key: cord-0684123-0e84tswn
authors: Paton, D.; Lowings, P.
title: Discrimination between transmissible gastroenteritisvirus isolates
date: 2014-04-05
journal: Arch Virol
DOI: 10.1007/s007050050191
sha: 8ad3378d2e8f82d5db1bafa7f4a16f938b27cad4
doc_id: 684123
cord_uid: 0e84tswn

Twenty TGEV isolates were compared by sequencing a 393–414 nucleotide stretch near the 5(′) end of the S gene, after amplification by RT-PCR. This part of the S gene is known to show considerable variation between porcine, canine and feline coronaviruses and is completely deleted from porcine respiratory coronaviruses. The discrimination achieved by nucleotide sequence analysis was compared with that obtained by monoclonal antibody typing. The viruses could be split into several clusters, and recent isolates of TGEV from England, The Netherlands and Belgium showed the greatest differences compared to earlier reference types. However, not all viruses with unique isolation histories were distinct, suggesting either genetic stability over many years, laboratory cross-contaminations or repeated introductions of similar viruses into the field. Firm conclusions on evolutionary trends cannot be drawn without obtaining a larger number of isolates, preferably from outbreaks with known epidemiological links. The sequences of some field isolates from the 1980s contained both nucleotide deletions and insertions. The latter included a short sequence of fourteen nucleotides with identity to a region of the TGEV polymerase gene.

Summary. Twenty TGEV isolates were compared by sequencing a 393±414 nucleotide stretch near the 5 H end of the S gene, after ampli®cation by RT-PCR. This part of the S gene is known to show considerable variation between porcine, canine and feline coronaviruses and is completely deleted from porcine respiratory coronaviruses. The discrimination achieved by nucleotide sequence analysis was compared with that obtained by monoclonal antibody typing. The viruses could be split into several clusters, and recent isolates of TGEV from England, The Netherlands and Belgium showed the greatest differences compared to earlier reference types. However, not all viruses with unique isolation histories were distinct, suggesting either genetic stability over many years, laboratory cross-contaminations or repeated introductions of similar viruses into the ®eld. Firm conclusions on evolutionary trends cannot be drawn without obtaining a larger number of isolates, preferably from outbreaks with known epidemiological links. The sequences of some ®eld isolates from the 1980s contained both nucleotide deletions and insertions. The latter included a short sequence of fourteen nucleotides with identity to a region of the TGEV polymerase gene.

* Transmissible gastroenteritis (TGE) is a highly contagious pig disease that may cause serious losses, especially in unweaned piglets [17] . The causative virus, TGEV, is a member of the Coronaviridae, and has a large, single-stranded, positive sense RNA genome. The virion is enveloped, and covered in projections caused by a glycosylated spike protein (S), of approximately 200 kDa. The S protein is a major target for antibodies, and monoclonal antibodies (mAbs) have helped to de®ne four principal antigenic sites, one of which (site A) is involved in antibody-mediated virus neutralisation [3] . Porcine respiratory coronavirus (PRCV) is a respiratory variant of TGEV, that has a deletion of over 670 nucleotides near the 5 H end of the S gene. This virus has become widespread in recent years, and its emergence has been associated with a reduction in the incidence and severity of cases of TGEV [14] .

Whilst numerous studies have examined antigenic differences between TGE viruses, only a small number of TGEV isolates have been compared genetically [16] . Here, we report nucleotide sequence comparisons of ampli®ed products from 20 TGEV isolates obtained from eight countries between 1946 and 1996 ( Table 1) .

Viral RNA for RT-PCR was extracted from cell lysates by the acid-phenol method [18] , or from culture supernatants using QIAamp HCV kits (Qiagen). Uninfected cell and water controls were processed concomitantly to check for unexpected ampli®cations. Establishment of the RT-PCR method has been described elsewhere (Paton et al., in press ). The target region of 886 nucleotides straddles a large deletion (672± 681 nucleotides) found exclusively in the S gene of isolates of PRCV, but not TGEV. Two mAb binding sites (B and C) have been mapped to this deleted region of the S protein [9] . The forward primer (F1121, 5 H -TATTTGTGGTYTTGGTYGTAATGC) is equivalent to nucleotide 11±344 of the S protein gene of TGEV, and the reverse primer (R1122, 5 H -GGCTGTTTGGTAACTAATTTRCCA) is complementary to nucleotides 896±873. Reverse transcription used random hexamers and M-MLV reverse transcriptase. The PCR used Taq polymerase and a temperature pro®le with 32 cycles (45 s at 95 C, 1 min at 50 C and 2 min at 72 C) followed by a ®nal extension time of 5 min at 72 C.

