Ticks (Acari: Ixodida) are unique among blood feeding arthropods in that epithelial cells lining the midgut lumen phagocytose and digest the blood meal intracellularly. That process is presumably significant to the close association of ticks with obligate intracellular bacteria within the order Rickettsiales (Balashov, 1972; Munderloh and Kurtti, 1995). The genus Rickettsia includes spotted fever group (SFG) pathogens of vertebrate and tick hosts such as Rickettsia rickettsii and Rickettsia conorii (Niebylski et al., 1999; Santos et al., 2002), as well as nonpathogenic endosymbionts such as Rickettsia peacockii that are maintained in ticks by transovarial transmission (TOT) and are believed to prevent ovarian superinfection by pathogenic rickettsiae (Burgdorfer et al., 1981; Niebylski et al., 1997; Azad and Beard, 1998; Macaluso et al., 2002; de la Fuente et al., 2003). The physiological mechanisms that mediate rickettsial interactions with arthropod hosts are poorly understood. Rickettsiae are notoriously difficult research subjects due to their small size, slow growth, instability outside host cells and resistance to genetic manipulation. Moreover, the pioneering efforts of early 20^th century researchers to establish tick host models of rickettsial infection dynamics (Harden, 1990) have been only minimally improved upon (Hayes and Burgdorfer, 1989). Recent advances in genetic manipulation of rickettsiae, coupled with development of improved host model systems, should lead to a better understanding of the physiology of rickettsial host range, tissue tropisms, transmission dynamics, pathogenesis versus symbiosis and superinfection resistance in arthropods.

Rickettsiologists have relied on stain-based light microscopy, transmission electron microscopy (TEM) and immunological detection methods to observe rickettsiae within host cells and tissues. Because those techniques detect fixed rickettsiae in fixed cells, they cannot fully define the interactions of rickettsiae and hosts, a task that will ultimately require live tissue techniques. Transformation of rickettsiae to express fluorescent markers will allow study of live rickettsiae within live host tissues, analogous to the advances achieved with such techniques in study of other arthropod-borne pathogens. For example, transformation with variants of the green fluorescent protein (GFP) reporter gene from the jellyfish, Aequorea victoria (Chalfie et al., 1994) allowed characterization of tissue distributions and infection dynamics of GFP-labeled arboviruses in mosquito hosts (Brault et al., 2004; Foy et al., 2004), of Leishmania protozoans in phlebotomine sand flies (Guevara et al., 2001) and of trypanosomes in tsetse flies (Bingle et al., 2001) and triatomine bugs (Guevara et al., 2005). Imaging of the movement of GFP-labeled malarial parasites in live mosquito hosts (Frischknecht et al., 2004; Vlachou et al., 2004) demonstrated the full power of this technology.

In the first reported studies of transgenic fluorescent bacteria in ticks, Ceraul and colleagues (2002) described evidence for encapsulation/nodulation of GFP-labeled Escherichia coli injected into Dermacentor variabilis ticks, while Matsuo and colleagues (2004) described clearance from the midgut of GFP-labeled E. coli ingested by Ornithodoros moubata ticks. However, E. coli, unlike rickettsiae, is not a normal constituent of the tick microflora and does not grow intracellularly. The recent development of electroporation techniques and transposon-based transformation vectors for rickettsiae (Rachek et al., 1998; Qin et al., 2004; Baldridge et al., 2005) has made their genetic manipulation a reality.

In this report, we describe the evaluation and selection of Ixodes scapularis from three North American tick species as a model host for Rickettsia monacensis transformed to express GFP (Baldridge et al., 2005). R. monacensis is a SFG rickettsial endosymbiont isolated from the European sheep tick, Ixodes ricinus (Simser et al., 2002). Phylogenetic analyses have shown that R. monacensis and related rickettsiae detected in I. ricinus populations by molecular means (Sreter-Lancz et al., 2005 and references therein) form a separate cluster within the SFG group distinct from its known pathogenic members (Ishikura et al., 2002, 2003; Kurtti et al., unpublished). I. scapularis is a member of the I. ricinus species complex (Keirans et al., 1999) and has been shown by phylogenetic analyses to very closely related to I. ricinus (Fukunaga et al., 2000; Xu et al., 2003). The two ticks are morphologically and behaviorally similar and, along with I. pacificus and I. persulcatus, are the principal vectors of the Lyme Disease agent Borrelia burgdorferi and related Borrelia in the northern hemisphere (Keirans et al., 1999; Fukunaga et al., 2003).

