Copyrightq1997, American Society for Microbiology
Detection of Colorado Tick Fever Virus by Using Reverse
Transcriptase PCR and Application of the Technique in
Laboratory Diagnosis
ALISON J. JOHNSON,* NICK KARABATSOS,ANDROBERT S. LANCIOTTI
Division of Vector-Borne Infectious Diseases, National Center for Infectious Diseases, Centers for Disease Control and Prevention, Public Health Service, U.S. Department of Health and Human Services, Fort Collins, Colorado 80522
Received 18 November 1996/Returned for modification 31 December 1996/Accepted 17 February 1997
Colorado tick fever (CTF) virus elicits an acute illness in humans, producing nonspecific flu-like symptoms and a biphasic fever in approximately 50% of patients. The disease is transmitted by the adult Rocky Mountain wood tick (Dermacentor andersoni), and therefore incidence is limited by the habitat and life cycle of that vector. The early symptoms of infection are difficult to distinguish from those of several other agents, especially Rickettsia rickettsii. Serologic testing is usually unable to provide evidence of CTF viral infection during the acute phase because of the late appearance of the various antibodies. Here we report the development and clinical application of a test to diagnose this disease during the acute stages. Oligonucleotide primers to the S2 segment of CTF (Florio) virus were made, and these were used in the amplification of a 528-bp fragment of DNA, transcribed from the double-stranded CTF virus RNA template by reverse transcriptase PCR. RNAs processed from 16 CTF virus isolates yielded similar results when analyzed on agarose gels. These were distinguishable from their antigenic relatives Eyach, S6-14-03, and T5-2092 and from other coltiviruses and an orbivirus but not from the antigenically distinct CTF virus-related isolate 720896. A mouse model demon-strated the utility of this method with whole-blood specimens, and CTF virus was successfully detected in human sera from the initial day of the onset of symptoms to 8 days later. The reverse transcriptase PCR method is a promising tool for the early diagnosis of CTF viral infection, or for ruling out CTF virus as the etiologic agent, in order to facilitate appropriate medical support.
Colorado tick fever (CTF) virus is the prototype species for the genus Coltivirus in the Reoviridae family, with a viral ge-nome consisting of 12 segments of double-stranded (ds) RNA (7). It is transmitted to mammals, including humans, princi-pally by the adult Rocky Mountain wood tick (Dermacentor
andersoni), and exposure to the virus is therefore restricted to
the vector habitat (20). Clinical cases peak between May and July (5). Historically, CTF has been the most frequently re-ported arboviral disease in the United States (15). Neverthe-less, it is considered to be an underdiagnosed condition. Viral maintenance is achieved via larval and nymphal stages of D.
andersoni and various rodent species (20).
CTF is a self-limiting illness. The acute phase typically lasts 5 to 10 days, followed by a convalescence, which may be pro-tracted, especially in adults. Although the mean incubation time from tick exposure to the onset of symptoms is about 4 days, the range is from,1 to 14 days (13). Disease diagnosis based on clinical signs and symptoms is rare, given their non-specific nature. Onset is sudden, with fever and chills, myalgia, photophobia, joint and retro-orbital pain, stiff neck, headache, malaise, and lethargy. Patients may also present with other, less consistent ailments, including a biphasic fever (,50% fre-quency), abdominal pain (20% frefre-quency), and transient rash (10% frequency) (8). Hospitalization occurs mostly among children, in whom severe complications have occasionally arisen (7).
The disease is considered to be underdiagnosed for a variety
of reasons. When the history of a tick bite is absent, a physician may overlook the possibility of CTF viral infection. Moreover, the presenting symptoms are easily confused with other dis-eases, such as Rocky Mountain spotted fever (RMSF) and relapsing fever. Finally, to date there has been no timely diag-nostic test available to identify CTF during the acute phase. A number of techniques are available for the diagnosis of CTF in humans; however, there are disadvantages associated with each. The late appearance of antibodies in the serum precludes any positive results from indirect fluorescent antibody, neutral-ization, and complement fixation tests (7) or from enzyme-linked immunosorbent assays (ELISA) (9) for about 10 to 14 days after the onset of symptoms. Immunofluorescent staining of erythrocytes, while suitable for early detection of CTF viral infection (12), is inherently subjective, and virus isolation (VI) from patient erythrocytes is time-consuming (9). State and local health laboratories in the Rocky Mountain region which perform diagnostic testing for CTF virus assign a positive result to patient sera having at least a fourfold rise in either indirect fluorescent antibody or complement fixation titer from acute-to convalescent-phase samples. Currently, diagnostic tests for other conditions with which CTF may initially be confused are unavailable for the early stages of those diseases. Hence, the development of a sensitive and objective diagnostic test for acute CTF viral infection would be a useful tool for the med-ical community.
