Development and Evaluation of PCR Assay for Detection of Low ...

JOURNAL OF CLINICAL MICROBIOLOGY, Jan. 1995, p. 166–172 0095-1137/95/$04.0010 Copyright q 1995, American Society for Microbiology

Vol. 33, No. 1

Development and Evaluation of PCR Assay for Detection of Low Levels of Cowdria ruminantium Infection in Amblyomma Ticks Not Detected by DNA Probe TREVOR F. PETER,1 SHARON L. DEEM,1 ANTHONY F. BARBET,2 R. ANDREW I. NORVAL,2† BIGBOY H. SIMBI,1 PATRICK J. KELLY,3 AND SUMAN M. MAHAN1* University of Florida/USAID/SADC Heartwater Research Project, Causeway, Harare,1 and Faculty of Veterinary Science, University of Zimbabwe, Mount Pleasant, Harare,3 Zimbabwe, and Department of Infectious Diseases, College of Veterinary Medicine, University of Florida, Gainesville, Florida 32611-08802 Received 14 July 1994/Returned for modification 14 September 1994/Accepted 20 October 1994

The sensitivities of a PCR assay and a DNA probe assay were compared for the detection of Cowdria ruminantium in Amblyomma ticks that were fed on C. ruminantium-infected, clinically reacting, and recovered carrier animals. The PCR assay and DNA probe detected infection in 86.0 and 37.0%, respectively, of 100 ticks fed on a febrile animal. In 75 ticks fed on carrier animals, PCR and the DNA probe detected infection in 28.0 and 1.33% of ticks, respectively. This demonstrates that the DNA probe has poor sensitivity for the detection of low levels of infection in ticks and that PCR is necessary for this purpose. The PCR assay had a detection limit of between 1 and 10 C. ruminantium organisms and did not amplify DNA from Ehrlichia canis, which is phylogenetically closely related to C. ruminantium, Theileria parva, or uninfected Amblyomma hebraeum or A. variegatum. PCR detected infection in A. hebraeum and A. variegatum adult ticks infected with one of six geographically different C. ruminantium strains. Amplification was also possible from desiccated ticks and ticks fixed in 70% ethanol, 10% buffered formalin, or 2% glutaraldehyde. The PCR assay supersedes the DNA probe and older detection methods for the detection of C. ruminantium in ticks, particularly those fed on carrier animals, and is suitable for both prospective and retrospective studies which require accurate detection of C. ruminantium in individual ticks. Application of the PCR assay should significantly improve the understanding of heartwater epidemiology, particularly through the determination of field tick infection rates.

unsuitable for routine screening of tick infections. For more convenient detection of infection, a C. ruminantium-specific DNA probe, pCS20, was developed and used successfully for the detection of C. ruminantium in A. variegatum and A. hebraeum ticks fed on experimentally infected, clinically ill (febrile) small ruminants (30, 58, 61). The probe detected infections in 47 to 93% of these ticks. However, questions remain as to whether the DNA probe-negative ticks in these studies were truly uninfected or a more sensitive test for the detection of low-level infections is required. Furthermore, prior to use in determining infection rates in field ticks, the DNA probe should be evaluated on ticks fed on carrier animals, which in areas where the disease is endemic represent the major source of infection for the vector. In our preliminary studies, batches of ticks fed on carriers and proven to be infected by the transmission of infection to small ruminants tested negative by the pCS20 DNA probe, suggesting that the low-level rickettsemias of carrier animals (2) can lead to low-level infections in ticks which are undetectable by DNA probe analysis. To be able to detect such infections, a PCRbased assay for C. ruminantium in Amblyomma ticks was developed. PCR has previously been shown to be a highly sensitive and specific method for the detection of rickettsial and other vector-borne agents in arthropods (3, 13, 16, 33, 40, 46, 49, 51). The PCR and DNA probe assays were compared by applying them simultaneously to ticks fed on a febrile animal and carrier animals. Additionally, the ability of the PCR assay to detect infection in dried and fixed ticks was investigated with the aim of developing a test suitable for examining archival specimens and preserved ticks collected from the field.

The tick-borne rickettsia Cowdria ruminantium is an obligate intracellular bacterium that parasitizes vascular endothelial cells, neutrophils, and macrophages of the mammalian host (10, 28, 42) and causes a disease in ruminants called heartwater. This disease is characterized by fever, nervous signs, hydrothorax, and hydropericardium and is frequently fatal in susceptible animals (55). Members of the Amblyomma tick genus, most importantly A. variegatum and A. hebraeum, are the only proven vectors (5, 11). Currently, heartwater occurs throughout most of sub-Saharan Africa and on three Caribbean islands (38, 41). The disease ranks as one of the most important vector-borne afflictions of livestock in Africa (53). The detection of C. ruminantium infection in Amblyomma ticks is essential for developing an understanding of the epidemiology of heartwater and devising effective control measures. Previously, this has mainly been achieved by the inoculation of tick extracts into, or the feeding of ticks on, susceptible small ruminants (2, 6, 8, 9, 35) or mice (14). This has traditionally been accepted as the ‘‘gold standard’’ for detecting C. ruminantium infection in ticks, although it is a laborious and expensive method. There have been additional reports on the use of light, electron, or fluorescence microscopy of tick tissues (4, 24–26, 60) or enzyme-linked immunosorbent assay techniques (34) to detect infections. All of these methods are cumbersome and time-consuming and have unknown sensitivities and specificities. They are therefore * Corresponding author. Mailing address: University of Florida/ USAID/SADC Heartwater Research Project, P.O. Box CY 551, Causeway, Harare, Zimbabwe. Phone and fax: 263-4-794980. † Deceased. 166

