SUMMARY. In this report, we describe a real-time reverse transcriptase–polymerase chain reaction (RRT-PCR) diagnostic test for infectious bronchitis virus (IBV) with the use of fluorescence resonance energy transfer (FRET) technology. Two primers that amplify a 383-base pair product between nucleotide positions 703 and 1086 relative to the start codon for the S1 gene of the Massachusetts 41 virus were designed and used to amplify the Beaudette, Massachusetts 41, Florida 18288, Connecticut, Iowa 97, Arkansas DPI, CA/NE95/99, DE/072/92, and GA/0470/98 strains of IBV. The primers were specific and did not amplify New Castle disease virus, Mycoplasma spp., or infectious laryngotracheitis virus. For RRT-PCR by FRET, an anchor probe conjugated to fluorescein and a detection probe conjugated to a red fluorophore were designed to anneal to a hypervariable region within the 383-base pair product. The level of sensitivity was 1 × 104 RNA molecules used as starting template. After amplification, a melting curve analysis was conducted to specifically identify IBV types. Because of sequence differences in the annealing position of the detection probe, the Arkansas, Connecticut, Beaudette, and Massachusetts 41 strains could be differentiated. No fluorescence was observed for the DE/072/92 and GA/0470/98 viruses with the anchor and detection probes. When the Beaudette strain was examined, two melting peaks were observed at 44 C and 51 C, indicating a quasispecies in that laboratory strain of IBV. Routine typing of vaccine strains of IBV was possible with this technology, but high standard deviations associated with the melting curve analysis of the FRET probes described herein made it difficult to use this test reliably for routine typing of IBV field isolates.
Infectious bronchitis virus (IBV) is a coronavirus that causes a highly contagious upper respiratory tract disease in chickens. The disease is worldwide in distribution and is extremely difficult to control because multiple serotypes and variants of the virus occur that are not cross protective. Live attenuated vaccines for the major serotypes of IBV can be used to control most outbreaks, but it is first necessary to isolate and definitively identify the serotype of the virus responsible for the outbreak. The disease caused by IBV can appear similar to infectious laryngotracheitis, avian influenza, and viscerotropic velogenic Newcastle disease, which are high priority diseases. This is particularly true when secondary pathogens like Escherichia coli are involved. Thus, it is important to be able to diagnose IBV rapidly so one can determine if it is or is not the cause of an upper respiratory disease outbreak.
Traditionally, the virus-neutralization test conducted in embryonating eggs was used to isolate and identify different serotypes of IBV. Today, the reverse transcriptase (RT)–polymerase chain reaction (PCR) is routinely used (4,5,7). The rapid RT-PCR/restriction fragment length polymorphism (RFLP) identification test has been used to identify and characterize traditional and variant types of IBV from commercial poultry and research laboratories all over the world (2,4,7). The RT-PCR/RFLP test is based on amplification and restriction enzyme digestion of the S1 portion of the spike glycoprotein gene. The restriction fragment patterns for the S1 gene are observed after electrophoresis on a 2% agarose gel and compared with patterns of known serotypes and variants of the virus.
In this report, we describe a real-time RT (RRT)-PCR diagnostic test for IBV by the LightCyclerTM PCR amplification and detection system (Roche Diagnostics Corp., Indianapolis, IN). To identify different IBV types, a set of probes complementary to a hypervariable sequence region of the S1 gene were designed to utilize the fluorescence resonance energy transfer (FRET) technology (1). The FRET technology uses an anchor probe that is conjugated to fluorescein and a detection probe conjugated to a red fluorophore (LightCyclerTM Red 640). Fluorescein on the anchor probe emits a green fluorescent light when it is excited by the ultraviolet light source in the LightCyclerTM. When that probe comes into close proximity to the serotype-specific probe, by hybridizing to the PCR product, energy from the green fluorescent light is transferred to and excites the red dye, which then emits light at 640 nm (Fig. 1). After each round of amplification, the probes are allowed to anneal to the product and the reaction is monitored for red dye emission at 640 nm. After RT-PCR amplification, the probes are allowed to anneal to the amplified product, and a melting curve analysis is conducted to determine the number of mismatches between the detection probe and the template. In the melting curve analysis, fluorescence is monitored while the temperature of the hybridization reaction is slowly increased. The melting temperature (Tm) of the detection probe is determined by a rapid decrease in fluorescence. The anchor probe, which is much longer than the detection probe, remains attached to the template because it has a higher Tm. Because the number of mismatches between the detection probe and the template will affect the temperature at which the probe disassociates from the template, that temperature can be used to distinguish different IBV types. A perfect match will dissociate at the highest temperature, a related virus where the probe has some mismatches will detach at a lower temperature, and a different serotype with the most mismatches will separate at the lowest temperature.