Reverse transcriptase (RT) PCR provides a powerful method of detection for RNA viruses and is gaining much popularity as a diagnostic technique. Its main advantages are its short completion time and sensitivity compared to those of other assays (10, 17, 18). Reoviruses bluetongue virus and African horse sickness virus have recently been successfully detected by RT-PCR (1, 11, 21). Presented here are data on
* Corresponding author. Mailing address: Division of Vector-Borne Infectious Diseases, National Center for Infectious Diseases, Centers for Disease Control and Prevention, P.O. Box 2087, Fort Collins, CO 80522. Phone: (970) 221-6469. Fax: (970) 221-6476. E-mail: AJJ1 @CDC.GOV.
1203
on May 15, 2020 by guest
http://jcm.asm.org/
the adaptation of this technique for detection of CTF and related viruses and its utility when used to diagnose human clinical infection.
MATERIALS AND METHODS
Virus strains and controls.Sixteen CTF isolates, four related viruses (720896, Eyach, S6-14-03, and T5-2092), and bluetongue virus serotype 17 (Table 1) were obtained from the collection at the Division of Vector-Borne Infectious Dis-eases, Centers for Disease Control and Prevention (CDC), Fort Collins, Colo., as was Venezuelan equine encephalitis virus, vaccine strain TC-83 (VEE TC-83), which was used as a control. The Indonesian coltiviruses (Jakarta [JKT] virus strains 6423, 6969, and 7043) (6) were supplied as a gift from Robert E. Shope. Viruses were replicated in Vero or BHK-21 cells with minimal essential medium and 5% fetal calf serum, with the exception of the JKT viruses, which were replicated in C6/36 insect cells in Dulbecco’s minimal essential medium with 10% fetal calf serum. Culture supernatants were clarified by centrifugation at 2,0003 g for 10 min. D. McKechnie kindly provided us with Rickettsia rickettsii DNA,
which was derived from a rickettsial culture.
RNA extraction.Viral RNA was extracted from 140ml of each clarified culture supernatant by using the QIAamp viral RNA kit (QIAGEN, Inc., Chatsworth, Calif.) according to the manufacturer’s instructions and eluted into 50ml of RNase-free water.
Oligonucleotide primers.The complete sequence of the S2 segment of CTF (Florio strain) virus, which encodes the smallest viral protein (VP12), was ob-tained from GenBank (accession no. U53227). From this, primers of the follow-ing sequences and genome positions were designed with Lasergene (DNAstar, Madison, Wis.) and synthesized on an Applied Biosystems (Foster City, Calif.) model 394 synthesizer.
The forward primer was the 21-mer 59GAGTGTCGCCGGGTTTTTGAA 39 (nucleotides 136 to 156). The reverse primer was the 21-mer 59CGTCCCGGG AGAATGATGCTA 39(nucleotides 643 to 663). The nested reverse primer was the 27-mer 59ATGTGAGGAGAGGCTGGCGGAGGATAG 39(nucleotides 466 to 492).
RT-PCR.Denaturation of ds RNA was achieved by heating 3.3ml of RNA with 1.7ml of 1.45% formamide to 958C for 5 min (23). For the RT reaction, 15 ml of a reagent cocktail was added to each sample for a total volume of 20ml. Final concentrations were 10 mM dithiothreitol, 200mM each deoxynucleotide triphosphate (dNTP), 103U of RNase inhibitor (Boehringer Mannheim
Bio-chemicals, Indianapolis, Ind.) perml, 5mM each forward and reverse primers in 50 mM Tris-HCl (pH 8.3), 75 mM KCl, and 3 mM MgCl2(13first-strand buffer;
GIBCO BRL, Grand Island, N.Y.). RTs Superscript II (GIBCO BRL) and RAV-2 (Amersham Corp., Arlington Heights, Ill.) were used interchangeably at concentrations of 104and 100 U/ml, respectively. Reaction mixtures were
main-tained at 458C for 30 min. For PCR amplification, the RT reaction mixtures were
diluted to 100ml with an additional 16 nmol of dNTPs and 2.5 U of AmpliTaq DNA polymerase (The Perkin-Elmer Corp., Norwalk, Conn.) in 1.5 mM MgCl2,
50 mM KCl, 10 mM Tris-HCl (pH 8.3), and 0.001% (wt/vol) gelatin (13PCR buffer; Perkin-Elmer). PCR amplification using a GeneAmp PCR System 9600 (Perkin-Elmer) was achieved with an initial denaturation (948C, 5 min) followed by 35 successive cycles, each consisting of denaturation (948C, 30 s), annealing (558C, 1 min), and extension (688C, 3 min). A final extension (688C, 12 min) concluded the thermal profile, and samples were held at 48C. To confirm the specificities of the RT-PCR products, DNA was purified with a QIAquick gel extraction kit (Qiagen, Inc.) and nested PCR was performed. DNA was diluted 1:100, and 5ml was added to 95ml of a reaction cocktail. Final concentrations were 5mM each forward and nested reverse primers, 0.16 mM dNTPs, and 25 U of AmpliTaq DNA polymerase per ml in 13PCR buffer. Thermal cycling for the nested reactions was the same as for the first-round PCR, except that the number of cycles was reduced to 20.
