INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Human granulocytic ehrlichiosis (HGE) is a severe or potentially fatal rickettsial infection which is emerging in the upper midwestern and northeastern United States (1, 3, 5). The disease is characterized by fever, severe headaches, and myalgia, as well as occasionally other signs, including respiratory compromise, gastrointestinal disturbance, organ failure, and possibly increased susceptibility to opportunistic infections such as candidiasis (6). There is a greater risk of severe disease in older patients and the case fatality rate has been estimated to be as high as 5% in some populations (7). HGE is caused by an unnamed ehrlichial pathogen which is serologically and morphologically indistinguishable from E. equi, the agent of equine granulocytic ehrlichiosis (EGE) (14).

EGE typically presents as fever, dependent edema, icterus, and ataxia (13). Infections were initially described in horses in northern California but are also found throughout the United States in habitats infested with the tick vectors Ixodes pacificus west of the Rocky Mountains and I. scapularis in the eastern United States (4, 11). The same ticks have been found to be vectors of the HGE agent. Experimental inoculation of horses with the HGE strain produced a clinical syndrome which was indistinguishable from naturally occurring EGE (16).

We and others have hypothesized that the agent of HGE is in fact Ehrlichia equi. This hypothesis was tested in the current study by comparing the sequences of three genes from horse isolates and two recent human cases in northern California (8).

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Field samples. Table 1 shows the origins of the E. equi isolates obtained from Sierra, Mendocino, Sonoma, and Marin counties and the HGE agent isolates obtained from Humboldt county (8) in California (Fig. 1). Briefly, Lily (CAMELI) was a 12-year-old quarterhorse mare from Mendocino county with icterus, petechial hemorrhages, lethargy, and a rectal temperature at the time of examination of 41.7°C. Lea Jubilee (CASOLJ) was a 10-year-old mare with mild edema bilaterally at the cannon-fetlock region, fever, icterus, and slight petechiae apparent on the vaginal mucosa and sclera. E. equi morulae were observed in a buffy coat smear stained with Giemsa. Clinical hematology indicated thrombocytopenia, anemia, and leukopenia. Blood samples from the remaining horses with clinical signs of EGE were collected during 1996 to 1998 in northern California. The blood samples were tested for E. equi by 16S rRNA nested PCR (6), and the DNA samples were stored until use. Samples from the human patients were obtained 1998 as described previously (8).

View larger version (23K): [in this window] [in a new window]   FIG. 1.   Map of California showing the counties where equine and human isolates were collected for PCR to detect E. equi () and HGE agent ().

DNA preparation. Genomic DNA was extracted from the buffy coat ACD-anticoagulated whole blood of horses as previously described (4). Briefly, the buffy coat was removed after centrifugation and kept at 20°C overnight. The erythrocytes in the buffy coat were lysed by adding 6 ml of 0.2% NaCl for 5 min and then adding 6 ml of 1.2% NaCl. The cell mixture was washed with phosphate-buffered saline, and the final pellet was diluted in 50 to 200 µl of lysis buffer (10 mM Tris-HCl, pH 8.3; 0.45% NP-40, 0.45% Tween 20, 100 µg of proteinase K per ml) and incubated in a 56°C water bath for 3 h. Finally, the DNA samples were incubated at 97°C for 15 min for inactivation of proteinase K and denaturation. The isolation of DNA from EDTA-anticoagulated blood from human cases was as previously described by Foley et al. (8). Briefly, DNAs were extracted from 100 µl of whole blood by using lysis buffer (10 mM Tris-HCl, pH 8.0; 1 mM EDTA; and 1% [wt/vol] sodium dodecyl sulfate) at 37°C for 1 h. The lysate was incubated at 37°C for 30 min with RNase A (10 µg/ml) and then incubated overnight at room temperature with proteinase K (50 µg/ml). Genomic DNAs were purified with use of Phase Lock Gel I Light (5 Prime  3 Prime, Boulder, Colo.). The positive control for PCR was BDS strain HGE agent (2); the negative control DNA was obtained from an E. equi-free horse.

