For molecular analyses, DNA from the infected cells of each of the 3 isolates was extracted by boiling at 100°C for 10 min; it yielded products of the expected size through PCR protocols selective for portions of 3 genes of the genus Coxiella: primers QR-FO (5′-ATTGAAGAGTTTGATTCTGG-3′) and QR-RO (5′-CGGCCTCCCGAAGGTTAG-3′) for the 16S rRNA gene (8); primers CAPI844F (5-ATTTAGTGGGTTTCGCGCAT-3′) and CAPI844R (5′-CATCAGCATACGTTTCGGGAA-3′) for the cap gene (9); and primers Cox-F-pry2 (5′-TTATTTACCAACGTTCCTGAGCCG-3′) and Cox-R-pry2 (5′-TTTATCCCGAGCAAATTCAATTATGG-3′) for the pyrG gene (9). PCR products underwent DNA sequencing in an automatic sequencer (Applied Biosystems/PerkinElmer, Foster City, CA, USA) according to the manufacturer`s protocol. We sequenced 1,386, 557, and 545 nt of the genes 16S rRNA, cap, and pyrG, respectively, which were identical to each other for each gene amplified from the 3 tick isolates. By BLAST analyses (www.ncbi.nlm.nih.gov/blast), these sequences were 99.9% (1,384/1,386 nt), 99.6% (556/558 nt), and 99.6% (452/454 nt) identical to the corresponding GenBank sequences of the North American C. burnetii genes 16S rRNA, cap, and pyrG, respectively (HM208383, CP001020, CP001020). Partial sequences (16S rRNA, cap, pyrG) from C. burnetii generated in this study were deposited into GenBank and assigned nucleotide accession nos. JQ740886–JQ740888, respectively.

Infected Vero cell monolayers were fixed in a modified Karnovsky solution, stained with uranyl acetate and lead citrate, and examined in a transmission electron microscope according to standard procedures. Ultrastructurally, Coxiella organisms were identified by morphologic features within heavily infected Vero cells. The organisms possessed typical bacillary morphologic characteristics and were observed inside vacuoles (phagolysosomes) of different sizes, proportional to the number of organisms (Figure, panels D, E). Intravacuolar organisms had a mean length of 0.55 ± 0.13 µm (range 0.42–0.85 µm) and a mean width of 0.25 ± 0.03 µm (range 0.22–0.32 µm).

Ticks negative for rickettsiae-like organisms by hemolymph testing were subjected individually to DNA extraction by the guanidine isothiocyanate-phenol technique (10) and screened for Coxiella spp. by PCR that targeted the pyrG gene, as described above. Although no A. parvum tick yielded amplicons, 40 A. tigrinum ticks yielded amplicons of the expected size for the pyrG gene. DNA sequences generated from these ticks were identical to the pyrG partial sequences obtained from the C. burnetii isolates mentioned above.

We found 41 (44.6%) of 92 ticks and 2 (15.4%) of 13 of the A. tigrinum and A. parvum adult ticks, respectively, to be infected by C. burnetii. Because these ticks were collected as engorged nymphs from wild rodents in a natural biome of Argentina, namely, the Chaco phytogeographic domain (5,6), our results indicate that C. burnetii is established in this part of the country where ticks possibly play an essential role in the enzootic cycle. Serologic evidence of C. burnetii infection has been found among goats and cattle in several areas of Argentina (3,4). Because free-ranging domestic cattle and goats are considered among the most likely hosts for A. parvum adult ticks in the Chaco domain (6), humans are likely being exposed to C. burnetii as well.