INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The zoonotic vector-borne diseases form a large proportion of the emerging bacterial infectious diseases. The most prominent of these diseases are Lyme disease, ehrlichiosis, and bartonellosis. In The Netherlands 10 to 35% of the Ixodes ricinus ticks are infected with Borrelia burgdorferi, the causative agent of Lyme disease (25, 26). In 1994, general practitioners in The Netherlands reported seeing 33,000 patients who had sustained tick bites and approximately 6,500 patients with erythema migrans (7). These findings not only underline the importance of borreliosis but also suggest that other vector-borne diseases may occur in The Netherlands.

Presently, two tick-transmitted Ehrlichia species have been shown to cause human disease. The first is Ehrlichia chaffeensis, which causes human monocytic ehrlichiosis and which is transmitted by Amblyomma americanum, a tick species found only in the United States. Until now, very few cases of E. chaffeensis infection in Europe have been described (5, 15, 21). The second Ehrlichia species pathogenic for humans is the human granulocytic ehrlichiosis agent (HGE). The exact nature of this organism is still unclear, but on the basis of its 16S rRNA sequence it is shown to be closely related to Ehrlichia phagocytophila and Ehrlichia equi. The HGE agent is transmitted by Ixodes scapularis, but possibly also by other vectors like I. ricinus. Again, the initial reports of disease of human patients with HGE came from the United States. Remarkably, there have been very few reports of cases of disease caused by HGE in Europe. The major indication that ehrlichiosis may play a role in Europe comes from serosurveys performed in several European countries including Sweden, Norway, Switzerland, and the United Kingdom (5, 7, 28). However, von Stedingk et al. (28) have recently detected Ehrlichia species in Swedish I. ricinus ticks, and Ehrlichia was also detected in a French I. ricinus tick (19). These findings indicate that Ehrlichia species that are pathogenic for humans may be present in Western Europe as well.

The clinical manifestations of HGE infection can vary from a flu-like disease to severe life-threatening acute febrile disease with thrombocytopenia, leukopenia, and elevated liver transaminase levels. Because of the diffuse nonspecific symptoms of this disease, diagnosis relies heavily on laboratory tests. Serology, particularly immunofluorescence, is commonly used, but serology often does not detect antibodies in the acute phase of disease. Culture of HGE is possible, but it is very labor intensive and has not been validated as far as sensitivity is concerned. Microscopic examination of stained blood smears can be used to detect characteristic enclosures in infected leukocytes. However, this method is insensitive and requires special expertise. The sensitivity and specificity of PCR for the detection of the Ehrlichia probably exceed those of the other methods. Several PCRs for detection of Ehrlichia species have been described (1, 6, 9, 14); however, all of these assays enable the detection of just a single species. In this report we describe a PCR-hybridization assay that enables the simultaneous detection and species identification of a variety of Ehrlichia, B. burgdorferi, and Bartonella species in a single sample. This method allowed us to screen a large number of Dutch tick samples for the presence of these tick-borne pathogens.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Ticks and bacterial strains. I. ricinus ticks were collected from infested roe deer (Capreolus capreolus) shot in the Flevopolder in The Netherlands, an area where roe deer are abundant (26). Immediately after collection, the ticks were immersed in 70% ethanol and stored. The four genomic groups of B. burgdorferi sensu lato were represented by B. burgdorferi sensu stricto HB4, Borrelia garinii AR-1, Borrelia afzelii A39S, and Borrelia valaisiana M19. Crude DNA extracts from the following Ehrlichia species were used: E. phagocytophila (kindly provided by F. Jongejan) and HGE, Ehrlichia canis, and E. chaffeensis (kindly provided by S. Dumler). The two Bartonella species used in this study were Bartonella henselae ATCC 49882 and Bartonella quintana 90-268 (3).

Preparation of DNA extracts from ticks. Ticks were processed as described before (10, 23). Briefly, the ticks were taken from the 70% ethanol solution, air dried, and boiled for 20 min in 100 µl of 0.7 M ammonium hydroxide to free the DNA. After cooling, the vial with the lysate was left open for 10 min at 90°C to evaporate the ammonia. The tick lysate either was used directly for PCR or was stored at 20°C until use.