Agarose gel electrophoresis.Ten microliters of each product was analyzed on a 1.5% agarose (Kodak, New Haven, Conn.) gel containing ethidium bromide and compared to molecular mass markers (low DNA mass ladder; Gibco BRL). Sensitivity.CTF virus strains Florio, R6225, 64V37, and 76-1-1007; antigenic relatives Eyach and S6-14-03; and bluetongue virus 17 were plaque titrated on six-well Vero cell plates as previously described (15) with 10-fold dilutions of clarified culture supernatant. To determine the relative sensitivity of the RT-PCR method, RNAs were extracted from similar 10-fold virus dilutions and RT-PCR was performed in the manner described above.
Detection of CTF virus in experimentally infected mice.Five outbred CDC colony mice (ICR CD-1), 4 weeks of age, were inoculated intraperitoneally with approximately 5,000 PFU of CTF (Florio) virus in 0.25 ml of phosphate-buffered saline. Mice were bled at the time of inoculation and marked for identification. Each animal was subsequently bled by retro-orbital puncture every second or third day postinoculation (p.i.) (Table 2). Sample tubes contained 40ml of 0.25 M EDTA (pH 8.5) (about 1/10 the volume of blood), with which the blood was immediately mixed upon collection to prevent clotting. The samples were washed twice with an equal volume of phosphate-buffered saline, and RNA was extracted from 140 ml of a 1:2 dilution of whole blood, as described above, and then amplified with the CTF virus primers and analyzed on agarose gels. Normal whole mouse blood from an uninfected animal was processed in the same manner and used as a negative control. For comparative VI, samples were diluted 1:5 in BA-1 medium (1% bovine albumin in M199 with Earle’s balanced salts) and homogenized with 20 strokes of a Tenbroeck apparatus. Vero cell six-well plates were inoculated with 100ml of the homogenate and successive 10-fold dilutions as described above.
[image:2.612.65.554.82.328.2]Sequence analysis of amplified DNA.On day 1 p.i., 70ml of amplified DNAs from CTF (Florio) virus (approximately 528 bp) and from mouse 4 (approxi-mately 420 bp) were excised from a 1.5% SeaKem (FMC Bioproducts, Rockland, Maine) agarose preparative gel. Fragments were purified with a Geneclean II kit
TABLE 1. Isolation specifics and RT-PCR amplification results with CTF virus strains and related viruses
Virus strain Source Location Amplicon size (bp) first-round PCR/nested PCR
Florio Human Colorado 528/356
720896 Human Idaho 528/356
R1575 Human Colorado 528/356
R11515 Human Colorado 528/356
69V28 Human Colorado 528/356
R19372 Human Colorado 528/356
R6225 Human Colorado 528/356
R19420 Human New Mexico 528/356
71V11 Human Colorado 528/356
Eyach Ixodes ricinus Germany 425/URa
77V5270 D. andersoni Colorado 528/356
64V37 D. andersoni Colorado 528/356
75V1906 D. andersoni Colorado 528/356
75V3863 D. andersoni Colorado 528/356
75V1843 D. andersoni Colorado 528/356
76-1-1007 Spermophilus lateralis Colorado 528/356
76-1-807 Tamiasciurus hudsonicus Colorado 528/356
76-1-1078 Eutamias minimus Colorado 528/356
T5-2092 Sciurus griseus California 425/UR
S6-14-03 Lepus californicus California 425/UR
JKT 6423 Culex vishnui Indonesia 360/UR
JKT 6969 Anopheles vagus Indonesia 360/UR
JKT 7043 Anopheles subpictus Indonesia 375/UR
Bluetongue 17 Sheep Wyoming 275/UR
aUR, nested PCR gave an uninterpretable result (multiple bands).
on May 15, 2020 by guest
http://jcm.asm.org/
(Bio 101, La Jolla, Calif.) and eluted into 50ml of water. The 420-bp fragment was dried completely and resuspended in 12ml of water to concentrate it. Twelve microliters of each preparation was ligated into pGEM-T vector system 1 (Pro-mega, Madison, Wis.), according to the manufacturer’s instructions. Ligations were transformed in Epicurian Coli XL-1 Blue Supercompetent cells (Strat-agene, La Jolla, Calif.) by following the recommended procedures, and 18 white colonies were picked from each transformation. Purified plasmid DNA was extracted with a QIAGEN spin plasmid miniprep kit and checked for the cor-rectly sized inserts by restriction endonuclease digestion, followed by agarose gel electrophoresis. Sequencing reactions were prepared with an ABI Prism dye terminator cycle sequencing core kit (Perkin-Elmer), and sequences were elec-trophoretically resolved with an ABI Prism 377 sequencer. The DNA sequences were analyzed and compared to the published sequence of CTF virus (S2 seg-ment) with Lasergene (DNAstar), and a GenBank search was performed to locate any sequences homologous with the 420-bp product.
