key: cord-0813182-22f04uqe
authors: Verbeek, A.; Dea, S.; Tijssen, P.
title: Genomic relationship between turkey and bovine enteric coronaviruses identified by hybridization with BCV or TCV specific cDNA probes
date: 1991
journal: Arch Virol
DOI: 10.1007/bf01316754
sha: 1f2c7fab75dd23f071966aef8ec6792e0c02a36f
doc_id: 813182
cord_uid: 22f04uqe

Genomic relationships between turkey and bovine coronavirus (TCV and BCV), which are currently placed in distinct antigenic groups, were demonstrated by hybridization using specific cDNA probes. BCV-specific recombinant plasmid probes p 52, p 27, and p 247, holding inserts derived from (probably nonstructural) genes, and plasmids pN 17 and pN 9 holding the N and M gene, respectively, permitted the detection of isolates of both BCV and TCV with similar sensitivities. Similarly, probing supernatants of cell cultures infected with several isolates of TCV, using probes pN 17 and pM 78, respectively holding the N gene of BCV and TCV, resulted in equally intense detection signals. Only a slight detection of MHV-3, which is antigenically related to BCV, was observed, whereas the probes did not allow the detection of IBV, TGEV, and HCV-229E, which are placed in antigenic groups separate from those of BCV and TCV. Detection of TCV was improved by hybridization with BCV-specific single-stranded (ss) probes holding sequences of the N and M genes and synthesized by the polymerase chain reaction. Diagnosis of TCV in 134 clinical samples by hybridization was better with PCR-produced ss BCV-specific probes than with ds PCR-produced probes or a combination of six recombinant plasmid probes holding non-overlapping BCV-specific cDNA sequences. Detection signals were absent when probing clinical samples with(32)P-labelled pUC-DNA.

Coronaviral enteritis (Bluecomb disease) causes significant economic losses to the turkey industry in the United States and Canada [6, 24, 28] . Turkey coronavirus (TCV) infects and destroys the major absorbtive cells of the small and large intestines, leading to severe diarrhea in young poults but mild symptoms in adults' [1, 22, 24] . Initially, the disease was diagnosed by oral infection of susceptible poults with filtered intestinal contents. Subsequently, detection of the virus was attempted by direct electron microscopy of clarified fecal samples [ 17] or detection of viral antigens by immunofluorescence in the cytoplasm of mucosal cells in frozen intestinal sections [19, 21, 23] . These techniques, however, are timeconsuming when large numbers of samples have to be screened and lack sufficient sensitivity. Recently, an indirect ELISA has been employed for detection of TCV in clinical samples [7] . Rabbit hyperimmune serum produced against purified egg-adapted and tissue culture-adapted Minnesota strain of TCV crossreacted strongly with bovine coronavirus (BCV) and murine hepatitis virus type 3 (MHV-3) although BCV and TCV are classified in different antigenic groups [20] . Subsequently, the existence of a strong antigenic relationship between BCV and TCV was confirmed by monoclonal antibodies produced against either virus [9I.

In the current study, we investigated the genomic relationship between BCV and TCV by hybridization. Several recombinant plasmids, containing BCVspecific cDNA sequences [33, 34] , were radioactively labelled by nick translation and used None or in combination for the detection of TCV. Polymerase chain reaction (PCR)-produced probes which were highly efficient in detection of BCV [35] , were also used to detect purified TCV or virus in supernatants of infected cell cultures. The use of one recombinant plasmid, either BCV or TCV specific, was usually unsufficient for the detection of TCV in clinical samples from diarrheic pours. Therefore, a combination of BCV-specific probes or inherently more sensitive PCR-produced probes were used to detect virus in clinical samples to confirm the positive detection signals obtained with tissue culture-propagated virus. These results emphasized a genomic relatedness between the two viruses and ruled out the possibility of the activation in vitro of a latent BCV homologous coronavirus.