PCR amplification. PCR amplifications were performed in an Omnigene thermal cycler (Hybaid Ltd., Teddington, United Kingdom). DNA amplification was done in 50-µl reaction volumes. For the amplification of Ehrlichia DNA, each reaction mixture contained 10 pmol of primer 16S8FE and B-GA1B, 1.25 U of SuperTaq DNA polymerase (HT Biotechnology Ltd., Cambridge, United Kingdom), 0.275 µg of the TaqStart antibody (Clontech Laboratories, Palo Alto, Calif.), and standard amounts of amplification reagents (each deoxynucleoside triphosphate at a concentration of 200 µM, 10 mM Tris · HCl [pH 9.0], 50 mM KCl, 1.5 mM MgCl₂, 0.01% gelatin, 0.1% Triton X-100). A 25-µl overlay of paraffin oil was added to the tubes, followed by the addition of 5 µl of the tick DNA extract. To minimize nonspecific amplification a touchdown PCR program was used: 3 min at 94°C, two cycles of 20 s at 94°C, 30 s at 67°C, and 30 s at 72°C, and then two cycles with conditions identical to the previous cycles but with an annealing temperature of 65°C. During subsequent two cycle sets the annealing temperature was lowered by 2°C until it reached 57°C. Then, an additional 40 cycles each consisting of 20 s at 94°C, 30 s at 57°C, and 20 s at 72°C, followed the touchdown program, were performed. The PCR was ended by an extra incubation for 7 min at 72°C. For the amplification of B. burgdorferi sensu lato DNA, conditions similar to those described above were used, except that 40 pmol of the primers 23SN2 and 5SCB and double the amounts of SuperTaq and TaqStart were used. In addition, the touchdown PCR temperature ranged from 60 to 50°C. For the amplification of Bartonella DNA, the previously described PCR protocol of Bergmans et al. (4) was used.

Reverse line blot hybridization. The reverse line blotting technique has been described before (11, 12, 25). Briefly, solutions with 5' amino-linked oligonucleotide probes ranging from 10 to 800 pmol were coupled covalently to an activated Biodyne C membrane in a line pattern by using a miniblotter (Immunetics, Cambridge, Mass.). After binding of the oligonucleotide probes the membrane was taken from the miniblotter, washed in 2× SSPE (360 mM NaCl, 20 mM Na₂HPO₄ · H₂O, 2 mM EDTA) with 0.1% sodium dodecyl sulfate (SDS) at 60°C, and again placed in the miniblotter with the oligonucleotide lines perpendicular to the slots. Ten microliters of the biotin-labeled PCR product was diluted in 150 µl of 2× SSPE-0.1% SDS, denatured for 10 min at 99°C, and cooled rapidly on ice. The slots of the miniblotter were filled with the denatured PCR product, and hybridization was performed for 1 h at 42°C. The membrane was removed from the miniblotter and was washed twice for 10 min each time in 2× SSPE-0.1% SDS at 51°C. Subsequently, the membrane was incubated for 30 min at 42°C with streptavidin-peroxidase (Boehringer Mannheim GmbH, Mannheim, Germany) diluted 1:4,000 in 2× SSPE-0.5% SDS and was washed twice for 10 min in 2× SSPE-0.5% SDS. Hybridization was visualized by incubating the membrane with enhanced chemiluminescence detection liquid (Amersham International plc, Den Bosch, The Netherlands) and exposing the membrane to X-ray film (Hyperfilm; Amersham). For species identification the biotinylated Ehrlichia PCR product was hybridized with seven different oligonucleotide probes in the reverse line blot assay. Similarly, the biotinylated spacer fragment of the Borrelia PCR was hybridized with five B. burgdorferi genospecies-specific oligonucleotide probes. For identification of Bartonella species a region between bases 964 and 1243 of the 16S rRNA gene was amplified and was used in a reverse line blot assay. All primers and probes are described in Table 1.

DNA sequencing and data analysis. The PCR products used for DNA sequencing were purified with Qiaquick PCR purification kits (Qiagen, Hilden, Germany). For DNA sequencing reactions, the fluorescence-labeled dideoxynucleotide technology was used (Perkin-Elmer, Applied Biosystems Division). The sequenced fragments were separated, and data were collected on an ABI 377 automated DNA sequencer (Perkin-Elmer, Applied Biosystems Division). The collected sequences were assembled, edited, and analyzed with the DNAStar package (DNAStar Inc., Madison, Wis.). The phylogenetic tree was constructed by using the Clustal analysis in the Megalign module of the DNAStar package.