Evidence of CTF viral infection in human sera by using RT-PCR as a diag-nostic method.Sera obtained from a collection of clinical specimens at the CDC were tested by RT-PCR for the presence of CTF virus. With a few exceptions, sera were 1 to 2 years old and had been frozen at2708C after initial testing. Eighteen sera were chosen and encompassed a range of collection dates after the onset of symptoms (Table 3). The method of RNA extraction with the Qiagen spin columns was modified slightly from the recommended procedure by treating 500ml of serum in 2 ml of lysis buffer and adding 2 ml of ice-cold ethanol to the
lysate. This mixture was loaded in successive aliquots onto a single spin column before the wash and elution steps were completed, in accordance with the manufacturer’s instructions. For comparison, all samples were evaluated by a 90% plaque reduction neutralization test (PRNT), as previously described (9), and by VI on Vero cell six-well plates with 100ml of undiluted serum in the first well, followed by 10-fold successive dilutions in BA-1 medium.
RESULTS
Specificity of the RT-PCR.ds RNA was isolated from viral cultures of a variety of human, tick, and rodent CTF virus isolates, along with other coltiviruses (JKT isolates), an orbi-virus (bluetongue serotype 17), and an irrelevant alphaorbi-virus, VEE (TC-83). These were subjected to the RT-PCR method described above. A DNA sample derived from a rickettsial culture was included to ensure that no amplicons were gener-ated by the CTF virus primers with this clinically pertinent organism. All the CTF virus isolates gave products of the predicted size of 528 bp when they were analyzed on agarose gels and visualized with ethidium bromide. Viral RNA ampli-fication of Eyach, S6-14-03, and T5-2092 yielded fragments of approximately 425 bp. JKT virus strain 7043 gave a weak band at about 375 bp, while strains 6423 and 6969 showed slightly smaller products, with a larger proportion of the amplicons presumably being nonspecific, at around 150 bp. Bluetongue virus 17 and R. rickettsii gave small bands at about 275 bp, which were also seen in some of the CTF virus strains as secondary products. These too were probably nonspecific, while VEE TC-83 showed no amplification. Nested PCR with the original forward primer but an alternate reverse primer confirmed the results of the first round of amplification. All CTF virus strains gave the expected band size of 356 bp, as did the antigenically distinct isolate 720896. The other CTF virus-related viruses, the JKT viruses, and bluetongue virus serotype 17 all gave ambiguous results which were uninterpretable due to multiple DNA bands when visualized on a gel (Table 1). Selected results of the first-round amplification are illustrated in Fig. 1.
[image:3.612.319.554.71.160.2] [image:3.612.59.298.90.253.2]Sensitivity of CTF virus detection from tissue culture prep-arations by RT-PCR.The limits of detection of CTF virus for a number of the viruses listed in Table 1 by the RT-PCR method were determined by comparing viral plaque titers in Vero cells and visibility on a gel of the DNA fragments am-plified from 10-fold dilutions of viral culture supernatants. Sensitivities were similar for CTF (Florio), Eyach, and S6-14-03 viruses. RNA extracted from 10 PFU/ml (approximately 0.1 viral plaque) of CTF (Florio) virus gave a low-intensity band when it was electrophoresed in an agarose gel containing ethidium bromide following RT-PCR (Fig. 2). Strong bands were visible from RNA extracted from a 1,000-PFU/ml starting culture. The test was not as sensitive for strains 76-1-1007, FIG. 1. Detection of RT-PCR products (10ml) from CTF virus isolates and controls with a 1.5% agarose gel. Lanes: A, low DNA mass ladder (labeled in base pairs); B, human isolate CTF (Florio) virus; C, tick isolate CTF (64V37) virus; D, mammal isolate CTF (76-1-1007) virus; E, Eyach virus; F, bluetongue 17 virus; G, JKT (7043) virus; H, R. rickettsii; I, VEE TC-83.