Advantages of the new LightCyclerTM system are that running agarose gels is no longer necessary, and identification is almost immediate. By this system, we identified and confirmed by cloning and sequencing a quasispecies in the Beaudette laboratory strain of IBV.
MATERIALS AND METHODS
We selected several common strains of IBV (Table 1) to demonstrate the feasibility of using RRT-PCR for detection and, possibly, diagnosis of the virus. The serotypes of these standard strains of IBV have been verified in our laboratory by RT-PCR/RFLP and routine serology in embryonating eggs (3).
We designed primers to the first third of the S1 gene, which contains three hypervariable regions. It is necessary to amplify the hypervariable regions because they contain serotype-specific sequences that will be identified during the actual PCR amplification process with fluorescent dye–labeled probes. The primers were designed to conserved regions outside the hypervariable regions with the Oligo v4.05 computer program (National Biosciences, Inc., Plymouth, MN). Primers were designated IBVLC5′ and IBVLC3′, synthesized at Research Genetics, Inc. (Huntsville, AL), and tested for their ability to amplify each of the viruses listed above.
The viral RNA from 200 μl of allantoic fluid from IBV-inoculated 10-day-old embryonating eggs (virus titer = approximately 1 × 104.5 50% embryo infectious dose [EID50]/ml) was extracted with the High Pure RNA isolation kit (Roche Diagnostics Corporation). The purified RNA was resuspended in 35 μl of diethylpyrocarbonate (DEPC)-treated water. The RRT-PCR amplification was conducted with the LightCycler-RNA amplification kit SYBR Green I (Roche Diagnostics Corp.). In that kit, amplification is monitored with SYBR Green I dye, which emits a fluorescent signal at 530 nm when it binds to the double-stranded DNA and is exposed to ultraviolet light. The 20-μl RRT-PCR reaction mixture contained 4 μl of reaction mixture (included in the kit), 1 μl of each primer (approximately 125 ng), 1.6 μl of MgCl2 (included in the kit), 0.4 μl of enzyme mixture (included in the kit), 4 μl of IBV RNA (from 300 ng to 800 ng per reaction), and 8 μl of DEPC-treated water.
The RRT-PCR amplification was done in a LightCyclerTM (Roche Diagnostics Corp.). The RT step was incubated at 55 C for 10 min, then the reaction mixture was heated to 95 C for 30 sec. Next, 45 cycles of 95 C for 0 sec, 52 C for 10 sec, and 72 C for 13 sec were conducted and SYBR Green I fluorescence was monitored after each cycle.
A Massachusetts 41 S1 gene RT-PCR amplified with previously reported primers (4) was cloned into the TOPO XL vector (Invitrogen, Carlsbad, CA). Runoff RNA transcripts were produced by the RiboMAXTM RNA production system (Promega, Madison, WI) according to the manufacturer's recommendations. The amount of RNA produced was determined with an Eppendorf BioPhotometer spectrophotometer (Hamburg, Germany). To determine the sensitivity of the test, serial 10-fold dilutions of the RNA product were made and used as template in the RRT-PCR reaction described above.
An anchor probe and a detection probe to distinguish the different types of IBV were designed with the Oligo v4.05 computer program (National Biosciences, Inc.) and synthesized at TIB Molbiol LLC (Adelphia, NJ). After each round of amplification, the probes were annealed to the product and the amount of amplified DNA was monitored real time by measuring for red dye emission at 640 nm. In addition, the probes were tested for their ability to detect different serotypes of IBV by running a melting curve analysis after amplification. After the RRT-PCR amplification, the reaction mixture was heated to 95 C for 1 sec and the two probes were allowed to anneal to the PCR product. The mixture was cooled to 30 C, then slowly heated at a rate of 0.1 C/sec to 92 C. Fluorescence at 640 nm was monitored in a stepwise fashion after each 0.1 C change in temperature.