PCR FOR C. RUMINANTIUM IN AMBLYOMMA TICKS

VOL. 33, 1995

MATERIALS AND METHODS Isolation of C. ruminantium organisms and DNA. Six C. ruminantium strains were used. These included three Zimbabwean strains (Crystal Springs [7], Mbizi [29], and Plumtree [29]), one South African strain (Ball-3 [17]), one Nigerian strain (Nigeria [57]), and one Caribbean strain (Gardel [54]). C. ruminantium organisms were isolated from supernatants of infected bovine endothelial cell cultures (7), purified on discontinuous Percoll gradients, stained with either fluorescein diacetate (live counts) or acridine orange (total counts), and enumerated by UV light microscopy (30). C. ruminantium DNA was obtained from Percoll-purified organisms by standard techniques (47). Isolation of DNA for specificity tests. Ehrlichia canis (Oklahoma strain) and Theileria parva (Zimbabwean Boleni strain [27]) organisms were isolated from infected canine macrophage cell line DH82 and bovine lymphoblastoid cell cultures, respectively. E. canis organisms were isolated as previously described (22). To isolate Theileria organisms, infected cells were frozen and thawed once and cellular debris was pelleted at 400 3 g for 15 min. The supernatant was centrifuged at 30,000 3 g for 30 min, and the pellet was washed three times with phosphate-buffered saline (PBS). DNA was purified from isolated E. canis and T. parva organisms by conventional methods (47). DNA was also extracted from uninfected cultured bovine endothelial and canine macrophage (DH82) cells, caprine and ovine leukocytes, and uninfected adult A. hebraeum and A. variegatum ticks. Source of ticks. A. hebraeum (Sengwe strain) and A. variegatum (Trafalgar strain) ticks from laboratory colonies established from original collections made in Zimbabwe were used. These tick colonies have been maintained by feeding on laboratory rabbits and heartwater-free cattle. Infection of animals and ticks. Six Merino sheep that were Western blot (immunoblot) negative for C. ruminantium infection (30) were obtained from a heartwater-free region of Zimbabwe. Each sheep was inoculated intravenously with one of six strains of a cryopreserved blood stabilate of C. ruminantium (36). Rectal temperatures were monitored daily, and all the sheep became febrile (rectal temperature, .40.58C) between 12 and 16 days after inoculation. C. ruminantium infection was confirmed by demonstration of colonies within brain capillary endothelial cells in brain crush smears prepared from brain biopsies taken on the third day of febrile reaction (50) and by Western blot analysis of the sheep that survived. Alternatively, heartwater was confirmed in the sheep that died (ovines infected with Ball-3 and Nigeria strains) by examination of brain crush smears prepared after death (43) and postmortem findings. Two hundred uninfected, unfed A. hebraeum or A. variegatum nymphs were placed in cloth bags attached to the shaven backs of each sheep daily from the 6th to the 10th day after infection. A. hebraeum nymphs were fed on sheep infected with the Crystal Springs, Mbizi, Plumtree, and Ball-3 strains, while A. variegatum nymphs were fed on sheep infected with the Gardel and Nigeria strains. Ticks completed engorgement and detached during the febrile reaction of each infected sheep. Uninfected A. hebraeum and A. variegatum nymphs fed on seronegative sheep from heartwater-free areas of Zimbabwe served as negative controls. All engorged ticks were incubated at 288C and 75% relative humidity and allowed to molt to the adult stage. To obtain A. hebraeum ticks infected with C. ruminantium by carrier animals, three 18-month-old seronegative Mashona cattle (Sanga breed) were inoculated intravenously with 5 3 107 viable C. ruminantium cell culture-derived, Percollpurified organisms (Mbizi strain). One bovine, no. 6, was febrile (rectal temperature, .39.58C) for 3 days, starting on day 13 after infection, and was then treated with oxytetracycline at 10 mg/kg for 3 days to prevent death. The two remaining bovines, no. 4 and 7, showed no significant febrile reaction and were not treated. The infection and carrier status of cattle was confirmed by Western blot analysis and tick transmission of infection to susceptible goats after recovery (see below). Six hundred uninfected A. hebraeum nymphs were placed on each bovine 45 days postinfection, which was estimated to be approximately 1 month after their recovery on the basis of the febrile reactions of bovine no. 6 and other cattle similarly infected. Uninfected nymphs were fed on clean cattle to provide negative controls. All engorged ticks were incubated until after molting as described above. Transmission of C. ruminantium by ticks. Thirty adult A. hebraeum ticks (15 males and 15 females) from each batch of nymphs that had fed to repletion on the three recovered cattle were fed on susceptible goats to allow transmission of C. ruminantium. Additionally, 30 adult ticks from a batch of nymphs that had engorged on a febrile sheep, no. 5112, infected with the Plumtree strain were also placed on a goat. This was done to prove that the cattle were heartwater carriers and to demonstrate that the ticks from these carriers and sheep no. 5112 were infected. The rectal temperatures of goats were monitored, and heartwater was confirmed by examination of brain crush smears and postmortem findings. Processing of ticks. Ticks were processed between 1 and 3 months after molting. The internal organs (including midgut and salivary glands) of each tick were dissected and transferred to individual microcentrifuge tubes. Crosscontamination between tick samples was prevented by the use of new scalpel blades and needles for each dissection. Dissected tissues were stored at 2708C until further processing. Tissues were frozen and thawed twice and digested at 378C for 30 min in 100 ml of digestion buffer (10 mM Tris HCl [pH 8.3], 50 mM KCl, 2.5 mM MgCl2, 0.5% Tween 20, 0.5% Nonidet P-40) containing 5 mg of lysozyme per ml. Tissues were digested for a further 16 h after the addition of