Nucleotide sequence accession number. The 16S rRNA gene sequence of the Ehrlichia-like organism found in this study is available in the GenBank database under accession no. AF104680.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Sensitivity of the PCR for Ehrlichia and Borrelia. The sensitivity of the PCR for Ehrlichia and Borrelia was assessed by spiking the samples with known concentrations of previously produced PCR products. Extracts from ticks that were negative by previous PCRs were spiked with serial dilutions of E. phagocytophila or B. burgdorferi PCR products. All experiments were performed in duplicate. Repeatedly, the detection limit for both assays in which the PCR yielded a positive result was five copies of the target sequence (data not shown). The detection limit of the PCR for Bartonella has been determined in another study and was shown to correspond to one genome copy (4).

Specificity of the reverse line blot hybridization. A reverse line blot hybridization assay was designed to differentiate the various Ehrlichia, B. burgdorferi, and Bartonella species. Some of the Ehrlichia species differ by only one nucleotide in the target sequence (Fig. 1). The oligonucleotide probes were designed in such a way that the melting temperature of all oligonucleotides was approximately 55°C under the conditions used. As a result the oligonucleotide probes differ in length. In order to obtain specific and sensitive signals in the assay the optimal oligonucleotide probe concentrations and hybridization conditions were determined empirically. To assess the specificity of the assay, control samples were amplified and used in the reverse line blot hybridization assay under stringent hybridization conditions. Figure 2 shows the reactivities of the control samples for Ehrlichia, B. burgdorferi, and Bartonella species in the optimized reverse line blot assay. No cross-hybridization between the various species occurred, indicating that the system distinguished target sequences that differed by only a single base pair.

View larger version (8K): [in this window] [in a new window]   FIG. 1.   Multiple alignment of the variable part of the 16S rRNA gene sequences of the E. phagocytophila group. The region where differences were detected is shown for E. phagocytophila, E. equi, HGE, and the variants of E. phagocytophila and HGE. The positions of the residues that differ in the 16S rRNA gene are shown below the multiple alignment. Residues identical to those of the E. phagocytophila sequence are indicated by a dot. The sequence of E. equi was obtained from GenBank (accession no. M73223). The 500 bp of the 5' end of the 16S rRNA gene sequences of E. phagocytophila and HGE were determined in this study and compared with the sequences in the GenBank and EMBL database and were found to be identical to published sequences (accession nos. M73320 and U02521, respectively). The 16S rRNA sequences of the E. phagocytophila variant and the HGE variant were determined in this study.

View larger version (47K): [in this window] [in a new window]   FIG. 2.   Reverse line blot hybridization assay analyses for the detection and identification of Ehrlichia, B. burgdorferi, and Bartonella spp. in ticks. (A) Membrane carrying Ehrlichia-specific oligonucleotide probes; (B) membrane carrying B. burgdorferi genospecies-specific probes; (C) combined membrane carrying probes for Ehrlichia, B. burgdorferi, and Bartonella species. The oligonucleotide probes are attached to the membrane in the horizontal direction, and the PCR samples were applied perpendicularly in the vertical direction. The numbered lanes represent tick-derived PCR products, and the lanes marked with letters show the PCR products obtained from the positive control samples. El, Ehrlichia-like; Eca, E. canis; Ech, E. chaffeensis; Epv, E. phagocytophila variant; Hv, HGE variant; Ep, E. phagocytophila; H, HGE; Bs, B. burgdorferi sensu stricto; Bg, B. garinii; Ba, B. afzelii; Bv, B. valaisiana; Bh, B. henselae; Bq, B. quintana.