The prototype egg-adapted Minnesota strain of TCV [26] , kindly supplied by Dr. B. S. Pomeroy (College of Veterinary Medicine, St. Paul, MN, U.S.A.), was initially propagated by inoculation into the amniotic cavity of 22-to 24-days old embryonating turkey eggs and serially propagated in HRT-18 cells, a continuous cell line derived from a human rectum adenocarcinoma [8, 14] . The origin of the tissue culture adapted TCV Quebec isolate 1713 has been described previously [9] . The NCDC (Mebus) strain of BCV was obtained from the American Type Culture Collection (ATCC, VR 874). The Mebus strain and Quebec BCV isolates, obtained from diarrheic calves, were also propagated in HRT-18 cells [14] . Turkey and bovine coronaviruses were purified from the supernatant fluids of infected cell cultures by differential and isopycnic ultracentrifugation on sucrose gradients as described previously [6, 7] . The Purdue strain of transmissible gastroenteritis virus (TGEV, ATCC, VR 763) and serotype 3 of murine hepatitis virus (MHV-3, ATCC, VR 262) were propagated, respectively, in swine testicle and mouse fibroblast L2 cells [7] . The Beaudette, Hollande, and Massachusetts strains of infectious bronchitis virus (IBV) were propagated by inoculation into embryonating chicken eggs [7] . Purified human coronavirus 201 HCV-229 E was a gift from Dr. P. Talbot (Virology Research Center, Institut Armand-Frappier).

Intestinal contents from 1-to 8-week-old turkey pours with mild to severe diarrhea were obtained from commerical flocks from different locations in Quebec, Canada. Specimens from 5 to l0 poxflts frnm the ~nme flock were pooled and used to make ten percent homogenates in 0.05 M Tris-HC1, pH 8.0. The homogenates were clarified by centrifugation in an Eppendorf centrifuge for 2 min at room temperature. Supernatants were stored at -70 °C until further use. Virus from clinical specimens of diarrheic turkey poults was propagated by one to three passages on HRT-18 cells as described previously [8] .

The preparation of BCV-specific clones (covering about 1/4 of the genome and dispersed over its total length) has been described elsewhere [33, 34] . To prepare TCV clones, the tissue culture-adapted Minnesota strain of TCV [8] was purified on sucrose gradients (20 to 55%) and pelleted by ultracentrifugation for 2h at 25,000rpm in a Beckman SW 27.1 rotor. Genomic RNA was extracted from concentrated virus and copied into eDNA for cloning as described for BCV [33] . Briefly, purified RNA, denatured by methyl mercuric hydroxyide treatment was used as template for first-strand eDNA synthesis using Molony murine leukemia virus reverse transcriptase (Pharmacia) in the presence of either oligod(T) or random primers (BRL) [3, 11, 16] . DNA polymerase I and RNase H were used for second-strand synthesis [11] . Homopolymer tailing of the eDNA molecules [27] , using dCTP precursors, was followed by insertion into Pst I linearized (dG)-tailed pUC-9 (Pharmacia) and transformation of JM 101 cells [13] .

About 300 TCV-clones from the eDNA library were screened for the presence of the genomic 3'-end and the N gene by colony filter hybridization [16] using BCV-specific radiolabelled recombinant plasmids, containing sequences of the N gene [34, 35] . One TCV-specific recombinant plasmid (pM 78) with an insert of 1.7 kbp containing the entire sequence of the TCVN gene ( Fig. 1 ) (manuscript submitted), was selected as a probe to detect TCV. BCV-specific recombinant plasmids p 52, p 27, and p 247, which served previously for the optimization of hybridization conditions for BCV-RNA detection [33] , and probes pN 17 and pN 9, holding respectively the BCV N and M gene ( Fig. 1) , were used in the detection of several coronaviruses in order to establish the presence of potential genomic homologies. The eDNA sequences of p 52 and p 27 of 664 bp and 1120 bp, respectively, were determined (unpubl. data) and compared with sequence data available for all structural genes of BCV [-4, 15, 18] , using the IBI Pustell sequence analysis programs. The translated sequence of the inserts revealed that they are parts of open reading frames. Although both probes were found to be specific for BCV [-33] , no sequence homologies were found between these inserts and the structural genes of BCV, suggesting that these inserts represent fragments of non-structural genes. Recombinant plasmid p 247, obtained by random priming of eDNA and holding an insert of 1.6 kbp was selected for its capacity to strongly detect BCV-RNA in colony filter hybridization assays. The insert of p 247 could be detected by probe p 27 in Southern blot analysis (not shown) and might represent at least a part of a non-structural gene.

