Previous work described an enzootic cycle of Borrelia burgdorferi sensu lato (hereafter referred to as B. burgdorferi) maintained by the rodent Neotoma mexicana and the tick Ixodes spinipalpis in northern Colorado. We investigated the incidence of coinfection among rodents with the agent of human granulocytic ehrlichiosis (aoHGE). aoHGE was detected in 23.5% of 119 rodent spleens examined. Biopsy results indicated that 78 (65.5%) of the 119 rodents were positive for B. burgdorferi, whereas 22 (78.5%) of the 28 animals that harbored aoHGE were also infected with B. burgdorferi. In 14 of 25 I. spinipalpis tick pools, aoHGE was detected by amplifying both the 16s rRNA and p44 gene of aoHGE. The ability of I. spinipalpis to transmit aoHGE was examined in C3H/HeJ mice. aoHGE was detected in their blood 5 days after I. spinipalpis infestation. This study confirms that both B. burgdorferi and aoHGE can be transmitted by I. spinipalpis ticks and that there is a high incidence of coinfection in rodents, predominantly Peromyscus maniculatus and N. mexicana, that inhabit the foothills of northern Colorado.
Since its first description in 1994 , human granulocytic ehrlichiosis (HGE) has been documented as an emerging tickborne illness in the United States. There have been >400 cases, with the highest incidence in the upper northeastern and upper midwestern portions of the United States . This disease incidence corresponds to the geographic distribution of the purported vector, Ixodes scapularis , whereas a small foci of human infection associated with I. pacificus has been reported in northern California .
Serosurveys indicate that Peromyscus leucopus is the main rodent reservoir for these agents in endemic areas [5–7]; however, in areas of the western United States other rodent species, predominantly Neotoma fuscipes and N. mexicana, are seropositive for the agent of HGE (aoHGE) and for Borrelia burgdorferi sensu lato (hereafter referred to as B. burgdorferi) . Likewise, I. spinipalpis is a competent vector of Lyme disease in enzootic cycles that involve Peromyscus species and N. fuscipes in California  and N. mexicana in Colorado . The incidence of coinfection with aoHGE has not been determined in these western habitats.
The recent discovery of an enzootic cycle of B. burgdorferi maintained by N. mexicana and I. spinipalpis and the speculation that an enzootic cycle of transmission of aoHGE exists in California  led us to investigate the possibility that a similar enzootic cycle exists that involves the transmission of aoHGE in Colorado.
Materials and Methods
Collection of vertebrates and ticks
Rodents were captured on the western edge of the Foothills Campus of Colorado State University, Fort Collins, Colorado, from February to October 1999. Captured animals were examined to identify individual species and then held over water so that replete ticks could be collected. Engorged ticks were identified by use of standard taxonomic keys, were held at 22°C, and were allowed to molt to the next stage. Skin biopsy specimens were obtained from the rodents for B. burgdorferi analysis, and spleens were harvested for determination of aoHGE infection.
Transmission of aoHGE by nymphal I. spinipalpis
Eight I. spinipalpis nymphs, which were originally collected as larvae from N. mexicana in the spring of 1999, were placed on individual C3H/HeJ mice (Jackson Laboratory, Bar Harbor, ME). Nymphal ticks were allowed to feed to repletion, and blood was collected from mice into heparinized tubes 5 and 10 days after tick feeding, to test for transmission of aoHGE.
Isolation of DNA from spleen and blood of rodents and from I. spinipalpis nymphs
For isolation of rodent DNA, individual spleens were teased apart, and DNA was extracted using a QIAamp DNA mini kit (Qiagen, Valencia, CA). Heparinized blood (200 µL) was spun at 450 g for 3 min, and the whole-cell pellet was subjected to DNA extraction, using a QIAamp DNA blood mini kit (Qiagen). For isolation of tick DNA, 125 nymphal ticks were pooled in groups of 5 within individual plastic bags, were macerated between metal plates, and were homogenized by adding 1 mL of DNA STAT-60 (Tel-Test, Friendswood, TX). The tick homogenate was incubated with chloroform, and DNA was precipitated with 100% isopropanol. All tick DNA extracts were tested for the presence of tick mtDNA .