167

proteinase K at 250 mg/ml and then for 1 additional h at 608C. Samples were boiled for 10 min, undigested material was pelleted at 12,000 3 g for 5 min, and the supernatant was transferred to a fresh tube. DNA was extracted once each with phenol, phenol-chloroform–isoamyl alcohol (24:1), and chloroform-isoamyl alcohol (24:1). Extracted tick DNA samples were stored at 2208C until PCR and DNA probe analysis. Desiccated dead ticks and ticks fixed in either 70% ethanol, 10% buffered formalin, or 2% glutaraldehyde in 0.12 M sodium cacodylate–0.06 M sucrose buffer were also processed for PCR. Adult A. hebraeum ticks from a batch fed as nymphs on a sheep acutely infected with C. ruminantium (Crystal Springs strain) were used for these analyses. Live ticks from the same batch had been shown to be 85% infected by DNA probe analysis (30). For fixation, 10 fresh ticks (5 males and 5 females) were cut in half and placed individually in each fixative. Fixed ticks were stored at room temperature for over 1 year before analysis. Internal tissues were removed from fixed ticks and transferred to microcentrifuge tubes. Tissues were washed three times with 1 ml of sterile PBS and pelleted by centrifugation at 12,000 3 g for 10 min after each wash. The final pellets were processed for PCR as described for fresh tick samples. To analyze desiccated specimens, ticks that had died naturally over a period of 1 year after molting were crushed in microcentrifuge tubes with plastic pestles (Kontes; Vineland, N.J.), and the ground material was digested and extracted for PCR as described for fresh tick tissue. PCR analysis. Oligonucleotide primers, AB 128 (59-ACTAGTAGAAATTG CACAATCTAT-39) and AB 129 (59-TGATAACTTGGTGCGGGAAATCC TT-39), which amplify a 279-bp region of open reading frame 2 of the 1,306-bp pCS20 sequence of C. ruminantium were used as primers for PCR amplification (31, 58). These primers do not specifically amplify bovine, Anaplasma marginale, Babesia bigemina, Trypanosoma brucei brucei, or Escherichia coli DNA (31). PCRs were performed with 5 ml of sample DNA in 50-ml reactions. PCR conditions were the same as described previously (31) except that the concentration of MgCl2 in each reaction was 3.0 mM and that of each oligonucleotide was 0.5 mM. Uninfected tick samples and positive- and negative-reaction controls, containing 1 ng of purified C. ruminantium genomic DNA (Mbizi stain) and no DNA, respectively, were included with each batch of PCRs. Completed PCR products were analyzed by Southern or dot blotting and pCS20 DNA probe hybridization. Twenty microliters of each PCR mixture was electrophoresed through 1.5% agarose gels which contained 0.4 mg of ethidium bromide per ml, and gels were photographed following transillumination with UV light. Electrophoresed DNA was transferred to nylon membranes (GenescreenPlus; Du Pont) by the capillary blot technique (47). Alternatively, 20 ml of each PCR mixture was denatured with 0.4 N NaOH for 15 min at 378C and dot blotted onto nylon membranes by vacuum filtration. Blotted DNA was hybridized with the pCS20 DNA probe as previously described (31, 61). The results were visualized by exposure to X-ray film (Kodak X-Omat) for 1 to 7 days. DNA probe analysis of ticks. After the removal of a 5-ml aliquot of each extracted tick DNA sample for PCR, the remainder of the sample was denatured with 0.4 N NaOH, dot blotted onto nylon membranes, hybridized with the [a-32P]dCTP-labeled pCS20 DNA probe as previously described, and exposed to X-ray film for 1 to 3 weeks (31, 61). Determination of PCR detection limit. A stock of Percoll-purified C. ruminantium (Mbizi stain) which contained 109 organisms per ml in PBS was frozen and thawed once, digested at 378C for 16 h with proteinase K at 100 mg/ml, and then boiled for 10 min. To determine the limit of detection for the PCR assay with tick samples, serial 10-fold dilutions of this stock were made and equivalents of 106 to 1021 organisms were added as templates to individual PCRs which contained 5 ml of extracted uninfected A. hebraeum male or female tick sample DNA. To compare the sensitivity of PCR in the absence of tick sample DNA, the same dilutions were used to spike reactions which contained 5 ml of distilled water instead of a tick DNA sample. In an analogous experiment to determine the detection limit with pure C. ruminantium DNA, serial 10-fold dilutions of genomic DNA from Percoll-purified organisms, 100 pg to 0.1 fg, were also prepared and used to spike individual PCRs which contained 5 ml of extracted male or female tick sample DNA or 5 ml of distilled water. Specificity testing of the PCR assay. Five nanograms of E. canis, T. parva, A. hebraeum, A. variegatum, bovine endothelial cell, or ovine and caprine leukocyte DNA was used as a template in individual PCRs. The specificity test was performed with E. canis DNA because of the close phylogenetic and serologic relationship between E. canis and C. ruminantium (22, 56). As a positive control for E. canis DNA, PCR was performed with 5 ng of E. canis DNA by using E. canis-specific primers, E2 (59-GTGGCAGACGGGTGAGTAATGC-39) and Ec1 (59-CAATTATTTATAGCCTCTGGCTATAGG-39), which are derived from 16S rDNA and amplify a 350-bp fragment (32). Reaction conditions for the E. canis PCR were the same as those for the C. ruminantium PCR. The E. canis primers were also tested against C. ruminantium DNA. Twenty microliters of PCR products from each specificity test was electrophoresed on agarose gels, Southern blotted, and hybridized with the pCS20 DNA probe. To evaluate the specificity of the PCR assay with batches of uninfected ticks, PCRs were conducted with ticks obtained from laboratory colonies. Thirty-six adult A. hebraeum (18 males and 18 females) and 34 adult A. variegatum (17 males and 17 females) organisms were tested. DNA from DNA probe-positive infected ticks were tested concurrently. Twenty microliters of products from these PCRs was separated on agarose gels, Southern blotted, hybridized with a

168

PETER ET AL.

J. CLIN. MICROBIOL.

FIG. 1. Removal of PCR-inhibitory substances from crude tick digests. (A) PCR before DNA extraction. The 279-bp C. ruminantium-specific DNA fragment was amplified from purified C. ruminantium genomic DNA (1 ng; Mbizi strain) in the presence of 10, 1, or 0.1 ml of unextracted tick DNA from two male ticks (lanes 1 to 3 and 4 to 6) and two female ticks (lanes 7 to 9 and 10 to 12). (B) PCR after DNA extraction. After phenol-chloroform extraction of the same tick digests used in panel A, 10, 1, or 0.1 ml, respectively, of each digest was added to PCRs which contained 1 ng of purified C. ruminantium genomic DNA. The sequence of samples is the same as in panel A. In each panel, the 123-bp DNA standard ladder and positive- and negative-reaction control PCR products are in lanes L, 1, and 2, respectively.

digoxigenin-UTP-labeled pCS20 DNA probe, and detected by chemiluminescence according to the manufacturer’s instructions (Genius DNA labeling and detection kit; Boehringer Mannheim, Indianapolis, Ind.). Specificity of pCS20 DNA probe against E. canis. To test the specificity of the pCS20 DNA probe against E. canis, serial dilutions of purified E. canis and uninfected canine macrophage (DH82 cell) DNA (400, 40, and 4 ng of each) were dot blotted onto a nylon membrane and hybridized with the probe. C. ruminantium genomic DNA (100, 10, 1, and 0.1 ng) was included on the blot as a positive control.