TABLE 2. Comparison of VI and RT-PCR techniques for detection of CTF viral RNA in whole mouse blood
Mouse Test Test result on day p.i.:
0 1 3 5 8 11 14 18 21 24
1 VI 2 2 1 1 2 2 2 2 2 2
RT-PCR 2 2 1 1 2 2 2 2 2 2
2 VI 2 2 1 pa p p p p p p
RT-PCR 2 2 1 p p p p p p p
3 VI 2 2 2 1 1 1 2 2 2 2
RT-PCR 2 2 1 1 1 2 2 2 2 2
4 VI 2 2 1 1 1 2 2 2 2 2
RT-PCR 2 2 1 1 1 2 2 2 2 2
5 VI 2 2 1 2 p p p p p p
RT-PCR 2 2 1 1 p p p p p p
ap, mouse dead (death unrelated to viral infection).
TABLE 3. Comparison of the utility of PRNT, RT-PCR, and VI in tissue culture in detecting CTF viral infection in human sera
Serum
sample no. No. of day(s) after onsetof symptoms PRNTtitera
Test result by: RT-PCR VI
1 0 5 1 2
2 0 5 1 2
3 1 5 1 2
4 1 5 1 2
5 3 5 1 2
6 3 5 1 2
7 5 5 1 2
8 6 5 1 2
9 7 5 1 2
10 8 5 1 2
11 12 40 2 2
12 13 320 2 2
13 17 160 2 2
14 19 10 2 2
15 21 10 1 2
16 28 160 2 2
17 52 20 2 2
18 135 160 2 2
aA PRNT titer of 5 is regarded as a negative result.
on May 15, 2020 by guest
http://jcm.asm.org/
[image:3.612.59.299.520.717.2]R6225, and 64V37. Limits of detection for these isolates ranged from 33103to 63105PFU/ml in the starting mate-rials. Morphological differences among the strains were noted: Florio, 76-1-1007, R6225, and 64V37 all had plaques approx-imately 2.5 to 4.0 mm in diameter, visible 4 days p.i.; S6-14-03 had a plaque size of 2 mm, visible 5 days p.i.; and Eyach had very small plaques (,0.5 mm), visible 5 days p.i.
CTF viral infection in experimentally infected mice.It is well documented that CTF virus persists in human erythrocytes for a longer period of time than it does in the serum (20). In the absence of appropriate infected human erythrocytes, a mouse model was used to determine whether the RT-PCR method was capable of detecting CTF virus in whole blood. Mouse blood treated with EDTA to prevent clotting was tested by RT-PCR and VI in Vero cells for the presence of CTF virus. Fragments of the expected size (528 bp) were generated by RT-PCR, and results from the two techniques were similar. In both cases, virus was detectable starting on day 3 and ending on day 8 (Table 2). A few exceptions were noted, however. Blood from mouse 3 tested positive for CTF virus on day 3 by RT-PCR but negative by VI. The same was true for mouse 5 on day 5. Virus was isolated from mouse 3 on day 11 but could not be detected by the RT-PCR assay. Amplification of RNA ex-tracted from the blood of CTF virus-infected mice on days 0 and 1 p.i. consistently generated DNA bands of approximately 420 bp, though those from day 0 p.i. were barely visible. Figure 3 illustrates the detection of CTF virus in whole mouse blood by RT-PCR, with mouse 4 blood as an example.
Sequence analysis of CTF (Florio) virus and a 420-bp band from amplified mouse blood RNA.A consensus sequence gen-erated by analysis of DNA fragments processed from RT-PCR amplicons of CTF (Florio) virus was compared to the pub-lished CTF (Florio) virus S2 segment sequence (GenBank accession no. U53227), from which the primers were derived. The sequences were virtually identical, with the one exception of AC at nucleotides 636 and 637 of the published sequence being replaced with T in the consensus sequence. The same comparison was made with a consensus sequence of the 420-bp fragment derived from mouse blood on day 1 p.i., and this sequence was found to be devoid of any homology to the CTF
virus S2 segment sequence. A search of GenBank was per-formed to determine if homology existed between our consen-sus sequence and any other in the database. The search re-vealed 89% homology with a 379-bp segment of DNA from a
Rattus norvegicus mitogenic regulatory gene, which is
transcrip-tionally suppressed by src and ras (GenBank accession no. 231460).
Human serum specimens.Eighteen human sera were tested for evidence of CTF virus by RT-PCR and PRNT, and the two techniques were compared. All specimens were from previ-ously confirmed CTF cases. A 528-bp DNA band was visual-ized after performing RT-PCR on each serum specimen, be-ginning with the initial day of onset of symptoms up to and including 8 days after onset. All RT-PCR results were negative for sera 12 days after onset and thereafter, with the single exception of a sample at 21 days after onset, which tested positive. Conversely, PRNT titers did not appear until 12 days after onset and neutralizing antibody was detected in the re-mainder of the specimens. All samples tested negative by VI (Table 3). Earlier experiments indicated that although CTF viral RNA could be detected with the standard 140 ml of starting material (data not shown), greater intensity in banding could be achieved without background interference by increas-ing the quantity of serum tested in the RNA extraction and by proportionally increasing the volume of viral lysis buffer and ethanol. Normal human serum was used as a negative control, and results were free of extraneous bands. Figure 4 illustrates an example of CTF viral RNA detection in human sera with the CTF virus primers.
