In Nara, Japan, to prevent injuries to park visitors, the Foundation for the Protection of Sika Deer in Nara Park captured 97 pregnant female sika deer in April 2005 and 20 male sika deer with antlers in November 2005. Blood specimens were collected from all 117 deer, and buffy coat fractions were stored at –80°C for further analysis. Genomic DNA was extracted from the buffy coat fractions by using a QIAamp tissue kit (QIAGEN, Valencia, CA, USA). Nested PCR amplification of genomic DNA was performed by using primer pairs designed to amplify the E. chaffeensis 16S ribosomal RNA (rRNA) gene and the E. chaffeensis groEL gene (Table).

Of the 117 specimens, 36 (31%) yielded E. chaffeensis 16S rRNA amplification products and 35 (30%) yielded E. chaffeensis groEL amplification products. Of 36 16S rRNA–positive specimens, 33 were positive for groEL (92% concordance rate). Of 35 groEL-positive specimens, 33 were positive for 16S rRNA (94% concordance rate). The E. chaffeensis sequences were conserved in that all sequences obtained from sika deer in Nara were nearly identical to those of a representative strain that we named NS101 and submitted to GenBank (accession no. AB454074). The sequence of the 16S rRNA gene from E. chaffeensis NS101 was most closely related (99.6% identity; 5 bases of 1,333 bp that can be aligned for comparison differed) to that of 5 human isolates: Arkansas (GenBank accession no. M73222) (8), Sapulpa (U60476) (9), 91HE17 (U23503) (10), St. Vincent (U86665) (11), and Jax (U86664) (11); the next closest sequence was from an E. chaffeensis isolate from A. testudinarium ticks in China (GenBank accession no. AF147752) (99.2% identity; 10 bases of 1,333 bp that can be aligned for comparison differed).

Figure 1

Figure 1. Phylogenetic relationship between Ehrlichia chaffeensis NS101 (in boldface) and other Ehrlichia spp. 16S rRNA gene sequences. GenBank accession numbers are shown in parentheses. Numbers above internal nodes indicate the number of...

When E. chaffeensis NS101 was compared with Ehrlichia spp. previously identified in Japan, the sequence of the 16S rRNA gene (a 1,328-bp segment that can be aligned for comparison) of E. chaffeensis NS101 was 98.9%, 98.9%, and 98.6% identical to that of E. muris AS145 strain, Ehrlichia sp. HF565 (the HF strain, Ixodes ovatus Ehrlichia) and Candidatus Ehrlichia shimanensis TS37, respectively. Phylogenic analysis concurred with the observation that E. chaffeensis from Nara sika deer has the highest identity to the E. chaffeensis human isolates from the United States (Figure 1).

Although longer 16S rRNA gene sequences are desirable for strain comparison, the sequence of the 16S rRNA gene of E. chaffeensis from H. longicornis ticks in Korea (GenBank accession no. AY350424, 390 bp) and from A. parvum ticks in Argentina (GenBank accession no. EU826516, 470 bp) were identical to a corresponding, but incomplete, segment (358 bp and 402 bp, respectively) of the 16S rRNA gene of E. chaffeensis NS101 and Arkansas strains. An Ehrlichia sp. was detected in blood samples from marsh deer (Blastocerus dichotomus) captured near the Parana River in southeast Brazil in 1998 (12). However, the 16S rRNA sequence from these marsh deer (GenBank accession no. DQ345720, 383 bp) shares 98.7% identity to a corresponding, but incomplete, segment (381 bp) of the E. chaffeensis NS101 and Arkansas strains. Because the sequences of the corresponding segments of Ehrlichia sp. HF565 and ‘Candidatus Ehrlichia shimanensis’ TS37 both had 99.0% identity to that of E. chaffeensis Arkansas strain, higher than the 98.7%, the Ehrlichia sp. in marsh deer from Brazil might not be E. chaffeensis.

Figure 2

Figure 2. Phylogenetic relationship between the Ehrlichia chaffeensis NS101 groEL sequence (1,208 bp) (in boldface) and other Ehrlichia spp. groEL sequences. GenBank accession numbers are shown in parentheses. Numbers above internal nodes indicate...

Among the E. chaffeensis groEL sequences available in current databases, only that from E. chaffeensis Arkansas has >1,000 bp for reliable comparison. The groEL DNA sequence (1,208 bp that can be aligned) of the NS101 strain (GenBank accession no. AB454077, 1,311 bp) was most closely related to that of E. chaffeensis Arkansas (GenBank accession no. L10917), followed by the Ehrlichia sp. HF565 strain (GenBank accession no. AB032712), and the E. muris AS145 strain (GenBank accession no. AF210459) (Figure 2). The deduced E. chaffeensis NS101 GroEL amino acid sequence (402 residues) was most closely related (99.5% identity) to that of E. chaffeensis Arkansas, followed by the Ehrlichia sp. HF565 strain (99.2% identity) and E. muris AS145 strain (99.0%).

On the basis of the long 16S rRNA and the groEL DNA sequences, our study demonstrates the presence of E. chaffeensis in sika deer from Nara, Japan. The genetic similarity of E. chaffeensis in the sika deer in Japan to strains isolated from HME patients in the United States raises the possibility that HME may exist in Japan. Of 1,803 serum samples collected from persons in metropolitan Tokyo from 1991 through 1995, when E. muris was used as antigen, 20 were seropositive (13). Because E. chaffeensis and E. muris antigens are highly cross-reactive (14), some of these persons might have been infected with E. chaffeensis. Of the 10 tick species found on sika deer (15), the primary tick species found on sika deer in Nara is H. longicornis, which is known to bite humans in Japan and to be infected with E. chaffeensis in South Korea (5). Because sika deer are abundant and increasing throughout Japan (16), this finding highlights need to survey sika deer and humans in Japan for E. chaffeensis infection.