We infected ticks by capillary feeding rickettsiae to adults and nymphs and by immersion of larvae in a rickettsial suspension. Fluorescent R. monacensis were readily visualized by epifluorescence microscopy (EFM) in dissected, unfixed tick tissues but could not be observed through the strongly autofluorescent tick cuticle. In adult female I. scapularis, but not in Amblyomma americanum or Dermacentor variabilis, R. monacensis disseminated from the gut and infected the salivary glands that mediate transmission to vertebrate hosts. R. monacensis was transstadially transmitted (TST) in I. scapularis and was maintained in apparently healthy adult ticks for up to 6 months. We did not obtain evidence for transovarial transmission (TOT), possibly due to superinfection resistance. R. monacensis infected the tick tracheal system that surrounds and penetrates tissues within the tick hemocoel. The tracheal system could potentially serve as a dissemination pathway and infection reservoir during the tick life cycle. Infection of I. scapularis with GFP-labeled R. monacensis provides a new model system for study of the interactions of a rickettsial endosymbiont with a tick host.

To infect ticks by feeding, drawn-out glass microcapillary pipets loaded with RPMI medium containing Rmona658 at 2 – 6 x 10⁵/μl were placed over hypostomes of ticks immobilized in Petri dishes as described for infection with Borrelia (Korshus et al., 2004). Following 3 – 4 hr feeding at 34 °C, ticks were recovered into shell vials and returned to colony maintenance conditions. To infect larvae by immersion, ticks were placed in L-15B300 medium containing Rmona658 at 8 x 10⁶/μl, vortexed at medium speed for 5 s and incubated at 34 °C for 30 min. The larvae were surface disinfected by immersion in a 0.1% bleach solution for 2 min, washed in distilled H₂O and returned to colony maintenance conditions.

Table 1 PCR and sequencing primers.

Three adult female A. americanum ticks collected as partially engorged ticks from a dog each ingested approximately 1.5 μl of medium containing Rmona658. Live single ticks were dissected and examined at 1, 2 and 3 weeks post-capillary feeding (PCF). At 1 week PCF, fluorescent Rmona658 were abundant throughout the midgut versus an estimated 10 – 100-fold fewer but still widely distributed rickettsiae in 2 week PCF gut. In 3 week PCF gut tissue, rickettsiae were rare and occurred in isolated areas (Fig. 1A and B). Rmona658 were not observed in any other tissue. Four of ten nymphs that had recently molted from larvae engorged with dog blood survived and were examined at 1 week PCF. Rmona658 were visible in midgut tissue only of 1 nymph. The results suggested that Rmona658 replicated poorly within and did not disseminate from the gut of partially fed adult or unfed nymphal A. americanum.

Figure 1 Epifluorescence imaging of GFP transformant Rickettsia monacensis (Rmona658) in Amblyomma americanum, Dermacentor variabilis and Ixodes scapularis adult female ticks. Scale bar at lower left in each image indicates 10 μm and small arrows indicate (more ...)

Of 47 unfed mixed sex adult D. variabilis collected by dragging, 45 survived 2 days PCF. At 1 week PCF, all remaining ticks were alive and of 7 examined, fluorescent Rmona658 that were primarily intracellular were observed in midguts only of 1 female and 1 male. At 2 weeks PCF, 2 ticks were dead and of 13 live ticks examined, 8 females and 2 males contained Rmona658 in midguts only. At 4 weeks PCF, 12 ticks were dead and 6 females and 3 males among the 11 survivors contained Rmona658 in midgut tissue only. We noted extensive tissue damage and presence of many extracellular rickettsiae in midguts of the most heavily infected ticks. In contrast to the Amblyomma ticks, the relative abundance of Rmona658 in Dermacentor midguts increased with time and replicating (doublet forms) as well as dense clusters of rickettsiae were common in the 2 and 4 week PCF ticks (Fig. 1C and D). The results suggested that Rmona658 replicated within and probably damaged the midgut of adult D. variabilis but did not disseminate to other tissues.