DISCUSSION
The genus Coltivirus was created in 1990, when a proposal recognizing the unique properties of ds RNA viruses with 12 genomic segments was accepted by the International Commit-tee on the Taxonomy of Viruses at the International Congress for Virology in Berlin, Germany. CTF (Florio) virus is the type species for this genus, which also includes numerous CTF virus variants and Eyach virus and variants, plus several isolates from Indonesia and China (7).
[image:4.612.333.540.70.153.2]Karabatsos et al. (15) reported significant antigenic variation among 20 virus isolates identified as CTF viruses, especially with respect to the human isolates. In addition, these research-ers found that the antigenic relatives Eyach and S6-14-03 were clearly distinct from the CTF viruses and from each other and that T5-2092 was antigenically identical to S6-14-03. Hence, there exists a great deal of potential variability among serolog-ical results, particularly where human subjects are concerned. Any useful diagnostic method for detection of CTF virus must therefore accommodate the existing diversity. The genomic segments of CTF virus isolates are identical by agarose gel FIG. 2. Sensitivity of detection of RT-PCR products of CTF (Florio) virus on
a 1.5% agarose gel. Lanes: A, low DNA mass ladder (labeled in base pairs); B, 104PFU/ml; C, 103PFU/ml; D, 102PFU/ml; E, 101PFU/ml; F, 100PFU/ml; G,
[image:4.612.75.282.71.126.2]negative reagent control.
FIG. 3. Agarose gel (1.5%) of RT-PCR products from whole mouse blood (mouse 4, samples taken at various days p.i.) and controls. Lanes: A, low DNA mass ladder (labeled in base pairs); B to I, CTF (Florio) virus prior to inoculation and on day 0 (C), day 1 (D), day 3 (E), day 5 (F), day 8 (G), day 11 (H), and day 14 (I) p.i.; J, normal mouse blood.
FIG. 4. Detection of CTF virus in human sera by the CTF virus RT-PCR assay. Samples were obtained from patients at various dates after the onset of symptoms. Lanes: A, low DNA mass ladder (labeled in base pairs); B, CTF (Florio) virus; C, day 0; D, day 3; E, day 7; and F, day 12 after onset of symptoms; G, normal human serum.
on May 15, 2020 by guest
http://jcm.asm.org/
[image:4.612.61.297.614.685.2]electrophoresis (5), and RNA-RNA blot hybridizations of CTF virus isolates indicate significant sequence homology (3). Thus, the use of viral genomic ds RNA on which to base a diagnostic test for CTF virus is attractive. The RT-PCR technique de-scribed here was able to amplify RNAs derived from all the isolates of CTF virus included in the study and also to distin-guish them from Eyach, S6-14-03, and T5-2092 viruses. It was not possible, however, to discriminate among these three an-tigenic relatives. One CTF virus-related isolate (720896) which is antigenically distinct from CTF virus by PRNT (unpublished data) was identical to the CTF virus isolates by the RT-PCR assay, indicating that for this virus, the amplified portion of the S2 segment was homologous to that of CTF virus and that neutralizing antibody was probably elicited by a viral protein other than VP12. The CTF virus strains were further differen-tiated from the Indonesian JKT viruses, which produced faint amplicons of significantly smaller sizes. The experimental de-sign included orbivirus bluetongue serotype 17 and R. rickettsia DNAs as controls for specificity. As expected, no amplicons of pertinent size were generated.
The ability to detect CTF viral RNA isolated from less than 100 PFU/ml when amplified by RT-PCR is relatively common, though to achieve this sensitivity, a method of detection other than electrophoresis, such as ELISA (10) or dot hybridization (1), is often necessary. The sensitivities of the test developed here for the prototype virus and some others equal or exceed the 100-PFU/ml limit only by using agarose gel electrophoresis, though the sensitivities for some of the strains tested fell short of this figure.
