Samples of X. cheopis were collected from 39 live-captured, peridomestic rodents and shrews from 3 localities in Malang, East Java, Indonesia, during an epidemiologic study conducted in 1994 (11). In this study the fleas were reidentified by using morphologic criteria, stored in fresh 70% ethanol, and subsequently evaluated for the presence of rickettsial DNA. DNA sample preparations were derived from triturates of 103 individual fleas in 100 μL PrepManUltra sample preparation reagent (Applied Biosystems, Foster City, CA, USA). DNA preparations of 1 to 5 fleas collected from the same rodent were pooled for testing. Reaction mixtures for the quantitative real-time PCR (qPCR) assays had a total volume of 25 μL and contained 3 μL DNA template. The master mixes were prepared for the 17-kDa Rickettsia–, R. typhi– and R. felis–specific qPCR assays in a separate, clean (DNA-free) room as previously described (6,12). The primer and probe sequences for the 17-kDa Rickettsia-specific and R. felis–specific assays have been reported (6,11). The R. typhi forward (Rt557F: 5´-TGG TAT TAC TGC TCA ACA AGC T-3´) and reverse (Rt678R: 5´-CAG TAA AGT CTA TTG ATC CTA CAC C-3´) primers and probe (Rt640BP: 5´-TET-CGC GAT CGT TAA TAG CAG CAC CAG CAT TAT CGC G-DABCYL-3´) sequences are listed here. Included in each run were 3 negative controls (GIBCO Ultrapure DNA-free distilled water, Invitrogen Corporation, Grand Island, NY, USA), 1 produced in the clean room and 2 in a biosafety cabinet in another laboratory where DNA templates were added. A TOPO TA plasmid (Invitrogen Corporation) that contained the target sequence at 10³ copies for each assay was used as a positive control. qPCR reactions were incubated in the SmartCycler (Cepheid, Sunnyvale, CA, USA) at 94°C for 2 min, followed by 50 cycles of a 2-step amplification protocol of 94°C for 5 s and 60°C for 30 s. Fluorescence was monitored during the annealing step of each cycle, and data were analyzed with SmartCycler software version 2.0c (Cepheid).

Rickettsial DNA was detected by 17-kDa qPCR in 7 of 39 pools containing 1–5 X. cheopis fleas. To determine whether R. typhi or R. felis infected these fleas, PCR assays specific for R. typhi and R. felis ompB partial sequence targets were performed (Table). Results of these assays showed that 5 of the 7 Rickettsia-positive X. cheopis fleas were infected with R. typhi, and 2 were positive for R. felis. The remainder of the 32 pools and all nontemplate controls were negative for R. typhi and R. felis. Additionally, 15 Rickettsia-free C. felis fleas (Heska Corporation, Loveland, CO, USA), evaluated at the same time and under the same conditions as the Malang fleas, were negative for R. typhi and R. felis.

To determine the identity of rickettsial agents infecting X. cheopis fleas collected from rodents and shrews in Malang, we assessed pools of 1 to 5 fleas from each animal. Our results confirm R. typhi in a known flea vector of murine typhus in a highly disease-endemic region of East Java, Indonesia (11,13). R. felis has been shown to infect fleas of peridomestic rodents (7,8) and fleas other than C. felis (1,5,14). However our report is the first of R. felis naturally infecting X. cheopis fleas, a vector of plague and murine typhus. Both R. felis–containing flea pools were derived from Rattus rattus, 1 from the suburban and 1 from the urban neighborhoods of Malang. R. rattus was the predominant species captured in urban and suburban environments (72%) and appears to be the primary host for R. felis– and R. typhi– infected X. cheopis. In the rural setting, where R. rattus was represented with far less frequency (14.3%), neither rickettsial agent was detected in collected fleas. These findings merit further epidemiologic investigation to better understand the relationship between R. felis, R. typhi, and X. cheopis and the transmission dynamics between flea and rodent.

Additionally, this report provides the first evidence of R. felis in the Indonesian archipelago. Investigations of rickettsial agents in Indonesia have been relatively few; to date, human infections with R. felis have not been reported from Indonesia. The lack of reports may be because a murine typhus–like disease associated with R. felis infection would not allow healthcare providers to clinically discriminate fleaborne spotted fever from murine typhus or other rickettsioses. In Indonesia, rickettsioses and typhoid fever are collectively referred to as tifus. Rickettsial tifus can be discerned by serologic tests or by observing when rickettsial tifus cases rapidly respond to treatment with a tetracycline or chloramphenicol. Furthermore, the inability to diagnose fleaborne spotted fever by laboratory means has been attributed to the cross-reactivity of antibodies to R. felis antigens with other rickettsial antigens (1). Consequently, serologic assays have been unable to differentiate fleaborne spotted fever from other rickettsioses. Thus, the high prevalence of murine typhus reported in Indonesia likely also includes fleaborne spotted fever. In addition, previously demonstrated serologic evidence of spotted fever group rickettsiae infection among residents of Gag Island, in eastern Indonesia (15), could have been due to R. felis. On the basis of data presented here and of recent reports of R. felis in other countries in Asia (2–5,8,9), healthcare providers in Indonesia should be alerted to the possibility of fleaborne spotted fever among their patients.