Of 15 unfed mixed sex adult I. scapularis collected from deer, 14 survived at two days PCF. At 1 week PCF, 7 of 11 survivors were examined and fluorescent Rmona658 were observed in midguts of 3 males and 3 of 4 females (Fig. 1E). At 2 weeks PCF, midguts of 2 of the 4 remaining females contained abundant Rmona658 that often occurred as replicating forms and in dense clusters (Fig. 1F and H). We observed some particularly dense clusters of Rmona658 in close proximity to tracheae associated with the gut (Fig. 1F). Areas of the gut containing high concentrations of hematin granules from blood ingested and digested during the nymphal stage contained fewer Rmona658 relative to gut tissue containing less hematin (Fig. 1G and H). We observed Rmona658 in salivary gland acini of a single female tick at 1 week PCF and in a second female at 2 weeks PCF (Fig. 1I and J). The rickettsiae were widely distributed within the acini as single and replicating doublet form cells. The dense clusters of Rmona658 observed in midguts were absent. In both of the 2 week PCF female ticks with midgut infections, we observed Rmona658 infection of the tracheal network. The rickettsiae were present in large tracheae with chitinous taenidia secreted by the surrounding epithelial cells and in small tracheoles without taenidia (Fig. 1K and L). In both cases dividing cell forms and clusters of rickettsiae were present, indicating active replication in epithelial cells of the tracheal network. Rmona658 were not observed in any other tissue.

We also assessed potential pathogenicity of Rmona658 capillary feeding dose to unfed male I. scapularis collected from deer. We capillary fed four groups of 26 males each with medium containing 0, or 5.6 x 10³, 10⁴ or 10⁵ rickettsiae per μl. An additional group of 20 ticks was included as a no treatment control. Over a period of 3 weeks, ticks died at approximately the same rate in all treatment groups. At 3 weeks PCF, the mean mortality was 74.6% and ANOVA indicated no significant differences between groups (P = 0.995). Infection rates, determined by PCR using CATF and CATRC primers, of dead and surviving ticks fed at the 10³, 10⁴ and 10⁵ concentrations were 100, 88 and 100 %, respectively (data not shown). EFM examination of tissues from the 26 surviving rickettsiae-fed males demonstrated Rmona658, including dividing forms, in healthy midguts of 23 ticks and within gut-associated tracheae of 1 of those ticks (data not shown). Rmona658 were not found in any other tissue.

Our preliminary results demonstrated that EFM could be used to visually monitor Rmona658 infection in unfixed tissues of dissected ticks. In unfed adult female I. scapularis infected with Rmona658 by capillary feeding, the rickettsiae were capable of disseminating from the midgut to infect the tracheal system and salivary glands. Rmona658 did not disseminate from the midgut of unfed adult male I. scapularis, unfed adult D. variabilis of either sex or partially fed adult female A. americanum under our test conditions. We selected I. scapularis for further experiments to evaluate the tick’s potential as a model host for R. monacensis expressing GFP.

Figure 2 Imaging of GFP fluorescent Rmona658 in I. scapularis ticks infected as nymphs by capillary feeding. Scale bars at lower left in each image indicate 10 μm and small arrows indicate Rmona658. (A) One week PCF nymph gut tissue. Large arrow indicates (more ...)