Clinical diagnosis of TCV was assayed with a pool of six 32p-labelled BCV specific recombinant plasmids (pBCV-pool), holding non-overlapping eDNA sequences, in order to amplify the detection signal. This pBCV-pool was superior in clinical diagnosis of BCV when compared with EM and ELISA [34] .

Recombinant plasmids were isolated from bacteria by alkaline lysis, purified on CsC1 gradients [16] , and labelled by nick translation in the presence of [0t32p]dCTP (3000 Ci/ mmol; 25) to specific activities of 5 to 10 x 107cpm/gg.

Recombinant plasmid pN 17 was linearized in the vector polylinker with SalI ( Fig. 1 ) and served as a template for single-stranded (ss) probe synthesis [35] . Primer PIORF 1, 5' TTA CAC CAG AGG TAG GGG TTC, complementary to the sequence located 950 to 970 nucleotides from the 3'-end of the viral genome, was used in repeated primer extension using the PCR thermal cycler to synthesize probe molecules of approximately 680 bases, corresponding to the internal open reading frame inside the N gene. Double-stranded probes to the same sequence were synthesized by PCR using pN 17 as template and PIORF 1 and PIORF2, 5' ATG GCA TCC TTA AGT GGG CCG, 1553 to 1573 nucleotides from the 3'-end, as primers. Plasmid pN 9, containing the BCV M gene (Fig. 1) , was restricted in the polylinker with Hind III for repeated primer extension in PCR, using the BCV-specific primer pXBAV, 5' GAA CAT TTC TAG ATT GGT CGG ACT G, 1526 to 1550 nucleotides from the 3'-end, resulting in probe molecules of approximately 800 bases. PCR mixtures for ss probe synthesis consisted of 110 ng template DNA, 10raM Tris-HCt, pH 8.3, 50 mM KCI, 1.5 mM MgCI2, 1 mM DTT, 200 I-tM of dATP, TTP, and dGTP, 50pmol primer, 13 gl [~3zP]dCTP (3000Ci/mmol, 3.3gM)and 1.25 U Taq DNA polymerase (Cetus) in a final reaction volume of 20 ~tl. About 10 ng template was used in case of ds probe synthesis, whereas the final dCTP concentration was adjusted with "cold" dCTP to 20 ~M. The reaction mixtures were overlaid with 30 gl paraffin oil and subjected Genomic relationship between turkey and bovine enteric coronaviruses 203 to 15 amplification cycles. Each cycle included a denaturation for 1 rain at 94 °C, an annealing for 1 min at 52 °C, and a primer extension period for 3 rain at 72 °C. The last cycle included a 10rain incubation at 72 °C, whereafter the products were separated from non-incorporated radionucleotides by spin-column chromatography, using Sephadex G-50. Ss and ds probes were labelled to specific activities of about 10 9 and t08cpm/gg, respectively.

Clarified clinical samples from diarrheic turkey poults were extracted once with Freon (1,1,2-trichloro trifluoroethane) before application to nitrocellulose membranes (pore size 0.20 gin), previously equilibrated in a 6 x SSC solution (20 x SSC = 175.3 g NaC1 and 88.2 g sodium citrate dihydrate per liter). Samples were spotted using a Minifold II slotblot apparatus (Schleicher and Schuell, Inc.) followed by baking of the blots at 80 °C for 1 to 2h under vacuum. Membranes were then rehydrated in a 6 x SSC solution and prehybridized at 42 °C for 4 h or overnight in a standard hybridization solution consisting of: 60% formamide (or as indicated in the individual experiment), 5 x Denhardt's components, 5 x SSPE salts, 0.1% SDS, and 100 gg/ml of denatured, sheared calf thymus DNA [16] . 32p-labelled recombinant plasmids and ds PCR-produced probes were heat-denatured and added to the prehybridization solution to a final concentration of 100 ng probe/ml. PCR-produced ss probes were added without previous heat-denaturation at concentrations indicated in individual experiments. Hybridization was for about 36 h at 42 °C, followed by blot-washing according to standard procedures [16] .