RESULTS Development and evaluation of the PCR assay. Initial PCRs performed with unextracted DNA from digests of infected tick tissues were unsuccessful, despite the strongly positive signals of DNA from the same ticks after dot blot hybridization with the DNA probe (results not shown). This suggested the presence of PCR inhibitors in these samples. To confirm this, purified C. ruminantium genomic DNA (1 ng; Mbizi strain) was amplified in the presence of 10, 1, or 0.1 ml of unextracted tick DNA samples from two male ticks and two female ticks. Amplification of the C. ruminantium-specific 279-bp product was inhibited in the presence of 10 and 1 ml of crude DNA from all ticks (Fig. 1A). To remove this inhibition, tick DNA samples were extracted once each with phenol, phenol-chloroform, and chloroform. When 10, 1, or 0.1 ml of the same tick digests after extraction was added to PCRs which contained 1 ng of purified C. ruminantium genomic DNA, strong uniform amplification of the 279-bp product occurred in all PCRs, regardless of the volume of tick DNA sample added, demonstrating the removal of PCR inhibition by conventional DNA extraction (Fig. 1B). All subsequent tick DNA samples for PCR were extracted prior to analysis. To determine the detection limit of the PCR assay with tick DNA samples after extraction, equivalents of 106 to 1021 C. ruminantium organisms were added to individual PCRs which contained extracted male or female uninfected tick sample DNA. Following agarose gel electrophoresis and ethidium bromide staining of the products from these PCRs, the detection limit of the PCR assay was found to be between 102 and 10 organisms (Fig. 2A). After dot blotting and hybridization with the pCS20 DNA probe, amplification was detected from one organism (Fig. 2B). On repetition of this experiment, the detection limit after hybridization varied between 1 and 10 organisms. Similar PCRs performed in the absence of tick sample DNA showed an identical detection limit (data not shown), demonstrating the lack of PCR inhibition by the tick DNA sample. As confirmation of these results, amplification

was also detectable from between 1 and 10 fg of purified C. ruminantium DNA, the genomic equivalent of between approximately one and seven organisms, assuming that the genome size of C. ruminantium is equivalent to that determined for the phylogenetically closely related rickettsia Anaplasma marginale (1,250 kbp [1]). As PCR can detect 10 organisms in the 1/20 of each tick sample used as the template for each reaction, the assay has a theoretical detection limit of 200 organisms per tick. The detection limit of the pCS20 DNA probe with DNA of C. ruminantium organisms purified on Percoll gradients from culture (with minimal bovine DNA contamination) is 0.1 ng, which is equivalent to the genomes of approximately 70,000 organisms. Thus, PCR increases the sensitivity for detection of C. ruminantium infection in a tick by at least 350-fold. To test the specificity of the PCR assay, PCRs were per-

FIG. 2. Detection limit of PCR. (A) PCR products from reactions which contained 5 ml of uninfected tick sample DNA and were spiked with 10-fold serial dilutions of digested C. ruminantium organisms, 106 to 1021 organisms (lanes 1 to 8, respectively), after agarose gel electrophoresis and ethidium bromide staining. The 123-bp DNA standard ladder and positive- and negativereaction control PCR products are in lanes L, 1, and 2, respectively. (B) Analysis of spiked PCRs after dot blotting and hybridization with pCS20 DNA probe. The sensitivity of detection after hybridization is increased 10-fold. The number of organisms used as the starting template for each PCR is given above each lane.

VOL. 33, 1995


169

FIG. 3. Specificity of the C. ruminantium PCR assay against E. canis. (A) Agarose gel electrophoresis of products from PCRs with the C. ruminantium primers and C. ruminantium DNA (lane 1) or E. canis DNA (lane 2) or with the E. canis-specific primers and C. ruminantium DNA (lane 4) or E. canis DNA (lane 5). The 123-bp DNA standard ladder and negative control reactions for the C. ruminantium and E. canis primers are in lanes L, 3, and 6, respectively. (B) Southern blot of gel in panel A after hybridization with the pCS20 DNA probe. Lane designations are the same as for panel A.

formed with uninfected A. hebraeum and A. variegatum ticks and purified DNAs from E. canis, T. parva, A. hebraeum, A. variegatum, bovine endothelial cells, and sheep and goat leukocytes. No amplification was detected from 36 A. hebraeum and 34 A. variegatum uninfected adult ticks following Southern blot hybridization with the pCS20 DNA probe, though the DNA-positive ticks tested simultaneously were positive by PCR (data not shown). No amplification was detected from E. canis DNA with C. ruminantium primers or from C. ruminantium DNA with E. canis-specific primers after agarose gel electrophoresis and Southern blot hybridization with the pCS20 DNA probe (Fig. 3). Amplification of a 350-bp product from E. canis DNA by E. canis-specific primers confirmed the identity of the E. canis DNA used for the specificity test (Fig. 3A). No amplification was detected from any of the remaining DNAs tested with C. ruminantium primers (data not shown). The pCS20 DNA probe, however, hybridized weakly to a dot blot of 400 ng of genomic DNA of E. canis, at a level lower than that to 1 ng of C. ruminantium DNA (Fig. 4), and did not hybridize to DH82 cell DNA. Detection of C. ruminantium infection in Amblyomma ticks by PCR. The utility of the PCR assay for detecting C. ruminantium infections in the two major vectors of heartwater, A. hebraeum and A. variegatum, was tested by performing PCRs with adult ticks that had fed as nymphs on sheep acutely infected with different strains of C. ruminantium. Five adult male A. hebraeum ticks infected with Crystal Springs, Mbizi, or Ball-3 strains and five adult male A. variegatum ticks infected

FIG. 4. Hybridization of the pCS20 DNA probe with C. ruminantium DNA (100, 10, 1, or 0.1 ng [lanes 1 to 4, respectively]), E. canis DNA (400, 40, or 4 ng [lanes 1 to 3, respectively]), and DH82 cell DNA (400, 40, or 4 ng [lanes 1 to 3, respectively]).