To further test for infection mortality as well as long-term infection maintenance, we capillary fed 100 laboratory colony F1 nymphs with Rmona658 in medium and 52 control nymphs with medium only. Two days PCF, 69 (69%) of Rmona658-fed and 37 (71%) of control nymphs survived, indicating no immediate infection induced mortality. At 3 weeks PCF, 35 (51%) of Rmona658-fed and 20 (54%) control nymphs survived, again indicating no Rmona658 induced mortality. Hamsters were infested with the nymphs. After 1 week, 27 (77%) Rmona658 fed and 14 (70%) control engorged nymphs were recovered. All molted to morphologically normal adults after approximately 1 month. After 6 months, 7 male and 7 female Rmona658 adults (52%) compared to 5 male and 4 female control adults (64%) survived. Rmona658 were observed by EFM in healthy midguts of 1 male (Fig. 2G) and 1 female tick (Fig. 2H) and in salivary glands of the female. In both ticks, Rmona658 tended to occur in the midguts as clusters rather than as the widely dispersed single and double cell forms typical of initial infections in adult ticks (Fig. 1H). The tracheal system of the female, but not the male tick, was massively infected with Rmona658 occurring in main and lateral trunk tracheae and in tracheoles (Fig. 2I and J). To further assess the rate of TST to the 14 adult ticks surviving from Rmona658 fed nymphs, tissue extracts were used as template for PCR assays. Using primers specific for the tick mitochondrial 16S rRNA gene, the predicted 460 bp product was obtained from extracts of all 14 adults (Fig. 3A, lanes 1 – 14), demonstrating presence of PCR competent DNA. Using Rmona658-specific GFP primers, only the EFM-positive female tick described above yielded a predicted first-round PCR 720 bp product (data not shown). However, nested PCR primer reactions using first round reactions as template yielded the predicted 400 bp product from the EFM-positive female and male ticks above and additional female and male ticks (Fig. 3B, lanes 3, 4, 8 and 9). Southern blot analysis of first round GFP PCR reactions confirmed results of the nested PCR analysis and indicated presence of Rmona658 in a third female tick (Fig. 3C, lanes 3, 4, 8, 9 and 10). In sum, the results indicated little or no lethal effect of capillary fed Rmona658 on I. scapularis nymphs, demonstrated maintenance of a disseminated TST infection in a 6 month post-eclosion adult female tick, and demonstrated TST to 36% of adult ticks assayed 6 months post-eclosion.

Figure 3 PCR and Southern blot detection of Rmona658 transstadial transmission from I. scapularis nymphs to adults. (A) PCR amplification of a 460 bp tick mitochondrial 16S rRNA gene product from individual adult male (lanes 1 – 7) and female (lanes 8 (more ...)

In view of rickettsial reactivation by bloodmeal (Hayes and Burgdorfer, 1982) and our demonstration of Rmona658 TST from larvae to nymphs (above), we infested hamsters with the remaining larvae and recovered nymphs after 1 month. Fluorescent Rmona658 were not observed in 20 dissected nymphs from each egg mass but presence of rickettsial DNA in 11 of 12 single nymph tissue extracts was indicated by positive PCR assay results using the general citrate synthase gene-specific RpCS.877p and RpCS.1273r PCR primer pairs (data not shown). PCR assays using the pan-SFG rickettsia 190–602n and 190-70p primer pair specific for the rompA gene indicated presence of SFG rickettsial DNA in tissue extracts of 3 pools of 5 nymphs each from each egg mass (Fig. 5, lanes P1 to P6, 532 bp band) and in all tissue extracts of 10 individual nymphs from each egg mass (lanes N1 to N20). However, PCR assays of the same DNA extracts using first round followed by nested CAT and GFP primers specific for Rmona658 and Southern blot analyses of the nested PCR reactions were negative, demonstrating that the SFG rickettsiae present in the nymphs were not Rmona658. The results demonstrated disseminated Rmona658 infections in capillary fed adult female I. scapularis ticks, the presence of SFG rickettsiae in their progeny, and that Rmona658 had not undergone TOT.

Figure 5 PCR detection of SFG rickettsiae in nymphal progeny of Rmona658-fed adult female I. scapularis. Upper and lower panels: amplification of a 532bp SFG rickettsial rompA gene product from DNA extracts of 6 pools of 5 nymphs each (lanes P1 to P6) and from (more ...)