The double antibody-sandwich ELISA was used for the detection of TCV in clinical specimens from diarrheic turkey poults as described elsewhere [7] .

Suspensions of purified virus of several coronavirus strains were adjusted to approximately 3 x 1015 virus particles per ml as determined by EM by admixture with known amounts of latex spheres. 100 gl of each suspension was spotted on nitrocellulose membranes and hybridized separately in triplicate under relatively stringent conditions (i,e., 65% of formamide) with 32p-labelled BCV specific recombinant plasmids p 52, p 27, and p 247. The three probes allowed individually the detection of the tissue culture-adapted Mebus strain of BCV as well as two BCV Quebec isolates. In addition, strong hybridization detection signals were obtained with the tissue culture-adapted Minnesota strain and Quebec 1713 TCV isolates (Fig. 2) . Similar results were obtained when probing identical blots with pN 17 and pN9, respectively, holding sequences of the BCV N and M genes (data not shown). These BCV-specific probes allowed also individually the detection of twenty other Quebec isolates of TCV (data not shown). Only slight detection signals were seen with the antigenically related MHV-3, whereas no detection signals were obtained with several strains of IBV and the human coronavirus HCV-229 E (Fig. 2) in antigenic groups distinct from the BCV and TCV groups. Detection signals were absent when testing spotted supernatant fluids of non-infected HRT-18 cells (Fig. 2) and when probing identical blots with 32p-labelled pUC-DNA (not shown), confirming the specificity of the hybridization signals obtained.

Aliquots of serially diluted suspensions of purified tissue culture-adapted Minnesota strain of TCV were spotted on nitrocellulose membranes. Figure 3 , lane 1, shows the detection signals obtained with the PIORF 1-primed probe, synthesized on linearized plasmid pN t7, whereas lane 2 represents the detection signals obtained with the PXBAV-primed probe, synthesized on pN 9. TCV detection, using a combination of both probes, resulted in an about 3-fold increase in sensitivity, as determined by the slope of the dose-response curve [32] , whereas detectability was increased about 5-fold (Fig. 3, lane 3) . A minimum corresponding to about 2 x 10 4 viral genomes (as estimated via protein content), could be detected by combining both probes. As expected, detection of spotted pUC-19-DNA was significantly reduced with the PCR-produced probes (Fig. 3 ) when compared to nick-translated recombinant plasmid probes (Fig. 2) , therefore increasing the specificty of detection.

The efficacity of TCV-and BCV-specific probes in detecting TCV was established using seventeen clinical isolates, propagated by three passages in HRT- Here, about 25 gl of each diluted reaction mixture was added to the prehybridization solution of the individual blot. The virus quantity was calibrated by the estimation of protein content rather than RNA due to the unstability of the latter. The amount of viral protein refers to the quantity of proteins measured according to Bradford [5] . An amount of 100 pg viral proteins is estimated to correspond to about 1.14 x 105 virus particles or about 1.1 pg genomic RNA. 100 ng pUC-19 DNA was spotted as a control on the hybridization itself. Exposure was overnight at -70 °C 18 cells. Hybridization of virus-containing supernatants with probes p N 17 and p M 78, respectively, holding the N gene sequences of BCV and T C V resulted in similar detection signals (Fig. 4, lanes 1 and 3) , confirming the genomic relatedness between BCV and T C V in this gene as suggested by serological evidence. Three cultured isolates (i.e., 32, 35, and 42), identified as weak-positive by ELISA [7] were only slightly positive or negative by hybridization with either of the two nick-translated recombinant plasmid probes. Hybridization with a combination of the PXBAV-and P I O R F 1-primed P C R -p r o d u c e d ss probes showed a strongly increased detection sensitivity and resulted also in the identification of all 17 samples as positive (Fig. 4, lane 2) . 

