RESEARCH Anaplasma phagocytophilum from Rodents and Sheep, China

To characterize the strains of Anaplasma phagocytophilum in wild and domestic animals in China, we isolated the organism from rodents and sheep in northeastern China. We isolated 3 strains (2 from rodents and 1 from sick sheep) through propagation in BALB/c mice and then cell culture in HL60 cells. The 3 isolates were identified by Wright-Giemsa staining, immunofluorescence, and electronic microscopy and were characterized by sequence analyses of the 16S rRNA gene, partial citrate synthase gene, major surface protein 4 gene, and heat shock protein gene. The multiple sequences of the 3 isolates were identical to each other but different from all known strains from other countries. The public health and veterinary relevance of the isolates deserves further investigation.

In May 2009, live rodents were captured in wire mesh traps in the hinterland of the Changbai Mountains (42°45′N, 130°35′E) in Jilin Province, China, where natural infections with A. phagocytophilum in ticks and rodents have been reported (8). After their species and sex were identifi ed, trapped rodents were euthanized and anatomized. The spleen was removed from each rodent and ground with sterile normal saline. Four dying sheep were found at the same site at the same time; blood samples were aseptically collected into tubes containing EDTA-K 2+ .

Propagation of A. phagocytophilum in BALB/c Mice
For isolation of A. phagocytophilum, the spleen suspensions of the rodents were pooled into 12 groups according to species, and 0.3 mL of spleen suspension was intraperitoneally injected into 48 BALB/c mice (4 in each group). Blood samples from the 4 sheep were also pooled and injected into a group of BALB/c mice by the same means.
After 7-14 days, blood samples were collected from each inoculated mouse and evaluated for infection by real-time PCR. All animal experiments were performed according to the approved Institutional Animal Care and Use Committee guidelines.

Wright-Giemsa Staining and Immunofl uorescence Microscopy
Slides of peripheral blood or the cultured cells were stained with Wright-Giemsa (BaSO DIAGNOSTICS, INC, Zhuhai, China). An indirect immunofl uorescence assay was performed after the slides of culture cells were fi xed for 10 minutes in a 1:1 solution of methanol and acetone as described (13). Horse anti-A. phagocytophilum serum (kindly provided by Jenet E. Foley, University of California, Davis, CA, USA) and fl uorescein isothiocyanateconjugated goat antihorse immunoglobulin G (Zhongshan Biotechnique, Inc., Beijing, China) were used for the assay. Serum samples from healthy horses were used as negative controls.

Electronic Microscopy
Infected HL60 cells were processed as previously described (14). Electron microscopic examination was conducted by using a Tecnai 10 electron microscope (Philips, Amsterdam, the Netherlands).

PCR and Sequence Analysis
Real-time PCR selective for the major surface protein 2 gene (msp2) was used as described by Drazenovich et al. (15). To characterize the A. phagocytophilum strains isolated in the study, we amplifi ed, purifi ed, sequenced, and compared the 16S rRNA gene (rrs), partial sequences of the citrate synthase gene (gltA), major surface protein 4 gene (msp4), and heat shock protein gene (groEL) as described (8,16,17). Phylogenetic analyses were performed and phylogenetic trees were constructed by using Mega 3.0 software (17,18).

A. phagocytophilum in BALB/c mice
A total of 47 live rodents-20 black-striped fi eld mice (Apodemus agrarius) and 27 great long-tailed hamsters (Tscherskia triton)-were captured. When tested 7-14 days postinoculation, every mouse in 5 of the 12 groups of inoculated BALB/c mice was positive for A. phagocytophilum according to real-time PCR selective for the msp2 gene; 3 groups were A. agrarius mice and 2 were T. triton hamsters. Two BALB/c mice inoculated with the anticoagulated blood samples from the 4 sick sheep were positive for A. phagocytophilum according to PCR. Typical morulae were observed in granulocytes of experimentally infected BALB/c mice ( Figure 1, panel A).