PCR detection of Ehrlichia, Borrelia, and Bartonella DNAs in ticks. The various PCRs were applied to DNA extracts from 121 I. ricinus ticks collected from 38 different roe deer. The majority of the ticks were adults, mainly females; and some were nonengorged, some were semiengorged, and some were fully engorged (Table 2). Regardless of whether these PCRs yielded a visible fragment on agarose gels, all samples were analyzed by the reverse line blot assay. Fifty-four of the 121 samples (45%) reacted with one or more of the Ehrlichia-specific probes (Table 2). Of these, 3 reacted with the HGE-specific probe, 3 reacted with the E. phagocytophila-specific probe, 11 reacted with the HGE variant-specific probe, 9 reacted with the E. phagocytophila variant-specific probe, and 7 reacted with both the HGE variant-specific and the E. phagocytophila variant-specific probes. In addition, 19 of the samples reacted solely with the Ehrlichia genus-specific probe. Sixteen of the same 121 tick samples (13%) reacted in the Borrelia PCR hybridization assay with the B. burgdorferi species-specific probes (Table 2). Coinfection with two different B. burgdorferi genospecies was detected in four tick samples. None of the ticks analyzed was infected solely with B. garinii or with B. burgdorferi sensu stricto. However, these genospecies were found in ticks coinfected with different genospecies.

DNA sequence analysis. In order to confirm the results obtained by the reverse line blot assay, 15 of the Ehrlichia-positive samples were also analyzed by DNA sequencing. Sequencing of the PCR product obtained by PCR for Ehrlichia revealed that we had correctly identified the various species by the reverse line blot hybridization assay. Furthermore, sequence analysis of the samples that reacted with both the HGE variant-specific and the E. phagocytophila variant-specific probes revealed an ambiguous nucleotide at position 100 in the 16S rRNA gene. Cloning of these PCR products and subsequent reverse line blot assay analysis and DNA sequencing of the clones revealed the presence of two types of cloned 16S sequences, indicating that these tick samples indeed carried a mixture of two different Ehrlichia sequences.

View larger version (21K): [in this window] [in a new window]   FIG. 3.   Dendrogram showing the phylogenetic relationships of the 16S rRNA gene sequences of the newly identified Ehrlichia-like and those of other rickettsiae. The tree was constructed by comparing sequences of the segment of the 16S rRNA gene ranging from bases 40 to 1434 (E. phagocytophila coordinates). The scale beneath the tree measures the distance between sequences expressed as the number of substitution events. The sequences used for comparison were obtained from the GenBank and EMBL database.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We developed a PCR-based reverse line blot hybridization assay in which Ehrlichia, B. burgdorferi, and Bartonella species can be detected and differentiated. The assay was specific enough to detect single-base-pair changes with immobilized oligonucleotide probes and enabled us to differentiate Ehrlichia variants. The reverse line blot technique is a relatively easy and rapid method for the simultaneous detection and identification of microorganisms in field samples such as ticks. In its present form we can combine the hybridization of PCR products obtained in separate PCRs. We are now developing a multiplex PCR that will enable us to have an even more convenient method for the screening of samples. These samples could be tick lysates but could also be other material such as blood from patients suffering from a febrile disease with an unknown origin.

In the study presented here we used this method to detect and identify Ehrlichia and B. burgdorferi species in Dutch I. ricinus ticks. Analysis of the ticks showed an unexpected high rate of infection with Ehrlichia species (45%). The high infection rate may be partly due to the fact that the ticks originated from roe deer, which may serve as a reservoir for Ehrlichia. However, there was no significant correlation between sex and engorgement of the ticks and infection with Ehrlichia species. In addition, ticks collected from the same roe deer carried a variety of Ehrlichia and Borrelia species. This suggests that the ticks may have been infected before feeding on the roe deer and that the Ehrlichia spp. originated from other reservoirs. In order to get a more accurate impression of the prevalence of Ehrlichia infection in Dutch ticks, we are now analyzing a large number of ticks collected from the vegetation. Whatever the reservoir may be, the results obtained in this survey suggest that Dutch ticks may pose a serious health threat to both humans and animals and should be used to warn clinicians to be aware of the possible presence of ehrlichiosis in The Netherlands.