with the Nigeria or Gardel strain were analyzed. A. hebraeum and A. variegatum ticks that had fed on uninfected sheep were tested concurrently. After DNA probe hybridization to dot blotted PCR products, amplification was detected from all Crystal Springs-, Mbizi-, Ball-3-, and Nigeria-infected ticks and from four of five Gardel-infected ticks (Fig. 5). No amplification was detected from uninfected ticks. C. ruminantium infection was successfully detected by PCR in dried and fixed A. hebraeum ticks that had acquired infection from a clinically ill sheep. All 10 dried ticks, all 10 glutaraldehyde-fixed ticks, all 10 ethanol-fixed ticks, and all 10 formalinfixed ticks were positive by PCR assay after agarose gel electrophoresis. A representative sample of these PCRs is shown in Fig. 6. Comparison of DNA probe and PCR assays for detection of C. ruminantium infection in Amblyomma ticks. PCR and DNA probe assays were performed with DNA samples from adult A. hebraeum ticks that had engorged as nymphs on a sheep clinically ill from C. ruminantium infection (Plumtree strain ovine no. 5112) and from A. hebraeum ticks that had fed on

FIG. 5. PCR detection of C. ruminantium in adult A. hebraeum and A. variegatum ticks fed as nymphs on febrile sheep infected with different strains of C. ruminantium. PCR products were dot blotted and hybridized with the pCS20 DNA probe. (A) A. hebraeum ticks infected with the Crystal Springs isolate (lanes 1 to 5), Mbizi isolate (lanes 6 to 10), and Ball-3 isolate (lanes 11 to 15). Lane 16 contains PCR products from an uninfected adult male A. hebraeum tick processed concurrently. (B) A. variegatum ticks infected with the Nigeria isolate (lanes 1 to 5) and the Gardel isolate (lanes 6 to 10). Lane 11 contains PCR products from an uninfected adult male A. variegatum tick processed concurrently. In each panel, positive- and negative-reaction control PCR products are in lanes 1 and 2, respectively.

170

J. CLIN. MICROBIOL.

PETER ET AL.

TABLE 2. Comparison of pCS20 DNA probe and PCR assays for detecting C. ruminantium infection in A. hebraeum ticks fed on carrier cattle Bovine no. and tick group

No. of ticks positive (% detected)

No. of ticks tested

DNA probe

Bovine no. 4 Male Female Bovine no. 6 male Bovine no. 7 Male Female Total FIG. 6. Detection of C. ruminantium in dried and fixed ticks. Agarose gel electrophoresis of PCR products representative of reactions with infected fresh (lanes 1 and 2), dried (lanes 3 and 4), and glutaraldehyde (lanes 5 and 6)-, ethanol (lanes 7 and 8)-, or formalin (lanes 9 and 10)-fixed A. hebraeum adult ticks. One male tick and one female tick from each category have been included. Lanes 11 and 12 contain PCR products from fresh, uninfected ticks (one male and one female) processed for PCR concurrently. The 123-bp DNA standard ladder and positive- and negative-reaction control PCR products are in lanes L, 1, and 2, respectively.

recovered carrier cattle infected with the Mbizi strain. Separate batches of 30 ticks from these animals transmitted heartwater fatally to susceptible goats, proving that all of these animals were infected. One hundred ticks (50 males and 50 females) from the febrile sheep were analyzed. The DNA probe detected infection in 22 of 50 females (44.0%) and 15 of 50 males (30.0%) (Table 1). The overall infection frequency as determined by the DNA probe was 37.0%. With the same ticks, the PCR assay detected infection in 46 of 50 females (92.0%) and 40 of 50 males (80.0%) (Table 1); the overall infection frequency detected by PCR was 86.0%. All except two of the DNA probe-positive, febrile animal-infected ticks were also positive by PCR. This may have been due to incomplete removal of PCR inhibitors from tick DNA samples. Thirty ticks each from two Mbizi strain-infected carrier bovines (no. 4 and 7) and 15 ticks from a third Mbizi-infected bovine (no. 6) were also tested by both assays. The DNA probe detected no infection in the 30 and 15 ticks from bovines no. 4 and 6, respectively, but detected 1 infected tick from the 30 bovine no. 7 ticks (Table 2). Applied to the same ticks, the PCR assay detected C. ruminantium infection in 4 ticks from bovine no. 4, 1 tick from bovine no. 6, and 16 ticks from bovine no. 7 (Table 2). The average percentages detected for all carrier ticks were 28.0 and 1.33% by PCR and the DNA probe, respectively. All DNA probe-positive carrier ticks were also positive by PCR.

TABLE 1. Comparison of pCS20 DNA probe and PCR assays for detecting C. ruminantium infection in A. hebraeum ticks fed on a clinically ill animal Group

Male Female Total

No. of ticks positive (% detected)

No. of ticks tested

DNA probe

PCR

50 50 100

15 (30.0) 22 (44.0) 37 (37.0)

40 (80.0) 46 (92.0) 86 (86.0)

8 22 15

0 (0) 0 (0) 0 (0)

15 15 75

1 (6.67) 0 (0) 1 (1.33)

PCR

1 (12.50) 3 (13.64) 1 (6.67) 6 (40.00) 10 (66.67) 21 (28.00)

DISCUSSION The data presented here describe the development and evaluation of a PCR assay for detecting C. ruminantium infection in Amblyomma ticks and demonstrate that PCR is necessary for the detection of low-level infections that are below the detection limit of the pCS20 DNA probe. The increased sensitivity of the PCR assay was evident when applied to ticks from both febrile and carrier animals, as it dramatically improved on the DNA probe detection of infection in these ticks. The PCR assay may also supersede older detection methods, such as microscopy and xenodiagnosis, because of its high sensitivity, specificity, and speed, though the proportion of PCR-positive ticks that bear viable, transmittable infections remains to be determined. Nevertheless, the application of PCR should improve the accuracy of estimates of vector infection rates. The PCR assay detected C. ruminantium in both A. hebraeum and A. variegatum ticks infected with geographically distinct strains of heartwater and failed to amplify DNA from batches of uninfected adults of these tick species. The PCR assay was specific for C. ruminantium and did not detect other hemoparasitic tick-transmitted parasites of livestock (reference 31 and this study) or DNA from phylogenetically closely related (56), serologically cross-reactive (22) E. canis organisms. However, weak hybridization of the pCS20 DNA probe with a 400-ng sample of E. canis genomic DNA was observed. Although this further substantiates the close relationship between E. canis and C. ruminantium (56), the weak hybridization of the probe with E. canis is unlikely to affect the detection of C. ruminantium in Amblyomma ticks, as E. canis is not transmitted by these ticks (45). Nevertheless, since the C. ruminantium primers do not amplify DNA from E. canis, this is the method of choice. The DNA probe and PCR analyses of ticks fed on carrier animals demonstrate that these ticks develop lower levels and lower rates of infection than those fed on febrile animals. This is likely due to lower rickettsemia levels in carrier animals (2) and perhaps also to limited replication and cell-to-cell spread from the original foci of infection by C. ruminantium, at least prior to tick feeding. This might have considerable implications for disease severity in areas where heartwater is endemic and ticks are most likely to acquire infection from carrier animals, though it remains to be seen whether lower levels of infection in unfed carrier-infected ticks result in less-acute disease. Studies of other diseases have demonstrated a positive correlation between donor animal parasitemia or viremia and vector infection rate (12, 15, 21, 48, 59). However, the final level of infection within vectors has been reported to be independent


