We designed a genus-specific set of primers (BOR1: 5´ TAA TAC GTC AGC CAT AAA TGC 3´ and BOR2: 5´ GCT CTT TGA TCA GTT ATC ATT C 3´) that targeted the Borrelia flaB flagellin gene (13). Each PCR mixture (25 μL) contained 3 μL of target DNA and was subjected to 1 min at 95°C, followed by 40 cycles of 56°C for 30 sec, 72°C for 30 sec, 94°C for 30 sec, and 5 min at 72°C for final elongation. DNA of B. duttonii and B. burgdorferi sensu stricto (strain B31) was used as controls. DNA of blood and ticks was extracted with the DNA easy tissue kit (Qiagen, Hilden, Germany). Tick samples were collected by using CO₂ traps in caves and were identified as Ornithodoros tholozani (Figure 1) by the Entomology Laboratory (Ministry of Health, Jerusalem). The tick specimens collected were tested either individually or as pools. Of 184 tick specimens collected from 5 different areas (Table 1), 94 were tested by BOR1-BOR2 PCR. One pool of 5, a pool of 4, and 6 individual specimens were positive; all produced a unique band 750 bp in length. The percentage of tick infection was variable, ranging from <2% in Ma'ale-Adumim to 40% in the Be'er Sheva region.

For patients, the TBRF diagnosis was established as previously reported (5). Eighteen samples of human blood were sent to the Parasitology Reference Center (Ministry of Health, Jerusalem); the samples corresponded to 15 confirmed cases (positive blood smear) and 3 associated cases of TBRF (negative blood smear). On receipt at the laboratory, fresh human blood samples were examined by darkfield microscopy for viable Borrelia and, if detected, 200 μL of blood was injected into 10-mL vials of BSK-H medium (14) and into 10-week-old ICR mice by the intraperitoneal route. In 4 patients, blood examined by darkfield microcopy showed 1–5 motile Borrelia per slide. In vitro cultivation was unsuccessful. However, Borrelia (1–5/field) were detected on day 4 (twice) and day 6 (twice) in the blood of mice injected intraperitoneally with patient blood. Cultivation attempts from positive mice blood were also unsuccessful. In contrast, all the samples were found positive by BOR1-BOR2 PCR, showing a unique band of 750 bp (data not shown).

PCR products were cloned in T7 plasmid by pGEM-T Easy vector SystemII (Promega, Madison, WI, USA). Plasmids containing inserts were purified and sent for 2-strand sequencing. Direct sequencing of DNA amplified by the BOR1 and BOR2 primers was performed later.

Phylogenetic and molecular evolutionary analyses were conducted by using MEGA version 3.1 (15). Among published flaB genes of TBRF Borrelia strains, only sequences for which a translated protein existed were taken in account. Because of the large number of available sequences for American Borrelia associated with TBRF, as well as for B. duttonii and B. recurrentis, a single sequence representative of each cluster was chosen for taxonomic analyses.

Three PCR amplicons (from 1 tick and 2 human samples) were sequenced after cloning, whereas 6 amplicons (from 2 ticks and 4 human samples) were analyzed by direct sequencing. These 9 sequenced samples showed 98%–100% homology between them and could be divided into 3 groups. The same DNA sequence was found in tick TG52 and in blood from 2 patients, HumanBlood2 and HumanBlood4. These 3 sequences had an additional triplet at the position 627. The second group of sequences, which consisted of tick samples TGd1 and CBkc7 and blood samples C1025B, FL1, and HumanBlood3, were identical, with only 3 minor substitutions between them. The third group consisted of the HumanBlood1 sample.

All the translated sequenced amplicons showed a very specific signature by the presence of a 7–amino acid (aa) gap at position 216 (Figure A1) when compared with previously described TBRF Borrelia flaB genes. In addition, the local TBRF Borrelia sequences could be grouped into 3 subtypes, according to variation at 7-aa positions (Table 2).

Comparison with published flaB protein sequences of TBRF Borrelia showed 88%–90% homology with B. duttonii and B. recurrentis, 85%–90% with B. crocidurae, 86%–88% with B. turicatae, 87%–89% with B. hermsii, and 85%–88% with B. parkeri. The sequences of the Israel TBRF Borrelia isolated from different samples clearly clustered in a separate group from the American and the African TBRF species (Figure 2).

Our results suggest that infection rates differ according to location, despite the small number of ticks tested and the use of pools. BOR1-BOR2 PCR was more sensitive than blood smear examination (100% vs 83%). An identical DNA sequence was found in both tick and patient samples and thus confirms, at the molecular level, the role of O. tholozani as the vector of TBRF in Israel.

A signature (7-aa gap) of the flaB flagellin defined the Israeli TBRF sequences as a homologous group different from other TBRF species. Despite the small number of samples studied, a clear polymorphism existed also at the protein level, resulting in 3 local types. This diversity can be explained by the use of direct sequencing of samples rather than through cultivation that reduces the biodiversity of isolates by selecting the most successful in vitro clone.

This study opens a new avenue in TBRF Borrelia studies by demonstrating a Middle East cluster in addition to the American and African groups. These results open encouraging perspectives for the better understanding of entomologic, epidemiologic, and bacteriologic aspects of this disease and may contribute to better diagnosis and treatment.