DNA typing of rickettsiae in naturally infected ticks using a polymerase chain reaction/restriction fragment length polymorphism system.

We used the polymerase chain reaction/restriction fragment length polymorphism (PCR/RFLP) rickettsial typing system of Regnery and others to rapidly identify rickettsiae in naturally infected ticks. Unlike previously described methods, our PCR assays type rickettsiae directly from tick tissues without first isolating the organisms. We collected 226 adult Dermacentor andersoni ticks in the Bitterroot Mountains of western Montana and analyzed them for possible rickettsial infection by hemolymph test using the Gimenez stain. Thirteen (5.8%) of these ticks were positive by hemolymph test and selected for further analysis using the above PCR/RFLP typing system. The PCR assays performed using the first primer set (RpCS) resulted in amplification of fragments of the predicted size from nine of the 13 hemolymph test-positive tick samples. Only four of these nine tick samples were also positive in similar PCR assays performed with a second primer set (Rr190) that is presumed to be spotted fever group specific. The RFLP analyses of material amplified from these four ticks indicated they were infected with Rickettsia rickettsii (one sample) and R. rhipicephali (three samples). The PCR/RFLP analyses of the five PCR-positive tick samples that were positive only in assays performed with the RpCS primer set indicated that these ticks were infected with R. bellii. The remaining four of 13 hemolymph test-positive tick samples gave negative PCR results with both the RpCS and Rr190 primer sets. Infected hemocytes from these PCR-negative ticks contained organisms of distinctive bacillary morphology that appeared similar to those described previously as long forms, and it is possible that these organisms belong to a genus other than Rickettsia. We also examined established laboratory isolates of tick-borne rickettsiae from different regions of North America to determine whether this typing system produces consistent results. Multiple isolates of R. montana (nine isolates), R. bellii (five isolates), R. rickettsii (Hlp-like) (four isolates), and R. canada (two isolates) were tested and no significant variations in PCR/RFLP patterns were observed between members of the same serotypes. However, among the five isolates of R. rhipicephali tested, two slightly different RFLP patterns were noted. Our results suggest that this PCR/RFLP typing scheme has wide applicability for identifying rickettsiae directly from D. andersoni or D. variabilis tick tissues.