Cloning and sequencing
The RRT-PCR product from the Beaudette strain of IBV was cloned with the TOPO TA cloning kit (Invitrogen) according to the manufacturer's recommendations. Clones containing the RRT-PCR product were identified by size on a 1% agarose gel. The cloned inserts were amplified with the LightCycler-RNA amplification kit SYBR Green I and the LightCyclerTM (Roche Diagnostics Corp.) as above but without the RT step. In addition, a melting curve analysis with the anchor and detection probes described above was conducted for each amplified clone. Finally, the amplified products from the clones were sequenced with the ABI Prism DNA sequencing kit (Applied Biosystems, Foster City, CA) with the IBVLC5′ and IBVLC3′ primers and an ABI Prism 310 genetic analyzer (Applied Biosystems). The nucleotide sequence data were analyzed with the MacDNASIS Pro v3.5 software (Hatachi Software Engineering Co., San Bruno, CA).
Testing samples from experimentally inoculated birds
To examine the ability of this new test to detect IBV, we tested samples taken from experimentally inoculated chickens. Briefly, chickens were inoculated intraocularly with 1 × 104 EID50 of the Arkansas virus at 4 wk of age, and tracheal swabs were collected at 5 days postchallenge. Tracheal swab material was inoculated into 10-day-old embryonating eggs as previously described (3), and the allantoic fluid was harvested 48 hr postinoculation and tested by the LightCyclerTM method described above.
We designed two primers to amplify a 383-base pair product between nucleotide positions 703 and 1086 relative to the start codon for the S1 gene of the Massachusetts 41 virus (GenBank accession no. X04722). The sequence of the 5′primer is 5′-ACTGGCAATTTTTCAGA-3′ and is designated IBVLC5′. The sequence of the 3′ primer is 5′-ACAGATTGCTTGCAACCAC-3′ and is designated IBVLC3′. The primers were used to amplify the Beaudette, Massachusetts 41, Florida 18288, Connecticut, Iowa 97, Arkansas DPI, CA/NE95/99, DE/072/92, and GA/0470/98 strains of IBV (Fig. 2). The level of sensitivity with RNA runoff transcripts of the Massachusetts 41 S1 gene was approximately 1 × 104 RNA molecules (0.01 pg RNA). The primers were tested for specificity against other upper respiratory tract pathogens, including New Castle disease virus, Mycoplasma spp., and infectious laryngotracheitis virus, and no amplification products were observed (data not shown). Different slopes of the amplification curves in Fig. 2 are due to the amount of RNA template used and the efficiency of the reaction in each tube.
Melting curve analysis
We designed an anchor and a detection probe based on the Massachusetts 41 sequence of the S1 gene (Fig. 1). A region of variability was selected for the detection probe based on comparison with sequences for the other viruses used in this study. The sequence of the detection probe and the mismatches with the Connecticut, Arkansas, and DE/072/92 viruses are shown in Fig. 3.
To determine if the anchor and detection probes could indeed be used to type IBV, the Connecticut, Arkansas, and Massachusetts 41 viruses were amplified with IBVLC5′ and IBVLC3′ primers, then a detection probe melting curve analysis was conducted on the RT-PCR–amplified products. On the basis of the melting curve analysis, a clearly different rate of change in fluorescence was observed for the Arkansas, Connecticut, and Massachusetts 41 strains (Fig. 4). However, the average detection probe Tms, based on a minimum of four tests for each virus examined in this study (Table 2), were statistically not significantly different for Massachusetts 41, Connecticut, and Florida strains.
No fluorescence was observed for the DE/072/92 and GA/0470/98 viruses with the anchor and detection probes. We tried to lower the initial temperature for the melting curve analysis to room temperature, but still no fluorescence was detected (data not shown).
The Beaudette strain of IBV was examined four different times with allantoic fluid from two different passages of the virus, and each time, two melting peaks were observed at 44 C and 51 C (Fig. 5). Sequence data obtained for nine clones of the amplified product from the Beaudette strain of IBV showed a base change from T to A at position 14 from the 5′ end of the detection probe in five of the nine clones. No differences were observed in the region of the anchor probe.