PCR amplification. The nested-PCR amplification for 16S rRNA gene fragment was done as previously described (4) (Table 2). Outer primers were used EE1 and EE2 for the first-round product (1,433 bp) and inner primers were used EE3 and EE4 for second-round product (928 bp). Primers EE5F and EE6R were designed to obtain an additional 517-bp fragment at the 3' end of the 16S rRNA gene (Table 2). The forward primer LA6 and reverse primer LA1 were used to amplify the 444 Ep-ank gene fragment in the ankyrin repeat region (Table 2) (P. Caturegli, K. Asanovich, J. Walls, J. Bakken, J. Madigan, and J. Dumler, submitted for publication). Cycling conditions were denaturation for 4 min at 94°C, followed by 94°C for 30 s, 62°C for 30 s, and 72°C for 30 s. The annealing temperature was stepped down four times by 2°C every two cycles. The final annealing temperature used was 54°C for 28 cycles, followed by a final extension for 5 min at 72°C.

Cloning and sequencing. The 16S rRNA or groESL gene amplicons were ligated into the plasmid vector pCR 2.1-TOPO and TOP10 One Shot competent cells transformed according to the manufacturer's recommendations (TOPO Cloning Kit; InVitrogen, San Diego, Calif.). The recombinant clones were verified by colony PCR amplification. Two clones of each isolate were arbitrarily chosen for sequencing the forward and reverse strands. Plasmid DNA for sequencing was prepared by using the Quantum Plasmid Miniprep Kit (Bio-Rad, Hercules, Calif.) according to manufacturer's instructions. Amplicons of 444 Ep-ank and the 3' end of 16S rRNA gene fragments were prepared for direct sequencing by using the QIAquick PCR Purification Kit (Qiagen, Hilden, Germany).

Nucleotide and translated amino acid sequence analyses. The sequence data were collected using ABI Prism Data Collection software (version 2.1), and the sequence data were analyzed using ABI Prism Sequence analysis software (version 2.1.1) and Chromas software (version 1.51) (Technelysium Pty., Ltd., Mt. Gravatt Plaza, Queensland, Australia).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

The 16S rRNA (two fragments of 928 and 517 bp), 444 Ep-ank (444 bp), and groESL (1,715 bp) genes were amplified from 13 GE isolates; 11 E. equi isolates from Sierra, Mendocino, Sonoma, and Marin counties; and two HGE agent isolates from Humboldt county in California (Fig. 1 and 2). Amplicons of 6 of the 13 isolates (1 isolate from each county) were selected from E. equi and compared with the two HGE agent isolates. The two HGE agent isolates had identical 16S rRNA gene sequences (1,395 bp) and were identical to the equine isolates from Mendocino and Marin counties. Two other equine isolates, CASITL and CASOLJ, have a C and G in nucleotide positions 34 and 180, respectively (Table 3). In comparison to the U.S. isolate sequences, which have an A at position 40, European isolates from horse, dog, cat, and human sources have a G at the same position. The overall sequence homology in the 16S rRNA gene among E. equi, the HGE agent, and E. phagocytophila is 99.9%.

View larger version (23K): [in this window] [in a new window]   FIG. 2.   PCR amplicons of 16S rRNA (928 bp), 444 Ep-ank (444 bp), and groESL heat shock operon (1,715 bp) genes from naturally infected with granulocytic ehrlichiae in horses from Sierra (CASIGR, CASIPE, CASIJA, CASICH, CASIRN, and CASITL), Mendocino (CAMEBS, CAMELI, and CAMEDO), Sonoma (CASOLJ), and Marin (CAMAWI) counties, California, and in a human from Humboldt county, California (CAHU-HGE1 and CAHU-HGE2). M, Molecular marker; C, double-distilled water; P, positive HGE agent control.

The sequences of 444 Ep-ank from all four E. equi isolates and both HGE isolates were 100% identical (Table 4). The 444 Ep-ank nucleotide sequences from California isolates differed from the BDS strain (GenBank accession number AF047897) from the eastern United States at positions 326 and 398, leading to predicted amino acid changes at positions 109 and 133 (Table 4). The California isolates differed from the USG3 strain (GenBank accession number AF020521) (from Rhode Island) in 16 nucleotide positions, leading to 9 predicted amino acid differences (Table 4).