Polymerase chain reaction (PCR) to amplify aoHGE
All rodent spleens and tick pools were tested for aoHGE by PCR, using primers Ehr 521 and Ehr 747 . Blood collected in experimental transmission studies, as well as spleen and tick DNA extracts, were also tested using primers MSP3F 5′-CCAGCGTTTAGCAAGATAAGAG-3′ and MSP3R 5′-GCCCAGTAACAACATCATAAGC-3′, which are specific for a 334-bp portion of the p44 gene of aoHGE. To amplify aoHGE-specific DNA by using the MSP3F and MSP3R primers, we added DNA (100 ng) separately to a 50-mL reaction volume containing 100 pmol each of primers MSP3F and MSP3R and a master mix described elsewhere . DNA was amplified by use of a 3-step cycling program set for an initial 4-min denaturation at 94°C, 40 cycles of denaturation at 94°C for 30 s, a 30-s annealing at 55°C, and a 1-min extension at 72°C.
Amplification products were visualized on ethidium bromide—stained 3% agarose gels. All positive amplicons were sequenced by adding 50 ng of purified amplicon and 3.2 pmol of primer to a dye-terminator cycle sequencing reaction (Perkin-Elmer Applied Biosystems, Foster City, CA) with AmpliTaq FS DNA polymerase (Perkin-Elmer). All PCR reactions were set up in a dedicated PCR site where aoHGE had never been handled; templates and products were handled at second and third sites, respectively, where aoHGE had never been investigated.
Culture of B. burgdorferi
Rodent ear biopsies were obtained by sterile surgical technique and cultured in BSK II medium (Sigma, St. Louis) . Cultures were examined weekly by dark-field microscopy for the presence of spirochetes.
Rates of single infection and coinfection were analyzed by use of Fisher's exact test, with P < .05 considered statistically significant.
Incidence of infection with aoHGE and B. burgdorferi in northern Colorado
A total of 119 rodents were captured between February and October 1999 (table 1). All spleens were examined by PCR for aoHGE, and the overall infection rate was 23.5% (28/119 were positive). In terms of B. burgdorferi infection, 65.5% of rodent biopsies were culture positive. Isolates of B. burgdorferi from this region had been characterized elsewhere as B. burgdorferi sensu lato , and no further characterization of B. burgdorferi was done in these studies. Of the 28 rodents that tested positive for aoHGE by PCR, 22 (78.5%) also were infected with B. burgdorferi. No significant difference in the rates of single infection with either aoHGE (P = .74) or B. burgdorferi (P = .11) was noted between N. mexicana, P. maniculatus, or Microtus ochrogaster.
The predominant rodent species captured (table 1) was P. maniculatus (n = 63); 13 (20.6%) had spleens that tested positive for aoHGE by PCR, and 9 (69.2%) of those 13 mice were coinfected with B. burgdorferi and aoHGE. Of 36 N. mexicana that were captured, 14 (38.8%) had spleens that tested positive for aoHGE by PCR. A significantly greater percentage of N. mexicana (33.3%) than P. maniculatus (14.2%, P < .04) were coinfected with aoHGE and B. burgdorferi. Five P. difficilis were captured, and only 1 tested positive for B. burgdorferi; no aoHGE was amplified from the spleens of these animals. The fourth species captured was M. ochrogaster (n = 15), of which 13 (86.6%) were infected with B. burgdorferi and only 1 (6.6%) was coinfected with aoHGE.
Incidence of aoHGE in I. spinipalpis nymphal ticks
A total of 25 I. spinipalpis nymphal tick pools were collected and analyzed by PCR. All 25 samples contained tick mtDNA. Both the 16s rRNA and p44 genes of aoHGE were analyzed for each tick pool, and 14 (56%) of the 25 were positive by PCR for one or both of these genes (figure 1A). All tick pools that were positive by PCR for the 16s rRNA gene were confirmed positive by amplification of the p44 gene. All 16s rRNA amplicons were then sequenced and found to be identical to the sequence of aoHGE (GenBank accession number U02521) .
Transmission of aoHGE by I. spinipalpis nymphal ticks
To confirm that I. spinipalpis nymphs were competent vectors for aoHGE, we placed 8 nymphs on each specific pathogen—free mouse (n = 4) and monitored them for their ability to transmit aoHGE. Five days after tick feeding, the p44 gene could be amplified from the peripheral blood of 3 of the 4 mice (figure 1B, lanes 2–5). By day 10 after tick feeding, no aoHGE could be detected in the peripheral blood of these mice (figure 1 B, lanes 7–10). Sequence analysis corroborated that aoHGE was transmitted and could be amplified from the peripheral blood of C3H/HeJ mice.
