serologic or culture confirmation of infection with the agent of LD,
Borrelia burgdorferi sensu stricto. Patients bitten by Ixodidae ticks in the southeastern or south-central United States may suffer from southern tick-associated rash illness (STARI) caused by an unknown etiologic agent or an inflammatory response (
1-
3,
5,
9,
10). Evidence suggests that persons in these regions bitten by the hard tick,
Amblyomma americanum, who develop a red, expanding rash with central clearing (indistinguishable from erythema migrans, the hallmark rash of LD) are infected with a spirochete named
Borrelia lonestari (
3,
4,
8,
18).
Since Barbour et al. first characterized the 16S rRNA and flagellin (
fla) genes of
B. lonestari in 1996 (
3), there have been few published reports extending our knowledge of this newfound spirochete. DNA evidence of
B. lonestari infections has been documented in the following: (i)
A. americanum ticks from Alabama, Arkansas, Delaware, Kansas, Kentucky, Maryland, Missouri, New Jersey, North Carolina, Tennessee, Texas, and Virginia (
1,
3,
4,
17,
18); (ii) blood samples from white-tailed deer,
Odocoileus virginianus, in Arkansas, Georgia, North Carolina, and South Carolina (
11); and (iii) the skin rash biopsy sample and attached
A. americanum tick from one human patient with possible exposure in Maryland or North Carolina (
8). However, there have been no published reports of successful cultivation, additional gene targets, or serologic assays developed to aid clinicians in determining a diagnosis of
B. lonestari as an etiologic agent of STARI.
In 1996, Schwan et al. reported the identification of an immunodominant enzyme in
Borrelia hermsii called glycerophosphodiester phosphodiesterase (GlpQ) and demonstrated that serum from humans infected with relapsing fever group (RFG) spirochetes reacted positively by immunoblotting with GlpQ recombinant antigen, while serum from patients with LD did not (
15). Later, Schwan et al. found by Southern blot analysis with
B. hermsii glpQ probes that eight species of RFG spirochetes contained a
glpQ ortholog and six LD
Borrelia species did not (
14). With these data, and given that
B. lonestari is more closely related to RFG spirochetes than to LD spirochetes based on 16S ribosomal DNA (rDNA) and
fla analyses (
3,
4,
7,
8,
13), we decided to look for an ortholog of
glpQ in
B. lonestari-positive tick DNA. The 16S rRNA and
fla genes are highly conserved and are present in all
Borrelia species, making them ill suited for development of rapid, differential diagnostic assays. If present in
B. lonestari,
glpQ could provide a target for distinguishing
B. lonestari- from
B. burgdorferi-infected ticks and human samples.
In 2003, we reported the presence of
B. lonestari DNA in
A. americanum collected from southeast Missouri, confirmed by sequence analyses of 16S rRNA and the complete
fla genes (GenBank accession no.
AY166715 and
AY166716) (
1). The gene encoding outer surface protein A (OspA) was not detected by PCR using
B. burgdorferi sensu lato-specific primers (
6). With few exceptions, OspA is present in all LD-causing borreliae, indicating that the samples likely did not contain DNA from these species. This provided ample
B. lonestari-positive tick DNA template for future studies. A subset of
A. americanum ticks from the same Missouri field collection was sent to the University of California—Irvine and processed for DNA extraction as reported in detail elsewhere (
3). An aliquot of this tick DNA extract was sent to the Rocky Mountain Laboratories for independent PCR and DNA sequence analysis (
1).
PCR primers that amplify the complete coding sequence of glpQ from B. lonestari and Borrelia miyamotoi were designed from B. hermsii noncoding sequences flanking the glpQ gene (forward, 5′GGTATGCTTATTGGTCTTC3′; reverse, 5′TTGTATCCTCTTGTAATTG3′). Identical coding sequences were determined for the B. lonestari glpQ gene by the Centers for Disease Control and Prevention and Rocky Mountain Laboratories. The complete coding sequence for B. miyamotoi glpQ was determined using genomic DNA of strain FR64B.
A comparison of the
glpQ sequences from
B. lonestari,
B. miyamotoi, and orthologs in other
Borrelia spp. determined previously (
12,
15) was performed using the ClustalV (PAM250) algorithm of MegAlign (DNASTAR, Inc., Madison, Wis.) (Fig. ).
B. lonestari glpQ was most similar to
B. miyamotoi glpQ (85% identical) and had less identity with orthologs found in
B. hermsii (83.4%),
Borrelia parkeri (81.9%),
Borrelia turicatae (81.3%),
Borrelia crocidurae (76.9%),
Borrelia coriaceae (79.5%), and
Borrelia recurrentis (76.6%). The dendrogram indicates a high level of similarity between
B. crocidurae and
B. recurrentis (99.7% identity) and between
B. parkeri and
B. turicatae (97.6% identity), as previously shown (
12). The
B. lonestari and
B. miyamotoi glpQ genes encode 336 and 334 amino acids, respectively. Both have putative signal peptide sequences followed by a cysteine cleavage site at amino acid 21, as reported for orthologs from other borreliae (
12,
15) (Fig. ).
B. lonestari contains a threonine at position 28 that is absent in
B. miyamotoi and a leucine at position 32 that is absent in all other sequences aligned.
Our discovery of
glpQ in
B. lonestari strengthens the notion that
B. lonestari is more closely related to the RFG spirochetes than to the LD-causing borreliae. A
glpQ ortholog has never been found in spirochetes that cause LD (
14). Others have shown that
B. lonestari,
B. miyamotoi, and the agent of bovine borreliosis,
Borrelia theileri, form a sister group based on 16S rDNA and
fla sequence analyses (
13). DNA evidence of another RFG spirochete (Yale spirochete), also closely related to
B. miyamotoi, has been obtained from
Ixodes scapularis (
16). Considering that these borreliae are the only RFG spirochetes known to infect hard ticks, we predict that
B. theileri and the Yale spirochete also contain a
glpQ ortholog. If so, screening field-collected hard ticks for
glpQ and
ospA by PCR could be more informative and faster than PCR, DNA analysis, or molecular subtyping based on 16S rDNA and/or
fla, since these targets are common to all borrelia.
B. lonestari infecting A. americanum was first reported in 1996, but the impact of this organism on human health remains uncertain. One probable case of human infection with B. lonestari was reported in 2001, but all other published research has focused on finding evidence of B. lonestari in vector ticks and mammalian hosts. Although ecologic studies are important in understanding the zoonotic cycle of this organism, more research is needed to determine whether the bite of A. americanum is associated with human borreliosis and whether B. lonestari is an etiologic agent of STARI. Because of the similarity in rash illness resulting after the bite of Amblyomma and Ixodes ticks and the lack of early antibody responses in many of these patients, researchers must rely on PCR-based assays for differential diagnosis of borrelia in rash biopsy samples, emphasizing the importance of finding glpQ in B. lonestari. However, a need to cultivate B. lonestari to identify specific antigens to aid development of a serologic test still exists, since human antibody to B. lonestari GlpQ is likely to cross-react with other RFG orthologs.
Here we identify the glpQ gene of B. lonestari and present its sequence with that of B. miyamotoi glpQ. There is enough sequence divergence among the borrelia glpQ genes to design DNA-based assays for rapid identification of B. lonestari in tick or mammal surveys and in human samples from STARI patients. To this end, efforts to develop and validate a rapid, sequence-based assay specific for B. lonestari glpQ are in progress.
This article has been 