Ten clinical samples, which were TCV-positive by ELISA only after virus cultivation in HRT-18 cell monolayers, were also positive in hybridization assays using one BCV-specific probe (p 52) on the supernatant fluid from the first passage (not shown). TCV in the original clinical specimens could hardly be detected with probe p 52 alone (not shown). Clinical diagnosis was attempted with either the pBCV-pool of recombinant plasmids to amplify the detection signal or with probes, synthesized by PCR as a result of labelling to specific activities about 20 times higher than nick-translated probes. A total of 134 clinical samples from birds with diarrhea were extracted with Freon to eliminate binding competition between the virus and macromotecules present in the samples for sites on the nitrocellulose. Treated samples were then spotted on nitrocellulose membranes for hybridization with BCV-specific probes (Fig. 5 ). Among these, 36 samples were identified as positive with ss PCRproduced probes, whereas 28 of these 36 were positive by hybridization with ds PCR-synthesized probes. Only 20 positive hybridization signals were obtained when the samples were probed with the pBCV-pool of recombinant plasmids. Three of these samples were not identified as positive with either of the two PCR-produced probes and may have been false-positives. However, similar blots, incubated with radioactive-labelled pUC-19-DNA to examine possible background hybridization with vector D N A did not reveal any significant hybridization detection signals, supporting the specificity of the signals obtained. An optimized ELISA [7] on all clinical specimens performed rather poorly (only 9 samples were identified as positive; data not shown), may be as a result of repeated freezing and thawing of the samples.

In the present study, genomic relatedness between BCV and TCV was demonstrated as BCV-specific probes to both structural and nonstructural genes could be used to detect isolates of TCV under relative stringent hybridization conditions. These results are in agreement with and extend previous findings on the serological relatedness between structural proteins of these two enteric coronaviruses as established by ELISA and Western immunoblotting [9] . Our BCV-specific probes could not detect the BCV-antigenically unrelated coronaviruses IBV and TGEV, whereas only slight detection signals were seen with MHV-3 (Fig. 2) , which is placed in the same antigenic group as BCV [20] . Shockley et al. [31] also reported a weak detection with another strain of MHV (MHV-A59), using a 32P-labelled insert probe corresponding to the BCV N and M genes [15, 31] . Their BCV-specific probe could in addition detect the serologically-related porcine hemagglutinating encephalitis virus and human coronavirus strain OC43.

Clinical diagnosis of TCV with a BCV specific 32p-labelted probe (p 52) was initially hampered by both the low concentration of virus in samples and by macromolecules, which clog the wells of the slot-blot apparatus and compete with the virus for binding sites on the nitrocellulose. In a previous report on the clinical diagnosis of BCV [33] , we investigated several options for the elimination of these macromolecutes in bovine diarrheic samples and for the amplification of detection signals. Similarly, TCV in clinical specimens, that were treated with Freon before spotting, could be detected when using the pBCV-pool of recombinant plasmid probes (Fig. 5) . Although not observed in this study, plasmid probes may in some cases cause difficulties in clinical diagnosis due to vector homology to plasmids in the sample [2, 10] . Therefore, PCR [12, 29, 30] was applied to synthesize probe molecules from inserts in linearized recombinant plasmids to reduce labelling of vector sequences. PCRproduced ss probes were more efficient than nick-translated recombinant plasmid probes in detection of purified TCV and extracellular virions present in TCV-infected cell cfilture supernatant fluids (cf. Figs. 3 and 4) . In combination, the PCR probes allowed the detection of a minimum amount of about 104 viral genomes (Fig. 3) . Hybridization signals obtained with spotted pUC-DNA (Fig. 3) were likely caused by molecules synthesized by polymerase-readthrough into vector sequences of incompletely digested templates. Detection of spotted pUC-DNA was, however,, significantly reduced with PCR-probes when compared to recombinant plasmid probes (Figs. 2 and 3) , improving the specificity of the detection assay. PCR-produced ss probes were superior to other probes used in clinical diagnosis, probably as a result of incorporation of label to high specific activities (109cpm/gg; [35] ) and the absence of self-aamealing among the probes. Double-stranded probes, synthesized in PCR with 1/3 of the dCTP being radiolabelled were efficient in BCV detection but resulted in strong background after only 8 h of autoradiography [35] . In our studies on detection of TCV, the background was reduced by addition of more "cold" dCTP for ds probe synthesis; in turn, however, fewer numbers of specimens were identified as positive. PCR-produced probes to the BCVN and M genes were, as in the detection of purified BCV [35] and TCV (Fig. 4) , also superior in clinical diagnosis of TCV (Fig. 5 ). In conclusion, BCV or TCV probes may be valuable in molecular hybridization for routine clinical diagnosis of either virus but may be problematic when these two viruses have to be distinguished. Studies are currently in progress to establish the nucleotide sequence of the TCV structural genes in order to determine the exact degree of genomic relatedness between the two viruses.