VOL. 33, 1995

of host parasitemia (16, 44), though infection levels in Rhipicephalus appendiculatus with T. parva appear to be dependent on host piroplasm parasitemia, at least in the early stage of disease (59). The development and multiplication of C. ruminantium in Amblyomma ticks may be stimulated by tick feeding or exposure to elevated temperatures (heat shock), as is the case for Anaplasma marginale (23). A preliminary study (61) has suggested that incubation at 378C for 3 days does not influence the final level of infection in adult A. hebraeum ticks, though a more thorough investigation of this aspect is required. The detection of C. ruminantium infection by PCR was possible in both desiccated and preserved ticks. Amplification from fixed or dried ticks permits the analysis of specimens unsuitable for traditional animal inoculation or microscopical tests and allows retrospective diagnostic studies to be performed on archival tick collections from the field. PCRs with DNA extracted from dried and ethanol- or formalin-fixed mammalian and tick tissues have been reported previously (13, 18–20, 39, 40, 52). Our data demonstrate that tick tissue contains PCR-inhibitory substances which can be removed by conventional DNA extraction without precipitation. PCR detection of other tickborne pathogens has been shown to be inhibited by tick tissues or to be optimal only after heat treatment or extraction of sample DNA (20, 49). Although the nature and origin of the PCR-inhibitory elements of Amblyomma tick tissue were not determined, it is suspected that they are derived from residues of the tick’s last blood meal. Electron micrographs have revealed deposits of hemoglobin within midgut cells of unfed Amblyomma ticks (25), and it has been shown that hemoglobin strongly inhibits Thermus aquaticus DNA polymerase (37). In conclusion, we have developed a PCR-based assay that reliably detects C. ruminantium in Amblyomma ticks fed on both acutely infected and carrier animals. This assay supersedes the DNA probe in its sensitivity and specificity. This fact and its ability to detect infections in dried or preserved specimens make it well suited for studies on the epidemiology of heartwater. Future research needs to compare the methods available for the detection of C. ruminantium infection in ticks to determine the most rapid, reliable, and relevant detection assay. ACKNOWLEDGMENTS This work was supported by U.S. Agency for International Development cooperative agreements AFR-0435-A-00-9084-00 and grant no. LAG-1328-G-00-3030-00 and USDA CBAG grant 91-34135-6178. Sharon Deem is a Howard Hughes Medical Institute Predoctoral Fellow. DH82 cells and E. canis (Oklahoma strain) were provided by the Centers for Disease Control and Prevention, Atlanta, Ga. K. Kanhai of the Food and Agriculture Organization kindly provided infected lymphoblastoid cultures of T. parva. We thank Lameck Chakurungama for rearing the ticks and Rensik Karunkomo for animal maintenance. REFERENCES 1. Alleman, R. A., S. M. Kamper, N. Viseshakul, and A. F. Barbet. 1993. Analysis of the Anaplasma marginale genome by pulsed field gel electrophoresis. J. Gen. Microbiol. 139:2439–2444. 2. Andrew, H. R., and R. A. I. Norval. 1989. The carrier status of sheep, cattle and African buffalo recovered from heartwater. Vet. Parasitol. 34:261– 266. 3. Azad, A. F., L. Webb, M. Carl, and G. A. Dasch. 1990. Detection of rickettsia in arthropod vectors by DNA amplification using the polymerase chain reaction. Ann. N. Y. Acad. Sci. 590:557–563. 4. Bezuidenhout, J. D. 1984. Demonstration of Cowdria ruminantium in Amblyomma hebraeum by fluorescent antibody techniques, light and electron microscopy. Onderstepoort J. Vet. Res. 51:213–215.