CTF virus has been isolated from a variety of mammals (15). In addition, experimental infections have been successfully in-duced in a number of Colorado species (4). Similarly, outbred laboratory mice are susceptible to the disease (19) and thus provide a convenient model with which to examine the effec-tiveness of the CTF virus RT-PCR technique as it applies to whole blood. A small number of mice were used for this model, as the object was to demonstrate the potential utility of the method with whole blood, as opposed to making any statistical observations. Mouse serum was not tested by RT-PCR, be-cause the volume required to obtain a reasonable signal would necessitate that the mouse be sacrificed and therefore the course of infection in each individual could not be monitored. CTF virus was detected in blood by the RT-PCR assay and in some instances was more sensitive than VI. Similar results were reported for RT-PCR of blood drawn from rams that had been experimentally infected with bluetongue virus serotype 11 (2). The RT-PCR may have been inhibited to some extent by the presence of hemoglobin, which has been reported to inhibit PCR. Although a protein-free product should have been ob-tained by the method used here, the initial treatment of whole blood with proteinase K, as opposed to guanidine thiocyanate, may have yielded greater sensitivity (16). The results suggest that whole blood may be an appropriate specimen type to request when RT-PCR is the diagnostic method of choice. It would have the potential of offering greater sensitivity and probably a longer period of usefulness after the onset of symp-toms than serum. The only drawback was the appearance of extra DNA bands in the products derived from nonviremic blood. Sequence analysis revealed that these were completely unrelated to CTF virus, and therefore this phenomenon may not occur with human blood. Because sufficient samples were lacking, no data that pertained to evidence of CTF virus infec-tion obtained by using RT-PCR as a diagnostic tool for human whole blood were reported here. However, preliminary find-ings from a few historical samples were promising (data not shown) and future experiments may reveal the utility of this
more concentrated source of viral RNA. The addition of an anticoagulant to the blood is important, as handling of the clotted erythrocyte fraction is difficult. The use of EDTA for this purpose is preferable to heparin, which is a known inhib-itor of PCR (14).
The serological methods developed for CTF diagnosis have previously been reviewed (7, 20). The application of each ex-isting technique is limited by the delayed appearance of de-tectable antibody in the serum. Our laboratory uses both PRNT and an unpublished modification of the immunoglobu-lin M ELISA of Calisher et al. (9). Of the methods available, these are probably the least subjective, but neutralizing anti-body does not appear until 10 days after onset and immuno-globulin M does not appear until day 17. Probably the most sensitive procedure remains VI from a blood clot obtained from a patient during the febrile stages of the infection. Washed cells are inoculated into 1- to 3-day-old suckling mice, which die in 4 to 8 days. Confirmation is done by PRNT or immunofluorescent staining of blood or brain tissue. Lack of timeliness remains a major drawback with these methods and is a problem, especially when more serious, but treatable, dis-eases are the possible causative agents of an infection. R.
rickettsii is the etiologic agent of RMSF. Early symptoms are
similar to those of CTF, and transmission of the disease is also by D. andersoni ticks. Consequently, clinical specimens are frequently coinvestigated for the presence of R. rickettsii and CTF virus. A PCR-based test for RMSF was recently published (22), but it suffered from a lack of sensitivity and the need for reamplification was frequently indicated. The CTF virus RT-PCR assay was shown to produce negative results with R.
rickettsii, and therefore false-positive banding will not be seen
if this agent is present in a clinical sample. The results of this study show that CTF viral RNA can be detected reliably in sera up to and including 8 days after onset, without background or the necessity of nested amplification. The RT-PCR protocol proved more sensitive than VI in tissue culture for these spec-imens, which failed to produce viral plaques. This is the first report of a rapid diagnostic technique to detect CTF virus in sera during the acute stages of infection.
The CTF RT-PCR assay is capable of offering either positive confirmation of CTF viral infection or evidence to rule out the possibility in a timely manner. Serum remains the most easily obtainable of specimens, and therefore a method that can detect CTF viral RNA from this source is likely to be readily embraced by diagnostic facilities. This technique will provide early CTF diagnoses, thus bridging the gap between clinical onset of the disease and the time frame where traditional serological methods may reliably be applied.
ACKNOWLEDGMENTS
We thank the following individuals: Robert Shope for providing the JKT viruses; Don McKechnie for supplying the rickettsial DNA; Denise Martin and Rich Tsuchiya for their excellent technical assis-tance; and Bill Wilson, Robert McLean, and John Roehrig for their valuable advice.
REFERENCES
1. Akita, G. Y., J. Chinsangaram, B. I. Osburn, M. Ianconescu, and R. Kauf-man.1992. Detection of bluetongue virus serogroup by polymerase chain reaction. J. Vet. Diagn. Invest. 4:400–405.
2. Akita, G. Y., J. Glenn, A. E. Castro, and B. I. Osburn. 1993. Detection of bluetongue virus in clinical samples by polymerase chain reaction. J. Vet. Diagn. Invest. 5:154–158.
3. Bodkin, D. K., and D. L. Knudson. 1987. Genetic relatedness of Colorado tick fever virus isolates by RNA-RNA blot hybridization. J. Gen. Virol. 68:1199–1204.