The groESL heat shock operon gene sequences of CASITL, CAMEBS, CAMAWI, and CAHU-HGE1 isolates were identical, but the sequences from CASOLJ and CAHU-HGE2 isolates had four and two nucleotide differences, respectively (Table 5). This nucleotide sequence difference was not predicted to result in amino acid differences between CAHU-HGE1 and CAHU-HGE2 but would result in expected amino acid differences at three positions when compared with CASOLJ. Similarly, up to 12 nucleotide differences were detected in comparing these northern California isolates with other HGE and E. phagocytophila isolates, but none except CASOLJ would be predicted to alter amino acid translation (Table 5).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We and others have suggested that the E. phagocytophila genogroup is a tightly related group of strains of ehrlichiae that share tick vectors, cross host species, and should be differentiated primarily on the basis of geographical and not host species origin. There are numerous lines of evidence suggesting that the HGE agent is a strain within this group.

The HGE agent is antigenically indistinguishable from E. equi by immunofluorescence assay and on Western blots (2, 7). The HGE agent and E. equi are geographically and seasonally coincident and are transmitted by the same tick. They are indistinguishable ultrastructurally, cause indistinguishable clinical conditions in animal models (9, 12, 16), and are cross-protective (19). Munderloh et al. (17) reported that, based on PCR and immunocytology analyses, the EGE agent isolated in tick cell culture was closely related or identical to the HGE agent.

Evidence from genetic analyses of E. equi and HGE agent strains supports the theory that the ehrlichiae are conspecific. Our group previously showed that there was variability in the DNA sequence of the 16S rRNA gene among tick-exposed horses in California (18). Also, equine isolates in New England have 16S rRNA gene sequences which are identical to the HGE agent (14). Our present results indicate that the 16S rRNA gene sequences from Mendocino (CAMEBS) and Marin (CAMAWI) county horse isolates were identical to two HGE isolates. However, other isolates had one nucleotide difference, and a previously isolated EGE from a horse (Alice) in California (18) was identical to CAMEBS, CAMAWI, and the two human isolates from California in the homologous fragment sequences available for analysis. There was a suggestion that geographical origin could account for some of the observed variability, in that horse and human isolates from California and E. phagocytophila (ovine and caprine) from Scotland had an A in position 40 (Table 3), but isolates of human, mice, and tick origin from the eastern and middle United States had a G in that position. However, DNA of horse, dog, cat, and tick origin from Europe also contained a G at position 40. The microheterogeneity consisting of one or two unique nucleotide changes among E. equi strains, HGE agent, and E. phagocytophila strains suggests that it is not possible to classify isolates solely on the basis of their 16S rRNA gene nucleotide sequences.

Comparison of DNA sequences of the groESL operon in the E. phagocytophila genogroup also indicated that variability was associated with geographic origin. In a previous study which also examined groESL variability, found slightly more variability was found in the 16S rRNA gene and European origin isolates tended to differ from isolates from the United States (21). DNA sequencing of the 444 Ep-ank gene also revealed that the E. equi and HGE agent isolates from northern California were identical but that California isolates differed significantly from the eastern U.S. BDS and NCH-1 strains. Moreover, there was predicted protein diversity in this gene product between western and eastern U.S. isolates. Another reasonable target for genetic analysis would be the 44-kDa major outer membrane protein of the HGE agent and E. equi (22).

These results for 16S rRNA and for 444 Ep-ank and groESL heat shock operon gene sequences suggest that the CAHU-HGE1 isolate and E. equi strains from Mendocino and Marin counties are invariant and that at least some HGE agent strains are conspecific with E. equi. However, the species designation should also incorporate data regarding infectivity or pathogenicity, and such studies are incomplete. Experimental manipulation of the equine model supports the hypothesis that HGE isolates do not differ in pathogenicity from strains of E. equi (16).

Our results indicate that the HGE agent and E. equi are conspecific, which is a very important finding because the risk of HGE to humans cannot be adequately evaluated without considering infections in domestic animals. HGE is an emerging infection in people, partly because of increased scrutiny, and yet E. equi is well established, suggesting that the banks of tick vectors in areas where horse, dog, and hoofstock disease is endemic are contaminated and pose risks to people in the area. In a seroepidemiological survey in 1985 and 1986 for E. equi in northern California, 10.4% of 335 horses were found to be positive for antibodies and 38 horses were diagnosed as having EGE based on clinical signs and the observation of inclusion bodies in neutrophils in blood smears (15). Antibodies to E. equi have been detected in dogs from Oklahoma (20), the midwestern United States (10), and California (J. Foley, unpublished data). Thus, surveillance for human disease will need to incorporate dynamics of infections in domestic animals as well.