A. phagocytophilum in HL60 cells
Three A. phagocytophilum strains were propagated in HL60 cells: 1 from A. agrarius mice was named China-C-Aa, 1 from T. triton hamsters was named China-C-Tt, and 1 from sheep was named China-C-Y. A. phagocytophilum was fi rst observed in Wright-Giemsa stain preparations 5 days after preparation of cultures ( Figure 1

A. phagocytophilum Isolate Sequences
The 1,431-bp nearly entire rrs sequences of the 3 A. phagocytophilum isolates from cultured cells were identical to each other and to the sequences amplifi ed from infected mice as well as from fi eld-collected rodents and sheep. The tested rrs sequences were also identical to sequences amplifi ed from ticks and rodents captured 3 years ago (GenBank accession nos. DQ342324 and DQ449948) in the same area (8) but different from all known A. phagocytophilum sequences deposited in GenBank.
Analysis of the partial sequences of gltA (348 bp), msp4 (779 bp), and groESL (428 bp) genes showed that the nucleotide sequences of gltA fragments amplifi ed from the 3 isolates were identical to each other and showed 84% -99% identity with previously reported A. phagocytophilum strains, with 3-52 bp differences and 83%-99% similarity of deduced amino acid sequences. Three clades were structured on a phylogenetic tree based on 348-bp nt of the gltA gene, including a clade of strains from the United States, the Russian Far East, and this study; a clade comprising strains from rodents in southeastern China; and a clade of other Anaplasma spp., such as A. centrale, A. marginale, and A. platys (Figure 2).
The sequences of 779-bp msp4 fragments amplifi ed from the 3 isolates were also 100% identical and had 98%-87% nt sequence identity and 99%-88% deduced 268-aa sequence identity compared with A. phagocytophilum strains available in GenBank. When compared with the sequences from rodents in southeastern China (GenBank accession no. EU008082), nucleotide identity was only 87% with a 95-bp difference, and induced amino acid identity was 88% with a 31-aa difference. Phylogenetic analysis placed the A. phagocytophilum isolates in this study on a separate branch and in the same clade as the strains from the United States and Europe but far from the strains from sheep in Norway (GenBank accession no. AY706391), mule deer in Montana (DQ674249), and rodents in southeastern China (EU008082) (Figure 3). The other Anaplasma spp. were in a separate clade. 766 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 16  When the 428-bp groEL sequences of the 3 isolates were compared with known A. phagocytophilum sequences in GenBank, the identity varied from 93% to 99%. The phylogenetic tree of groEL showed the A. phagocytophilum isolates in this study on a separate branch. The strains from humans in China and the United States, horses in United States, dogs in Slovenia, roe deer in Poland and Austria, and ticks in Germany were in another clade ( Figure 4); however, their deduced amino acid sequences were identical to those from patients and rodents in southeastern China.

Discussion
We isolated 3 strains of A. phagocytophilum from black-striped fi eld mice, great long-tailed hamsters, and sheep in northeastern China. The availability of the isolates in a cell line will permit studies on the genetic, proteomic, and pathogenic characteristics of this agent.
A. phagocytophilum is reportedly maintained in various animal reservoirs, such as white-footed mice (19), woodrats (7), goats, sheep, and horses (5,20). Our isolation of 3 A. phagocytophilum strains from A. agrarius and T. triton rodents and from sheep indicates that both small wild animals and domestic animals may act as competent reservoirs of A. phagocytophilum in northeastern China. Although we found cultivation of this organism from experimentally infected mice to be reliable, the sensitivity of cultivation from wild and domestic animals is uncertain. In addition, the specimens used for isolation were pooled. Consequently, we were unable to ascertain the exact prevalence of infection in the rodents collected for this study. In a previous survey, we found a natural infection rate of 8.8% for A. phagocytophilum in rodents in the same area (8). To determine the level of infectivity in rodents as well as domestic animals, further studies are needed.
The nucleotide sequences of the 3 strains in this study were identical to each other in corresponding genes. The 1,431-bp nearly entire rrs sequences were most closely related to those detected in rodents from southeastern China (8,9), but they differed from other known strains. The sequence divergences and the phylogenetic analyses of partial gltA, msp4, and groESL genes indicated that a novel strain of A. phagocytophilum might be prevalent in northeastern China.
Different A. phagocytophilum strains seem to have special host tropisms (21). Strains from sciurids and whitefooted mice infect various laboratory animals and perhaps humans as well. A. phagocytophilum-variant 1 and the strains from woodrats are found in association with wildlife only; human infections with these strains have yet to be identifi ed. A. phagocytophilum-variant 1 has been unable to infect white-footed mice or SCID (severe combined immunodefi ciency) mice but could infect goats by experimental inoculation (22). Holden et al. have documented that  the pathogenicity of an A. phagocytophilum strain causing human disease waned with mouse passage in C3H mice but could be resurrected by passage in SCID mice (23). In our study, A. phagocytophilum strains with the same molecular characteristics were isolated not only from wild rodents but also from domestic sheep. Furthermore, they could propagate in BALB/c mice in the laboratory. The host tropisms and pathogenicity of the isolates remain to be clarifi ed, and the relevance of these fi ndings to public health and veterinary medicine deserves further investigation.