Multispacer Typing of Bartonella henselae Isolates from Humans and Cats, Japan

To determine genotypic distribution of and relationship between human and cat strains of Bartonella henselae, we characterized 56 specimens using multispacer typing (MST). Of 13 MST genotypes identified, 12 were grouped into cluster 1. In Japan, human infections can be caused by B. henselae strains in cluster 1.


The Study
During 1997 through 2008, we collected 56 B. henselae specimens from western Japan, mainly from Yamaguchi prefecture; the specimens included 1 B. henselae isolate from a patient with endocarditis (7), 24 clinical specimens from CSD patients who had test results positive for B. henselae DNA, and 31 B. henselae isolates from domestic cats (8). The 24 clinical specimens included 5 lymph node specimens and 16 pus specimens from patients with typical CSD, 1 blood specimen from a patient with bacteremia, 1 liver specimen from a patient with hepatic granuloma, and 1 spleen specimen from a patient with splenic granuloma. Total genomic DNA was extracted from the specimens by using the QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany). B. henselae DNA was detected by using PCR with specifi c primers for the 16S-23S rRNA intergenic spacer (9) and the htrA gene (10), and the 16S rRNA genotype was confi rmed by partial sequencing of the 16S rRNA gene (11).
Subsequently, we analyzed the phylogenetic relationships of the 7 novel MST genotypes identifi ed in this study with the 50 previously identifi ed genotypes. Multiple sequence alignment of the concatenated spacer sequences was performed by using ClustalW (www.ebi.ac.uk/clustalw). Finally, a phylogenetic tree was constructed by using the unweighted pair-group method with arithmetic mean (UP-GMA) in MEGA4 (12). This phylogenetic tree is grouped into 4 clusters (Figure). Of the 13 MST genotypes identifi ed in this study, 12 genotypes belonged to cluster 1, but one genotype (MST genotype 52) belonged to cluster 4.

Conclusions
This study showed that MST genotypes in Japan were mainly grouped into 1 lineage (cluster 1), which was composed of Asian and American strains of B. henselae, and that the genotypic distribution of human strains coincided with that of cat strains. Although only 1 human strain from the West Indies belonged to cluster 1 before this study (6), we discovered that all 25 of our human strains from Japan were grouped into cluster 1. These results demonstrate that human infections can be caused by B. henselae strains in cluster 1, which differed from clusters corresponding to the Houston-1 and Marseille type strains.
The genotypic distribution of the human strains in this study differed from that reported by Li et al. (6) because their strains isolated in France were grouped under 2 lineages (Houston-1 and Marseille). However, we found that the lineages of human strains matched those of cat strains in each country. These results are consistent with the role of cats as the major reservoir of B. henselae (13).
In this study, we identifi ed 2 cat strains that were classifi ed into cluster 4. These strains belonged to 16S type II, which is rare in Japan (3). In previous MST studies, strains in cluster 4 were isolated from cats and belonged to 16S type II (5,6). Intriguingly, similar lineages consisting of 16S type II isolates from cats were observed in other genotyping studies involving the use of multilocus sequence typing (MLST) (14) and multiple locus variable number tandem repeat analysis (MLVA) (15). Thus, these lineages may be less pathogenic for humans. However, further studies are needed to investigate this hypothesis.
When we characterized the strains in this study by MLST we found that almost all of them shared the same sequence type as Houston-1 (8). In contrast, we identifi ed 13 MST genotypes that belonged to different clusters than Houston-1. The lower resolving power of MLST is mostly likely due to sequence conservation in the 8 housekeeping genes selected for the method. MST has a higher resolving power because the spacers used in this method are more variable than MLST markers. As a result, MST is better suited for evaluating the population structure of closely related B. henselae strains.
We conclude that the MST genotypes in Japan are mainly grouped into cluster 1 and that the genotypic distribution of human strains coincides with that of cat strains. In Japan, human infections can be caused by B. henselae strains in cluster 1, distinct from clusters containing the Houston-1 and Marseille type strains. These results improve our understanding of the population structure of and geographic relationship between human and cat strains of B. henselae.    Search past issues of EID at www.cdc.gov/eid