Establishment of specific rickettsial infections in ticks by feeding on rickettsemic animals (Niebylski et al., 1999) is complicated by use of hosts with different rickettsial titers and variable immune status as well as potential antimicrobial activities derived from blood digestion products (Sonenshine et al., 2005). Intracelomic inoculation of rickettsial suspensions (Santos et al., 2002) is complicated by injury and the wound response that includes induction of the innate immune system (Nakajima, et al., 2005). Infection by capillary feeding (Macaluso et. al., 2001) avoided such complications.

In contrast to the successful infection of A. americanum nymphs by intracoelomic inoculation of Rickettsia parkeri and subsequent TST to adults (Goddard, 2003), our results suggested that capillary feeding of R. monacensis to A. americanum or to freshly molted nymphs did not result in establishment of infection. Macaluso and colleagues (2001) capillary fed Rickettsia montanensis and Rickettsia rhipicephali to D. variabilis and demonstrated disseminated infections, but capillary feeding of R. monacensis to D. variabilis did not result in disseminated infections. However, I. scapularis, supported disseminated infections and TST. The infection characteristics of Borrelia burgdorferi in the same three ticks resemble those of R. monacensis and may be determined by differences in innate and blood meal-derived immune factors and presence or absence of appropriate receptor/ligand interactions between host and microbe (Soares et al., 2006).

Rickettsial interactions with tick hosts have been most thoroughly studied in Dermacentor andersoni infected with the virulent pathogen, R. rickettsii, and/or the nonpathogenic endosymbiont, R. peacockii (reviewed in Marchette, 1982; Hayes and Burgdorfer, 1989). R. peacockii is the most closely related of SFG rickettsiae to R. rickettsii (Baldridge et al., 2004), suggesting that subtle differences may underlie pathogenicity versus nonpathogenicity. In adult female D. andersoni, R. peacockii primarily infects ovaries and midgut posterior lobes but is not found in adult males. It infects salivary glands of some larvae, but not those of nymphs or adults and is thus primarily acquired and maintained by TOT. It may exclude ovarian infection by R. rickettsii (Burgdorfer et al., 1981; Niebylski et al., 1997). In contrast, R. rickettsii is acquired by blood feeding or TOT and can be widely distributed in tissues of both sexes. It is highly pathogenic to ticks, implying a major role for horizontal transmission in its maintenance and little or no significant contribution by TST (Burgdorfer et al., 1981; Hayes and Burgdorfer, 1982; Niebylski et al., 1999). Our analysis of R. monacensis infection dynamics in I. scapularis revealed characteristics that shared aspects of the endosymbiotic and pathogenic rickettsial infection paradigms described above.

Consistent with the endosymbiotic paradigm, R. monacensis infection of I. scapularis did not result in the obvious tissue damage and high mortality caused by the spotted fever pathogens R. rickettsii in D. andersoni (Niebylski et al., 1999) and R. conorii in Rhipicephalus sanguineus (Santos et al., 2002). However, infection and replication within midgut epithelial cells usually occurred throughout that tissue rather than being confined to its posterior lobes as in R. peacockii infections. In ticks harboring pathogenic rickettsiae, the salivary glands are among the most heavily infected tissues (Santos et al., 2002). In adult female ticks, but not in males, R. monacensis disseminated from the gut and infected salivary gland acini and ducts, although we have not demonstrated presence of R. monacensis within the duct lumen. Hamsters inoculated with R. monacensis raised a strong antibody response (Simser et al., 2001) but in this study we observed no evidence of infection in a rabbit or hamsters used to blood feed infected ticks or for an immune response by the hamsters. Recently, another SFG member not associated with disease, R. massiliae strain Bar29, was shown to infect the salivary glands of field collected Rhipicephalus turanicus (Matsumoto et al., 2005). Sera of eight spotted fever patients reacted to antigens of both R. conorii and R. massilae (Cardenosa et al., 2003). Reactivity of polyclonal sera to heterologous rickettsial antigens is well documented and interferes with assessment of epidemiological situations.