171

5. Bezuidenhout, J. D. 1987. Natural transmission of heartwater. Onderstepoort J. Vet. Res. 54:349–351. 6. Birnie, E. F., M. J. Burridge, E. Camus, and N. Barre. 1985. Heartwater in the Caribbean: isolation of Cowdria ruminantium from Antigua. Vet. Rec. 116:121–123. 7. Byrom, B., and C. E. Yunker. 1990. Improved culture conditions for Cowdria ruminantium (Rickettsiales), the agent of heartwater disease of domestic ruminants. Cryobiology 4:285–290. 8. Camus, E., and N. Barre. 1987. Epidemiology of heartwater in Guadeloupe and in the Caribbean. Onderstepoort J. Vet. Res. 54:419–426. 9. Camus, E., and N. Barre. 1992. The role of Amblyomma variegatum in the transmission of heartwater with special reference to Guadeloupe. Ann. N. Y. Acad. Sci. 653:33–41. 10. Cowdry, E. V. 1925. Studies on the etiology of heartwater. I. Observation of a rickettsia, Rickettsia ruminantium (n. sp.), in the tissues of infected animals. J. Exp. Med. 42:253–274. 11. Cowdry, E. V. 1925. Studies on the etiology of heartwater. II. Rickettsia ruminantium (n. sp.) in the tissues of ticks transmitting the disease. J. Exp. Med. 42:231–245. 12. Davies, C. R., L. D. Jones, and P. A. Nuttall. 1990. A comparative study of the infection thresholds of Thogoto virus in Rhipicephalus appendiculatus and Amblyomma variegatum. Am. J. Trop. Med. Hyg. 43:99–103. 13. De Kok, J. B., C. d’Oliveira, and F. Jongejan. 1993. Detection of the protozoan parasite Theileria annulata in Hyalomma ticks by the polymerase chain reaction. Exp. Appl. Acarol. 17:839–846. 14. Du Plessis, J. L. 1985. A method for determining the Cowdria ruminantium infection rate of Amblyomma hebraeum: effects in mice injected with tick homogenates. Onderstepoort J. Vet. Res. 52:55–61. 15. Eriks, I. S., D. Stiller, and G. H. Palmer. 1993. Impact of persistent Anaplasma marginale rickettsemia on tick infection and transmission. J. Clin. Microbiol. 31:2091–2096. 16. Gage, K. L., R. D. Gilmour, R. H. Karstens, and T. G. S. O. Schwan. 1992. Detection of Rickettsia rickettsii in saliva, haemolymph and triturated tissues of infected Dermacentor andersoni ticks by polymerase chain reaction. Mol. Cell. Probes 6:333–341. 17. Haig, D. A. 1952. Note on the use of the white mouse for the transport of strains of heartwater. J. S. Afr. Vet. Med. Assoc. 23:167–170. 18. Honma, M., Y. Ohara, H. Murayama, K. Sako, and Y. Iwasaki. 1993. Effects of fixation and varying target length on the sensitivity of polymerase chain reaction for detection of human T-cell leukemia virus type I proviral DNA in formalin-fixed tissue sections. J. Clin. Microbiol. 31:1799–1803. 19. Impraim, C. C., R. K. Saiki, H. A. Erlich, and R. L. Teplitz. 1987. Analysis of DNA extracted from formalin-fixed, paraffin embedded tissues by enzymatic amplification and hybridization with sequence-specific oligonucleotides. Biochem. Biophys. Res. Commun. 142:710–716. 20. Johnson, J. B., C. M. Happ, W. M. Mayer, and J. Piesman. 1992. Detection of Borrelia burgdorferi in ticks by species-specific amplification of the flagellin gene. Am. J. Trop. Med. Hyg. 47:730–741. 21. Kamio, T., K. Fujisaki, and T. Minami. 1989. Correlation between the infection rate of the vector tick, Haemaphysalis longicornis and the parasitemia of cattle infected with Theileria sergenti. Ann. Trop. Med. Parasitol. 83:77–83. 22. Kelly, P. J., L. A. Matthewman, S. M. Mahan, S. Semu, T. Peter, P. R. Mason, P. Brouqui, and D. Raoult. 1994. Serological evidence for antigenic relationships between Ehrlichia canis and Cowdria ruminantium. Res. Vet. Sci. 56:170–174. 23. Kocan, K. M., S. J. Barron, D. Holbert, S. A. Ewing, and J. A. Hair. 1982. Influence of increased temperature on Anaplasma marginale Theiler in the gut of Dermacentor andersoni Stiles. Am. J. Vet. Res. 43:32–35. 24. Kocan, K. M., and J. D. Bezuidenhout. 1987. Morphology and development of Cowdria ruminantium in Amblyomma ticks. Onderstepoort J. Vet. Res. 54: 177–182. 25. Kocan, K. M., J. D. Bezuidenhout, and A. Hart. 1987. Demonstration and ultrastructural features of Cowdria ruminantium in midgut and salivary glands of nymphal Amblyomma hebraeum. Onderstepoort J. Vet. Res. 54: 87–93. 26. Kocan, K. M., S. P. Morzaria, W. P. Voigt, J. Kiarie, and A. P. Irvin. 1987. Demonstration of colonies of Cowdria ruminantium in midgut epithelial cells of Amblyomma variegatum. Am. J. Vet. Res. 48:356–360. 27. Lawrence, J. A., and P. K. I. MacKenzie. 1980. Isolation of a non-pathogenic Theileria of cattle transmitted by Rhipicephalus appendiculatus. Zimb. Vet. J. 11:27–35. 28. Logan, L. L., T. C. Whyard, J. C. Quintero, and C. A. Mebus. 1987. The development of Cowdria ruminantium in neutrophils. Onderstepoort J. Vet. Res. 54:197–204. 29. Mahan, S. M., G. E. Smith, and B. Byrom. 1994. Concanavalin A-stimulated bovine T-cell supernatants inhibit growth of Cowdria ruminantium in bovine endothelial cells in vitro. Infect. Immun. 62:747–750. 30. Mahan, S. M., N. Tebele, D. Mukwedeya, S. Semu, C. B. Nyathi, L. A. Wassink, P. J. Kelly, T. Peter, and A. F. Barbet. 1993. An immunoblotting diagnostic assay for heartwater based on the immunodominant 32-kilodalton

172

31.

32. 33. 34.

35. 36. 37. 38.

39. 40. 41. 42. 43. 44. 45. 46.