To our knowledge, this is the first report of an enzootic cycle of dual infection involving B. burgdorferi and aoHGE in the western United States. Our results demonstrate that P. maniculatus and N. mexicana are major reservoirs for dual infection in Colorado. The incidence of infection with B. burgdorferi and aoHGE (69.2%) in P. maniculatus was similar to that reported in a highly endemic area of Connecticut, where rates of dual infection in P. leucopus were 53.8%–68.2% .
As reported here, the overall infection rate for aoHGE (38.8%) was similar to that reported in the dusky-footed wood-rat (N. fuscipes; 34%) in California , but it was higher than the rate of infection reported among Peromyscus species in Colorado (20.6%), in California (8%) , or in an endemic area of Minnesota (11.4%) . Our data indicate that both N. mexicana and P. maniculatus are reservoirs for this infection in Colorado; however, in California, only N. fuscipes was infected, and only 1 P. maniculatus was seropositive . Of interest is the fact that aoHGE could not be amplified from P. maniculatus in these California studies , yet a significant rate of infection was found among P. maniculatus in Colorado. Whether this was a true difference in California P. maniculatus or just a reflection of the difference in sampling techniques (peripheral blood vs. spleen) remains to be determined.
Although the prairie vole (M. ochrogaster) in Colorado appeared to harbor the highest incidence of B. burgdorferi infection (86.6%), this was not statistically significant (P > .05). Only 1 of 15 animals tested was also infected with aoHGE, which possibly indicates that this species may be refractory to aoHGE infection, unless previously or simultaneously infected with B. burgdorferi. Although M. pennsylvanicus has been shown to be a competent reservoir for B. burgdorferi , to our knowledge this is the first report of M. ochrogaster as a reservoir species for B. burgdorferi infection.
Previous studies have suggested that I. spinipalpis ticks may be a western vector of aoHGE ; however, the studies presented here are the first documentation that I. spinipalpis ticks harbor and can transmit aoHGE effectively. The only tick species recovered from rodents captured in the foothills of Colorado was I. spinipalpis. Our results indicate that ticks derived from N. mexicana were positive by PCR for both the 16s rRNA and p44 genes of aoHGE. Likewise, sequence analysis of 16s rRNA amplicons indicated a 100% sequence homology to aoHGE . Moreover, this is the first study to demonstrate that I. spinipalpis can transmit aoHGE to specific pathogen—free mice in the laboratory. We successfully amplified aoHGE from the peripheral blood of 3 of 4 mice exposed to I. spinipalpis nymphs 5 days after tick feeding. The likelihood that coinfection in this part of Colorado is maintained by a second tick species, Dermacentor andersoni, is unlikely, because these ticks have not been shown to transmit B. burgdorferi. Similarly, D. variabilis was incapable of transmitting aoHGE in the laboratory .
Because I. spinipalpis has been described as nidicolous, the significance to human health of I. spinipalpis transmitting and maintaining cycles of B. burgdorferi and aoHGE infections in rodent reservoirs is unclear. There are several reports of I. spinipalpis found on humans in geographic regions as diverse as New Mexico (Centers for Disease Control and Prevention, unpublished data) and Oregon and Canada . Likewise, the low infection rate of I. pacificus in areas of California with reported equine ehrlichiosis and a high prevalence of infection in N. fuscipes  suggests that I. spinipalpis may be a bridging vector of human ehrlichiosis in California. Moreover, it has been hypothesized that I. spinipalpis, which has a greater infection rate than does I. pacificus for B. burgdorferi, may pose a greater risk to humans in the Pacific Northwest , where questing behavior of these ticks outside of rodent nests has been recorded .
This study confirms that N. mexicana, P. maniculatus, and M. ochrogaster are the principal rodent reservoirs for B. burgdorferi in northeastern Colorado. Likewise, both the deer mouse (P. maniculatus) and the Mexican woodrat (N. mexicana) displayed rates of infection with aoHGE and B. burgdorferi similar to those demonstrated in endemic areas of Connecticut . Although the incidence of human infection with B. burgdorferi in the southwestern United States is minimal and no human cases of HGE have been confirmed in this geographic region, future research should be oriented toward surveillance studies that adequately assess the risk of I. spinipalpis ticks to transmit B. burgdorferi or aoHGE to humans.
We thank Bradley Schneider for expert technical assistance in purifying DNA and for performing polymerase chain reactions (PCRs) and Christine Happ for culturing and monitoring rodent skin biopsy samples (Bacterial Zoonosis Branch, Centers for Disease Control and Prevention [CDC], Atlanta) and Robert Lanciotti (Arboviral Diseases Branch, CDC) for helpful discussions regarding PCR and sequencing techniques.