Transstadial transmission is important in the maintenance of rickettsial endosymbionts in ticks. We demonstrated TST of R. monacensis from I. scapularis larvae to nymphs and from nymphs to adults of both sexes that was not accompanied by the high mortality and developmental abnormalities observed in TST of R. conorii in Rh. sanguineus (Santos et al., 2003). We did not obtain evidence for TOT of R. monacensis in I. scapularis. Rickettsial endosymbionts closely related to R. monacensis (Ishikura et al., 2002, 2003; Kurtti; unpublished) occur in I. scapularis populations throughout the United States (Benson et al., 2004). In I. scapularis collected in or near Minnesota, the endosymbionts were restricted to ovaries and underwent TOT to larvae and TST to females, but not males (Noda et al., 1997). Because endosymbionts occur at near 100% incidence in female ticks from the source population for our experiments (Kurtti et al., 2005), we speculate that ovarian superinfection resistance may explain our inability to demonstrate TOT of R. monacensis in I. scapularis. R. monacensis did not infect testicular tissues, consistent with absence of I. scapularis rickettsial endosymbionts in testicular tissues (Noda et al., 1997) and in contrast to presence of R. rickettsii in D. andersoni and R. helvetica in I. ricinus testicular tissues (Hayes, et al., 1980).

Ultrastructural TEM studies have led to reports of “rickettsia-like organisms” (RLOs) within transverse sectioned tracheae in Ixodid ticks (Tarasevich and Rehacek, 1972; Rehacek et al., 1976; Sutakova and Rehacek, 1989) and insects (Huger, 1964; Mitsuhashi and Kono, 1975; Adams and Azad, 1990). In longitudinally sectioned Siphonapteran tracheae, RLOs, later suggested to be Rickettsia felis (Bouyer et al., 2001), were observed in epithelial cells that surround and secrete the chitinous tracheae (Ito and Vinson, 1980). We have shown that R. monacensis infected the tracheal system in I. scapularis and disseminated from the midgut of female ticks to the salivary glands. Hayes and Burgdorfer (1989) proposed that structural changes induced by blood meal digestion allow rickettsiae to penetrate the gut basal lamina and enter the hemocoel to infect hemocytes that disseminate the infection to other tissues. Our demonstration of R. monacensis infection of the tick tracheal system suggests that it might serve as a dissemination pathway for rickettsiae. That hypothesis derives support from the demonstrations that nuclear polyhedrosis viruses disseminate from the gut of larval Lepidoptera via the tracheal system (Engelhard et al., 1994; Barret et al., 1998; Rahman and Gopinathan, 2004) and that arboviruses penetrate the mosquito gut basal lamina to disseminate via tissue conduits involving tracheae and visceral muscle (Romoser et al., 2004; Scholle et al., 2004). The epithelial cells that surround and secrete tracheae are metabolically active and envelope and penetrate tissues with high oxygen demand (Wigglesworth, 1972), a characteristic of the well-tracheated tick midgut (Fielden et al., 1999). The tracheal system ramifies to all internal tissues, thus offering rickettsiae a protected intracellular environment for dissemination to other tissues. Tracheal tissue might also serve as an infection reservoir during tick morphogenesis. Tick larvae are atracheate. As they prepare to molt, future tracheal epidermal cells enlarge and fuse into multinucleate cells (Balashov, 1972) in which rickettsiae could remain in an intracellular environment during tissue reorganization and then invade other tissues after development of the nymphal tracheal system. At the final molt, the tracheal system remains largely intact in contrast to almost completely regenerated tissues such as the salivary glands (Balashov, 1972). The tracheal system might thus serve as both a rickettsial dissemination pathway and infection reservoir during the tick life cycle.

Infection of I. scapularis with R. monacensis expressing GFP establishes a new model system for exploration of rickettsial tissue tropisms and infection dynamics.

This research was supported by NIH grant RO1 AI49424 to U.G.M.