PETER ET AL. protein of Cowdria ruminantium detects false positives in field sera. J. Clin. Microbiol. 31:2729–2737. Mahan, S. M., S. D. Waghela, T. C. McGuire, F. R. Rurangirwa, L. A. Wassink, and A. F. Barbet. 1992. A cloned DNA probe for Cowdria ruminantium hybridizes with eight heartwater strains and detects infected sheep. J. Clin. Microbiol. 30:981–986. Matthewman, L. A., K. Sumption, P. J. Kelly, N. Nally, and D. Raoult. Unpublished data. Mullis, K. B., and F. A. Faloona. 1987. Specific synthesis of DNA in vitro via a polymerase catalyzed chain reaction. Methods Enzymol. 155:335–350. Neitz, A. W. H., G. J. Viljoen, J. D. Bezuidenhout, P. T. Oberem, W. Van Wyngaardt, and N. M. J. Vermeulen. 1986. The detection of antibodies to Cowdria ruminantium in serum and C. ruminantium antigen in Amblyomma hebraeum by means of an enzyme-linked immunosorbent assay. Onderstepoort J. Vet. Res. 53:39–41. Norval, R. A. I., H. R. Andrew, and C. E. Yunker. 1990. Infection rates with Cowdria ruminantium of nymphs and adults of the bont tick collected in the field in Zimbabwe. Vet. Parasitol. 36:277–282. Oberem, P. T., and J. D. Bezuidenhout. 1987. The production of heartwater vaccine. Onderstepoort J. Vet. Res. 54:485–488. Panaccio, M., and A. Lew. 1991. PCR based diagnosis in the presence of 8% (v/v) blood. Nucleic Acids Res. 19:1151. Perreau, P., P. C. Morel, N. Barre, and P. Durand. 1980. Existence de la cowdriose (heartwater) a Cowdria ruminantium, chez les petits ruminants des Antilles Francaises (la Guadeloupe) et des Mascareignes (La Reunion et Ile Maurice). Rev. Elev. Med. Vet. Pays Trop. 33:21–22. Persing, D. H., S. R. Telford, P. N. Rys, D. E. Dodge, T. J. White, S. E. Malawista, and A. S. O. Spielman. 1990. Detection of Borrelia burgdorferi DNA in museum specimens of Ixodes dammini ticks. Science 248:1420–1423. Persing, D. H., S. R. Telford III, A. Spielman, and S. W. Barthold. 1990. Detection of Borrelia burgdorferi infection in Ixodes dammini ticks with the polymerase chain reaction. J. Clin. Microbiol. 28:566–572. Provost, A., and J. D. Bezuidenhout. 1987. The historical background and global importance of heartwater. Onderstepoort J. Vet. Res. 54:165–169. Prozesky, L., and J. L. Du Plessis. 1987. Heartwater. The development and life cycle of Cowdria ruminantium in the vertebrate host, ticks and cultured endothelial cells. Onderstepoort J. Vet. Res. 54:193–196. Purchase, H. S. 1945. A simple and rapid method for demonstrating Rickettsia ruminantium (Cowdry, 1925) in heartwater brains. Vet. Rec. 57: 413–414. Purnell, R. E., M. A. Ledger, P. L. Omwoyo, R. C. Payne, and M. A. Peirce. 1974. Theileria parva: variation in the infection rate of the vector tick, Rhipicephalus appendiculatus. Int. J. Parasitol. 4:513–517. Ristic, M., and D. L. Huxsoll. 1984. Tribe II. Ehrlichieae Philip 1957, 948AL, p. 704–709. In N. R. Krieg and J. G. Holt (ed.), Bergey’s manual of systematic bacteriology, vol. 1. The Williams & Wilkins Co., Baltimore. Saiki, R. K., D. H. Gelfand, S. Stoffel, S. J. Scharf, R. Higuchi, G. T. Horn, K. B. Mullis, and H. A. Ehrlich. 1988. Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239:487–491.

J. CLIN. MICROBIOL. 47. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 48. Singh, K. R., and C. R. Anderson. 1968. Relation of Haemaphysalis spinigera larval infection rates and viremia levels of Kyasanur Forest disease virus. Indian J. Med. Res. 56:137–141. 49. Stich, R. W., J. A. Bantle, K. M. Kocan, and A. Fekete. 1993. Detection of Anaplasma marginale (Rickettsiales, Anaplasmataceae) in haemolymph of Dermacentor andersoni (Acari, Ixodidae) with the polymerase chain reaction. J. Med. Entomol. 30:781–788. 50. Synge, B. A. 1978. Brain biopsy for the diagnosis of heartwater. Trop. Anim. Health Prod. 10:45–48. 51. Tassanakajon, A., V. Boonsaeng, P. Wilairat, and S. Panyim. 1993. Polymerase chain reaction detection of Plasmodium falciparum in mosquitoes. Trans. R. Soc. Trop. Med. Hyg. 87:273–275. 52. Tobolewski, J., M. J. Kaliszewski, R. K. Colwell, and J. H. Oliver. 1992. Detection and identification of mammalian DNA from the gut of museum specimens of ticks. J. Med. Entomol. 29:1049–1051. 53. Uilenberg, G. 1983. Heartwater (Cowdria ruminantium infection): current status. Adv. Vet. Sci. Comp. Med. 27:427–480. 54. Uilenberg, G., E. Camus, and N. Barre. 1985. A strain of C. ruminantium isolated in Guadeloupe (French West Indies). Rev. Elev. Med. Vet. Pays Trop. 38:494–498. 55. Van de Pypekamp, H. E., and L. Prozesky. 1987. Heartwater. An overview of the clinical signs, susceptibility and differential diagnosis of the disease in domestic ruminants. Onderstepoort J. Vet. Res. 54:263–266. 56. van Vliet, A. H. M., F. Jongejan, and B. A. M. van der Zeijst. 1992. Phylogenetic position of Cowdria ruminantium (Rickettsiales) determined by analysis of amplified 16S ribosomal DNA sequences. Int. J. Syst. Bacteriol. 42:494–498. 57. Van Winkelhoff, A. J., and G. Uilenberg. 1981. Heartwater: cross immunity studies with strains of Cowdria ruminantium isolated in West and South Africa. Trop. Anim. Health Prod. 13:160–164. 58. Waghela, S. D., F. R. Rurangirwa, S. M. Mahan, C. E. Yunker, T. B. Crawford, A. F. Barbet, M. J. Burridge, and T. C. McGuire. 1991. A cloned DNA probe identifies Cowdria ruminantium in Amblyomma variegatum ticks. J. Clin. Microbiol. 29:2571–2577. 59. Young, A. S., S. Morzaria, T. Dolan, and F. Mwakima. 1992. Development of Theileria parva in the tick, p. 11–12. In ILRAD 1992 Annual Scientific Report. International Laboratory for Research on Animal Diseases, Nairobi, Kenya. 60. Yunker, C. E., K. M. Kocan, R. A. I. Norval, and M. J. Burridge. 1987. Distinctive staining of colonies of Cowdria ruminantium in midguts of Amblyomma hebraeum. Onderstepoort J. Vet. Res. 54:183–185. 61. Yunker, C. E., S. M. Mahan, S. D. Waghela, T. C. McGuire, F. R. Rurangirwa, A. F. Barbet, and L. A. Wassink. 1993. Detection of Cowdria ruminantium by means of a DNA probe, pCS20, in infected bont ticks, Amblyomma hebraeum, the major vector of heartwater in Southern Africa. Epidemiol. Infect. 110:95–104.