Detection in experimentally inoculated birds
When chickens were inoculated with the Arkansas strain of IBV, all of the samples (10/10) taken at 5 days postchallenge were positive for that serotype of the virus, and virus was not detected (0/10) in the negative control birds (data not shown) with the LightcyclerTM and FRET anchor and detection probes.
In this study, we used RRT-PCR to detect several common strains of IBV. The test takes 38 min (35 min for RT-PCR reaction and 3 min for melting curve analysis) to run and, with the FRET probes, costs approximately $19.00 per test. The RRT-PCR reaction with SYBR Green I alone costs approximately $8.00. We found that the RRT-PCR test could be used to detect 1 × 104 RNA molecules (0.01 pg RNA). It is difficult to compare the sensitivity of this test with other RT-PCR tests for IBV because of differences in the efficiency of viral RNA extraction and RT-PCR reactions. However, the test reported herein appears to be adequately sensitive for detection of laboratory strains and bird isolates of IBV propagated in 10-day-old embryonating eggs.
We found that detection and differentiation of the common IBV types (Massachusetts 41, Arkansas, and Connecticut) were reliable when the appropriate controls were used. However, the relatively high standard deviation calculated for the detection probe Tm for other types of IBV made it difficult to distinguish among them even when the appropriate controls were used. It appears that the number and type of nucleotide mismatches, as well as the relative position of the mismatches and the surrounding sequence, affect the Tm of the detection probe. This made it difficult to distinguish among IBV types with a similar number of G/C mismatches.
No fluorescence was detected when the FRET anchor and detection probes were used with RRT-PCR product from the DE/072/92 and GA/0470/GA98 strains. Because those strains were amplified with the IBVLC5′ and IBVLC3′ primers, we assume that the detection probe, which had seven mismatches with the amplified product of those strains, was not hybridizing. Lowering the initial temperature of the melting curve analysis did not help.
Because of the high standard deviations associated with the melting curve analysis, we could not reliably develop this test for routine diagnosis of IBV types. We have developed the test for use in the laboratory, however, and found it to be extremely reliable and accurate at identifying IBVs after experimental inoculation of chickens or in vaccine/challenge studies. In this study, we were able to detect and identify the Arkansas strain of IBV in all of the birds experimentally exposed to that virus. The ability to detect and identify IBV type with a test that takes about 1 hr is extremely useful when testing a large number of samples like those generated in in vivo experiments.
Each time the Beaudette strain of IBV was examined, melting peaks were observed at 44 C and 51 C, indicating that two populations or quasispecies of that virus occur in the same sample. In the region of the detection probe, the sequence of the Beaudette strain is reported to be the same as that of Massachusetts 41 (GenBank accession no. AJ311362), and a Tm of 51 C ± 1.4 calculated for the Massachusetts 41 strain is consistent with that sequence. On the basis of the area under the two peaks for the Beaudette strain, 59% of the population has a sequence different from the reported sequence. This percentage is consistent with the sequence data generated from cloned RRT-PCR product of the Beaudette strain, which showed a T to A nucleotide change at position 14 from the 5′ end of the detection probe in 55% (5/9) of the clones examined. Considering that the Connecticut strain with two mismatches only dropped the Tm of the detection probe to 49 C, that one A/T nucleotide mismatch for the Beaudette strain dropped the Tm of the detection probe to 44 C was unexpected. However, and as previously mentioned, the number, type, and position of nucleotide mismatches can significantly affect the Tm of the detection probe.
While conducting RRT-PCR on different IBV strains, we fortuitously identified the presence of a quasispecies in the sample of the Beaudette strain from our laboratory. Quasispecies have been reported for coronavirus (6,8,9). Those studies have shown that coronaviral RNAs exist as a diverse population, which is important for viral evolution and persistence of the virus. This diversity can contribute to the emergence of new virus types in the field and continued pathogenicity of the virus for the host (8). Our data confirm the existence of quasispecies in IBV, but the significance of two major populations of virus in a highly attenuated laboratory strain of IBV is not known.