Antigenic and genomic relationships among tissue culture-adapted turkey enteric coronavirus (TCV) isolates, three strains of avian infectious bronchitis virus (IBV), and mammalian coronaviruses were investigated. Immunoblotting and immunoprecipitation experiments using polyclonal antisera showed that the four major structural proteins of TCV cross-reacted with the four homologous proteins of bovine enteric coronavirus (BCV), the N and M proteins of mouse hepatitis virus serotype 3, and the N protein of IBV. Close antigenic relationships between TCV and BCV were also established by seroneutralization and hemagglutination-inhibition. Of 49 monoclonal antibodies produced against either TCV or BCV, 11 differentiated the two viruses. Five of these monoclonal antibodies had neutralizing activities and were directed to either the peplomeric S (gp200-gplOO) or hemagglutinin HE (gpl40-gp65) glycoproteins. BCV cDNA probes tested on purified viral preparations and coronavirus-positive (by electron microscopy) fecal samples from diarrheic turkey poults confirmed the relatedness of TCV and BCV. The two viruses produced distinct cytopathic changes in HRT-18 cells in the presence of trypsin, whereas only TCV isolates were able to reproduce the clinical symptoms in turkey poults. Their matrix (M) proteins undergo different glycosylation processes.

The 3' end of the turkey coronavirus (TCV) genome and the gene encoding the nucleocapsid protein (N) were cloned and sequenced. The gene encoding the membrane protein (M) was obtained by cloning a polymerase chain reaction (PCR)-amplified fragment obtained using bovine coronavirus (BCV)-specific primers. Furthermore, five TCV DNA fragments, obtained by PCR on RNA from clinical specimens and corresponding to either the N terminus of the M protein or the complete M protein were also cloned and sequenced. The sequence revealed a 3' non-coding region of 291 bases, an open reading frame (ORF) encoding the N protein with a predicted size of 448 amino acids, or an Mr of 49K, and an ORF encoding the M protein with a predicted size of 230 amino acids and an Mr of 26K. A third ORF, encoding a hypothetical protein of 207 amino acids with an Mr of 23K was found within the N gene sequence. The amino acid sequences of both the N and M proteins were more than 99% similar to those published for BCV. Extensive similarity was also observed between the amino acid sequences of the TCV N protein and those of murine hepatitis virus (MHV) (70%) and human respiratory coronavirus strain OC43 (HCV-OC43) (98%) and between the amino acid sequences of the predicted M proteins of TCV and MHV (86 %). Such striking identity suggests that BCV, TCV and HCV-OC43 must have diverged from each other only recently. A potential N-glycosylation site was found at the N terminus of the TCV M protein and is situated at the same location in BCV, MHV and transmissible gastroenteritis virus.

cDNA, synthesized on bovine coronavirus (BCV) genomic RNA templates, could be used to detect very small quantities (i.e. 1 pg) of viral RNA by hybridization with either radioisotopic-labelled or biotinylated recombinant plasmids. Virus was optimally attached to nitrocellulose membranes when spotted in 1 x SSC, whereas 20 x SSC was superior for viral RNA. Denaturation and RNA fixation of both RNA, still encapsidated in virus particles and isolated genomic RNA, was achieved by baking of the blots in vacuum. Virus detection in the supernatant of infected HRT-18 cells was feasible, but improved significantly after proteinase K treatment. No homology was observed between virus cDNA with either plasmid DNA or nucleic acid isolated from non-infected HRT-18 cells. Hybridization with radioisotopic-labelled probes in higher formamide concentrations (up to 60%) increased the detection signals, possibly by reducing reassociation of the probe. Significant detection amplification (30-50 times) was achieved in the case of biotinylated probes by stimulation of hyperpolymer formation on already hybridized target sequences, by additional hybridization with biotinylated pUC-19. A detection amplification was also obtained when hybridization was done with two probes (pBC-52 and pBC-247), containing non-overlapping viral sequences. Although the detectability was surpassed by biotinylated probes, sensitivity was superior in radioisotopic virus detection.

The polymerase chain reaction (PCR) was used to synthesize ds and ss probes for the detection of bovine coronavirus (BCV) using recombinant plasmids as template molecules. The ds probes detected a minimum of about 2 X 10(5) viral genomes after exposure for 1 h, a detection limit similar to nick-translated probes after exposure of the films for 60 h. More than 8 h exposure to blots probed with these ds probes resulted in complete darkening of the film. The ss probes, synthesized by asymmetric PCR on linearized plasmids, permitted the detection of 5 X 10(4) genomes, which equalled the capacity of random-primed probes. Prolonged exposure did not increase the background as in case of ds PCR-probed blots. Probes, synthesized by asymmetric PCR and random-priming were labeled to similar specific activities and were better in terms of sensitivity and detectability as opposed to nick-translated probes. However, the specificity of detection with ss probes as to random primed probes was increased further. About 10 viral genomes, after fragment-specific amplification by PCR, were detected by agarose-gel analysis. PCR-probe synthesis was simple, highly reproducible, and allowed the synthesis of probes for specific fragments.

Molecular hybridization, previously optimized for purified bovine coronavirus (BCV), was adapted for detection of virus in clinical specimens. For this purpose, the accuracy of the existing tests had to be improved and suitable means for removal of extraneous molecules had to be established. Six radioactive probes were used to obtain adequate detection signals. These probes, containing the complete N and E1 gene sequences and other sequences, hybridized to about 1/4 of the total length of the viral RNA. Genomic RNA could be detected after direct spotting of samples, but prior Freon-extraction or centrifugation of specimens on a cushion of sucrose improved considerably the positive identification of virus containing samples. RNA detection in positive clinical specimens was significantly better by hybridization than immunological detection of BCV by ELISA, although differences were slight after two passages of the virus on HRT-18 cell monolayers. Consequently, the reliability of positive and negative test results in hybridization tests on Freon extracted specimens was better than in ELISA. However, results after extraction with other organic solvents were inferior. The accuracy of ELISA was surpassed by hybridization assays. Background signals, due to vector homology were found to be negligible in all the samples analyzed.