Isolation of transmissible enteritis agent of turkeys in avian embryos

The vector homology problem in diagnostic nucleic acid hybridization in clinical specimens

The use of a random priming procedure to generate cDNA libraries of infectious bronchitis virus, a large RNA virus

Nucleotide sequence of the glycoprotein Sgene of bovine enteric coronavirus and comparison with the S proteins of two mouse hepatitis virus strains

A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding

Coronaviruses associated with outbreaks of transmissible enteritis (Bluecomb) of turkeys in Quebec: hemagglutination properties and cell cultivation

Detection of turkey enteric coronavirus by enzyme-linked immunosorbent assay and differentiation from other coronaviruses

Isolation and trypsin-enhanced propagation of turkey entric (Bluecomb) coronavirus in a continuous human rectal tumour (HRT-18) cell line

Antigenic and genomic relationships among turkey and bovine coronaviruses

Cossart YE (t986) False-positive results with hepatitis B virus DNA not-hybridization in hepatitis B surface antigennegative specimens

A simple and very efficient method for generating cDNA libraries

Generation of single-stranded DNA by the polymerase chain reaction and its application to direct sequencing of the HLA-DQA locus

Studies on transformation of Eschericia coli with plasmids

Une lign6e cellulaire particuli~rement sensible fi le replication du coronavirus ent6tique bovin: les cellules HRT-18

t987) Sequence analysis of the bovine coronavirus nucleocapsid and matrix protein

Molecular cloning: a laboratory manual

Isolation and characterization of viruses associated with transmissible enteritis (Bluecomb) of turkeys

Cloning and in vitro expression of the gene for the E3 haemagglutinin glycoprotein of bovine coronavirus

Fluorescent antibody test for rapid diagnosis of coronaviral enteritis of turkeys (Bluecomb)

Antigenic relationships of the feline infectious peritonitis virus to coronaviruses of other species

Indirect fluorescent antibody test for the diagnosis of coronaviral enteritis of turkeys (Bluecomb)

Bluecomb disease of turkeys

Combined immunofluorescence and transmission electron microscopic studies of sequential intestinal samples from turkey embryos and poults infected with turkey enteritis virus

Coronaviral enteritis of turkeys

Labeling deoxyribonucleic acid to high specific activity in vitro by nick translation with DNA polymerase I

Electron microscopy of coronavirus-like particles characteristic of turkey Bluecomb disease

Terminal transferase-catalysed addition of nucleotides to the 3' termini of DNA

Enteric viruses in diarrheic poults

Primer-directed enzymatic amplification of DNA with a thermostabte DNA polymerase

The generation of radiolabeled DNA and RNA probes with polymerase chain reaction

Diagnosis of porcine and bovine enteric coronavirus infections using cloned cDNA probes

Tijssen P (t985) Practice and theory of enzyme immunoassays

Biotinylated and radioactive cDNA probes in the detection by hybridization of bovine enteric coronavirus

Genomic relationship between turkey and bovine enteric coronaviruses 211

Detection of bovine enteric coronavirus in clinical specimens by hybridization with cDNA probes

Polymerase chain reaction for probe synthesis and for direct amplification in detection of bovine coronavirus

This research was supported by the Conseil des Recherches et Services Agricoles du Qu6bec. AV acknowledges the support received from the World University Service of Canada. This report was taken in part from a dissertation submitted by AV to the Department of Virology, Institut Armand-Frappier, Universit6 du Qu6bec, in partial fulfillment of the requirements for the Ph.D. degree.

Authors' address: P. Tijssen, Centre de Recherche en Virologie, Institut Armand-Frappier, 531, boulevard des Prairies, Laval-des-Rapides, Qu6bec H7V 1B7, Canada.Received October 21, 1990