During the spring and summer seasons of 2004 and 2005, rodents were captured with baited snap traps at 5 sites (Figure): Inner Mongolia Autonomous Region, Heilongjiang Province, Jilin Province, Xinjiang Autonomous Region, and Zhejiang Province. The first 3 sites are forested highlands in northeastern China. The study site in Xinjiang Autonomous Region is grassland in northwestern China. The study site in Zhejiang Province is wooded foothills in southern China.

After the species of the captured rodents was determined, a small piece of the spleen (500 mg) from each animal was used to extract DNA. Briefly, each spleen specimen was crushed with Trizol (Invitrogen, Carlsbad, CA, USA) to separate DNA from RNA after centrifugation, according to the manufacturer's instructions. Nested polymerase chain reaction (PCR) that targeted the FopA gene of F. tularensis was first performed as described previously (8). To further identify the agents in the samples by nested PCR and determine their genotype, PCR using the primer pair of C1 and C4 was performed to amplify the ppI-helicase region of F. tularensis gene structure (9,10). Then the products were gel purified by using Qiaquick Gel Extraction Kit (Qiagen, Hilden, Germany) and sequenced with an automated DNA sequencer (ABI Prism 377, Perkin-Elmer, Foster City, CA, USA). To minimize contamination, DNA extraction, reagent setup, amplification, and agarose gel electrophoresis were performed in separate rooms.

Of 420 rodents tested by nested PCR, 20 were positive for the FopA gene of F. tularensis. The overall infection rate was 4.76% with a 95% confidence interval (CI) of 2.72%–6.78%. A total of 14 species of rodents were identified in the study. Seven species, including Apodemus peninsulae, A. agrarius, Cricetus migratorius, C. triton, Eutamias sibiricus, Meriones lybicus, and Clethrionomys rufocanus, were positive for F. tularensis (Table). Although E. sibiricus and M. lybicus were most likely to be infected, with positive rates of 25% and 22.2%, respectively, no significant difference in infection rate was found among the 7 positive rodent species (χ² = 11.82, degrees of freedom [df] = 6, p = 0.066).

The prevalence of F. tularensis in rodents varied with the geographic origin (Table). The infection rate was highest in rodents in Jilin Province (11.65%), followed by Xinjiang Autonomous Region (10.00%), Heilongjiang Province (6.54%), and Inner Mongolia Autonomous Region (1.76%). No F. tularensis infection was found in rodents collected from Zhejiang Province. The geographic difference in infection rate was significant (χ² = 20.91, df = 4, p = 0.0003). PCR assay, targeting the ppI-helicase region in combination with sequence analysis, identified F. tularensis in rodents in China as type B. The nucleotide sequences of the 20 positive specimens were identical to the published sequences of F. tularensis subsp. holarctica strain (GenBank accession no. AF247642.2).

This study is the first PCR-based study on F. tularensis infection in rodents in China. Heilongjiang Province, Inner Mongolia, and Xinjiang Autonomous Regions were recognized as the natural foci of tularemia ≈40 years ago. The high prevalence of F. tularensis infection in rodents indicates that tularemia natural foci still exist in these areas. F. tularensis was detected for the first time in Jilin Province, which borders Heilongjiang Province and had similar landscape characteristics. Whether this newly described natural focus in Jilin Province is associated with human infection should be further investigated. In any case, this finding demonstrates that F. tularensis seems to be distributed more widely in China than expected. The extensive presence of F. tularensis indicates a potential threat to human health. The fact that rodents trapped from Zhejiang Province (southern China) were negative for the bacterium verifies our belief that F. tularensis is present only in northern China. The geographic variation in infection rate is likely attributable to the difference in biologic characteristics of each study site or a selection bias. Further studies are needed to clarify this question.

Rodents are efficient natural reservoirs for F. tularensis (2,11). In recent years, reports on human tularemia transmitted from small animals have been increasing (12). In this study, several rodent species were found to harbor F. tularensis, but which species are the main hosts in China is still unknown because no significant difference in infection rate was observed among rodents, regardless of their geographic origin. Systematic epidemiologic studies are required to investigate characteristics of natural foci and the role of both wild and domestic animals in transmission of F. tularensis to humans.