The majority of the Ehrlichia species found in this study belong to the E. phagocytophila group. As expected, neither E. canis nor E. chaffeensis was found in any of the ticks. Analysis of PCR products revealed that the 16S rRNA gene sequences of the E. phagocytophila group showed slight variations. In total, four types of E. phagocytophila-like sequences were found: species with the E. phagocytophila or the HGE 16S rRNA gene sequences and two variants of these sequences that carried a substitution of a single base pair at position 92 of the 16S rRNA gene. This corroborates the findings of a Swedish group (28) and a group from the United States (2) that also found Ehrlichia species in which the A at position 92 of the 16S gene was substituted by a G. It remains to be determined whether the 16S rRNA variants represent different Ehrlichia species. It is possible that the HGE agent, E. phagocytophila, and the variants found in this study all belong to the same species and should be designated E. phagocytophila subspecies. Furthermore, it is unclear whether these variants can cause disease in humans or animals. It was remarkable that in none of the samples of the E. phagocytophila group from which the 16S rRNA gene sequences were determined was a C found at position 49 in the 16S rRNA gene. The presence of a C at this position may be characteristic for E. equi. This would corroborate earlier observations that E. equi was not found in Europe.

More than 6% of the ticks were infected with an Ehrlichia-like organism not described before. This organism is closely related to but clearly distinct from the monocytic group of Ehrlichia species and C. ruminantium. It is unclear whether this organism can cause disease in mammals, but experimental infection of animals may confirm its infectious nature. The newly identified organism may represent an endosymbiont. Examples of such endosymbionts in ticks are the Francisella and Wolbachia species, which are found at high rates in particular tick species (16-18). However, the relatively low frequency of infection of the ticks would argue against this hypothesis.

Analysis of the 121 ticks showed that 13% of the ticks carried B. burgdorferi species and confirmed earlier findings that 10 to 35% of the Dutch I. ricinus ticks are infected with B. burgdorferi genospecies (24). Interestingly, 5 of the 121 ticks were coinfected with Ehrlichia and two genospecies of B. burgdorferi. Due to its immunosuppressive nature, coinfection with Ehrlichia and B. burgdorferi may increase the severity of Lyme borreliosis.

Transmission of Bartonella species by ticks is speculative. However, at least one study reports on three patients with B. henselae bacteremia. These patients had no history of contact with cats but sustained tick bites prior to the bacteremia (13). From the study presented here it is clear that a large proportion of the ticks carry Bartonella species or species closely related to Bartonella but not the human pathogens B. henselae and B. quintana. The Bartonella species found might originate from small rodents on which the ticks may have been feeding. This could indicate that transmission of Bartonella species between rodents is, at least in some part, tick mediated. Further studies with other arthropods such as body lice and perhaps also blood from rodents such as rats may disclose the reservoirs and vectors for B. quintana.

Until now there have been no reports of ehrlichiosis in Dutch patients. Therefore, the high rate of infection of Dutch ticks with Ehrlichia species raises the question of whether human ehrlichiosis does occur in The Netherlands. It is known that Ehrlichia species cause infections in cattle, sheep, and dogs in Europe. However, until now there have been very few reports on human ehrlichiosis in Europe (15, 20, 27). In fact, only recently was the first case of granulocytic ehrlichiosis infection reported, and that was in Slovenia (20). Although the seroprevalence in several European serosurveys suggest that infections with Ehrlichia do occur in Europe, there seems to be a paucity of reported cases. There may be several explanations for this phenomenon. First, it is possible that there really are very few cases of human ehrlichiosis. Second, the majority of cases may go unnoted because they are caused by less virulent variants of HGE that result in a mild course of disease. Finally, cases of ehrlichiosis may remain unnoted because clinicians do not recognize the disease. Relatively few clinicians know that the disease exists and therefore cannot make the correct diagnosis. Furthermore, the tools used to diagnose ehrlichiosis are usually lacking. Very few laboratories in The Netherlands are equipped to perform serology studies for Ehrlichia, and PCR is performed in none of these laboratories. Therefore, at least in The Netherlands, ehrlichiosis may have been overlooked. Recently, a Swedish group reported on three PCR-confirmed cases of HGE infection in humans (PROMED file 980418193622). Two of the three patients were seronegative, which forewarns us that serology may not suffice for the diagnosis of ehrlichiosis. The patients showed a variety of clinical symptoms, of which only fever and headache were seen in all three patients. Remarkably, the initial diagnosis for one of the patients was neuroborreliosis, and the patient was treated for this condition. These findings indicate that HGE infections do occur in Europe and suggest that there may indeed be an underdiagnosis of ehrlichiosis and that surveillance is required to determine the true extent of the problem.