4. Bowen, G. S., R. B. Shriner, K. S. Pokorny, L. J. Kirk, and R. J. McLean. 1981. Experimental Colorado tick fever virus infection in Colorado
on May 15, 2020 by guest
http://jcm.asm.org/
mals. Am. J. Trop. Med. Hyg. 30:224–229.
5. Brown, S. E., B. R. Miller, R. G. McLean, and D. L. Knudson. 1989. Co-circulation of multiple Colorado tick fever virus genotypes. Am. J. Trop. Med. Hyg. 40:94–101.
6. Brown, S. E., B. M. Gorman, R. B. Tesh, and D. L. Knudson. 1993. Colti-viruses isolated from mosquitoes collected in Indonesia. Virology 196:363– 367.
7. Brown, S. E., and D. L. Knudson. 1995. Coltivirus infections, p. 329–342. In J. S. Porterfield (ed.), Exotic viral infections. Chapman & Hall, London, United Kingdom.
8. Burgdorfer, W. 1977. Tick-borne diseases in the United States: Rocky Moun-tain spotted fever and Colorado tick fever. Acta Trop. 34:103–126. 9. Calisher, C. H., J. D. Poland, S. B. Calisher, and L. A. Warmoth. 1985.
Diagnosis of Colorado tick fever virus infection by enzyme immunoassays for immunoglobulin M and G antibodies. J. Clin. Microbiol. 22:84–88. 10. Chang, G.-J. J., D. W. Trent, A. V. Vorndam, E. Vergne, R. M. Kinney, and
C. J. Mitchell.1994. An integrated target sequence and signal amplification assay, reverse transcriptase–PCR–enzyme-linked immunosorbent assay, to detect and characterize flaviviruses. J. Clin. Microbiol. 32:477–483. 11. Dangler, C. A., C. A. de Mattos, C. C. de Mattos, and B. I. Osburn. 1990.
Identifying bluetongue virus ribonucleic acid sequences by the polymerase chain reaction. J. Virol. Methods 28:281–292.
12. Emmons, R. W., and E. H. Lennette. 1966. Immunofluorescent staining in the laboratory diagnosis of Colorado tick fever. J. Lab. Clin. Med. 68:923– 929.
13. Goodpasture, H. C., J. D. Poland, B. Francy, G. S. Bowen, and K. Horn. 1978. Colorado tick fever: clinical, epidemiologic, and laboratory aspects of 228 cases in Colorado in 1973–1974. Ann. Intern. Med. 88:303–310. 14. Holodniy, M., S. Kim, D. Katzenstein, M. Konrad, E. Groves, and T. C.
Merigan.1991. Inhibition of human immunodeficiency virus gene amplifi-cation by heparin. J. Clin. Microbiol. 29:676–679.
15. Karabatsos, N., J. D. Poland, R. W. Emmons, J. H. Mathews, C. H. Calisher, and K. L. Wolff.1987. Antigenic variants of Colorado tick fever virus. J. Gen. Virol. 68:1463–1469.
16. Kox, L. F. F., D. Rheinthong, A. M. Miranda, N. Udomsantisuk, K. Ellis, J. van Leeuwen, S. van Heusden, S. Kuijper, and A. H. J. Kolk.1994. A more reliable PCR for detection of Mycobacterium tuberculosis in clinical samples. J. Clin. Microbiol. 32:672–678.
17. Kuno, G., C. J. Mitchell, G.-J. J. Chang, and G. C. Smith. 1996. Detecting bunyaviruses of the Bunyamwera and California serogroups by a PCR tech-nique. J. Clin. Microbiol. 34:1184–1188.
18. Lanciotti, R. S., C. H. Calisher, D. J. Gubler, G.-J. Chang, and A. V. Vorndam.1992. Rapid detection and typing of dengue viruses from clinical samples by using reverse transcriptase-polymerase chain reaction. J. Clin. Microbiol. 30:545–551.
19. McClean, R. G. Personal communication.
20. Poland, J. D. 1985. Colorado tick fever, p. 195–197. In R. B. Conn (ed.), Current diagnosis, 7. The W. B. Saunders Co., Philadelphia, Pa. 21. Sakomoto, K., R. Punyahotra, N. Mizukoshi, S. Ueda, H. Imagawa, T.
Sugiura, M. Kamada, and A. Fukusho.1994. Rapid detection of African horsesickness virus by the reverse transcriptase polymerase chain reaction (RT-PCR) using the amplimer for segment 3 (VP3 gene). Arch. Virol. 136:87–97.
22. Sexton, D. J., S. S. Kanj, K. Wilson, G. R. Corey, B. C. Hegarty, M. G. Levy, and E. B. Breitschwerdt.1994. The use of polymerase chain reaction as a diagnostic test for Rocky Mountain spotted fever. Am. J. Trop. Med. Hyg. 50:59–63.
23. Wilson, W. Personal communication.