For each of 20 TGEV isolates, RT-PCR was performed in triplicate, and ampli®ed products were pooled and cleaned using Wizard PCR prep columns (Promega). The PCR templates were sequenced using a FS dye primer kit (Perkin Elmer) and internal primers F1174 (5 H -CGCAATAATAGTAATGACCTTTAT) and R950 (5 H -TCAAACCACCAAAGGTCTACAA), identical or complementary to S protein gene nucleotides 208±231 and 684±665. In instances when the sequence data was ambiguous the external primer, R1122 was also used. Sequencing reactions were visualised by the use of a commercial service (University of Durham, UK) which employed an ABI 373 automated sequencer (Perkin Elmer/ABI). Raw sequencing data was edited and aligned using the DNASTAR (DNASTAR Inc.) and GCG [4] computer packages. All sequences used for comparative purposes were con®rmed by sequencing in both directions. Aligned sequence data was compared with previously published data and phylogenetic trees were constructed by maximum likelihood, using the DNAML program from the PHYLIP package [6] .

For mAb typing. TGEV infected LLCPK1 cultures in 96 well microtitre plates were ®xed at 12±72 h post infection, before cytopathic effect became extensive. The ®xative, 20% acetone in PBS, was applied for 10 min and then plates were dried under a bench lamp at 37 C for 3±5 h. Immunostaining was carried out either immediately or after thawing plates stored at À20 C in sealed plastic bags. Optimised dilutions of antisera or mAbs were applied and these were detected by an appropriate anti-species peroxidase conjugate (Dako) and the substrate 3-amino 9-ethyl carbazole [10] . The mAbs were courtesy of D. Garwes (produced using TGEV 70±772; DA3, 4B1, 3C1, 7A6, 7C6, 9D2 [7, 8] ), L. Enjuanes (produced using TGEV Purdue; 1D.B12 [11] ) and A. van Nieuwstadt (produced using TGEV Purdue; 57.22, 57.110 [19] ). All isolates were recognised by mAbs DA3 (anti-nucleocapsid), 3C1, 7C6 and 9D2 (anti-S). The pattern of reactivity with mAbs able to discriminate between the isolates (all anti-S) is shown in Table 2 . Reliable typing results could not be obtained for the 96±1993 isolate due to its poor growth in cell culture.

The length of sequence obtained for each TGEV isolate ranged from 393 to 414 nucleotides, depending on the presence of insertions and or deletions. The new data were aligned with the published sequences of the prototype Purdue isolate (nucleotides 250±648) and that of TGEVs 70±772, Toy56, Neb72, Miller [16] and TF1 [2] , as shown in Fig. 1 . Our own sequence of Purdue, from a virus obtained via The Czech Republic (V91) differed at one nucleotide from the published sequence. Similarly, a previous sequence of 70±772 [1] differed at a single nucleotide compared to our own data. Four isolates had either additional or deleted nucleotides, none of which altered the reading frame of the S gene. The English isolate 83±3289 had a deletion of six nucleotides between position 286 and 291, overlapping antigenic site B [9] . The two Dutch isolates Erica and Slagharen, shared the same insertion of 15 nucleotides between positions 496 and 497. The two nucleotides preceding the insertion and the ®rst twelve nucleotides of the insertion itself, are identical in sequence to a region within the published polymerase Ia gene of the Purdue isolate of TGEV (nucleotides 1 054±1 065 [5] ). The last three nucleotides of the insert are the same as those at positions 494±496 which immediately precede the insertion. The Belgian isolate V63 had the most complex rearrangement, in which ®ve nucleotides were inserted, also after position 496, followed by a deletion of eleven nucleotides at position 497±507. The ®ve nucleotides inserted into V63 were the same as the ®rst ®ve nucleotides inserted into the Dutch isolates Slagharen and Erica. The 5 H end of the S protein of TGEV is that which differs greatest from canine and feline coronaviruses which otherwise have sequences very similar to those of TGEV. PRCV and other coronaviruses have been found to contain large deletions in this part of the S gene [14, 15] , whilst it has been reported that the S protein of the Purdue and Nebraska 72 isolates from America were two amino acids shorter than in other TGE viruses [16] . Our new data provide further evidence for the plasticity of this region of the S gene. The three isolates, Erica, Slagharen and V63 all share similar and has also been omitted from the alignment. The two boxed areas indicate areas of the gene which have been shown to encode parts of antigenic site B of the S protein [9] rearrangements to one another and are geographically and temporally linked, suggesting that the changes arose in vivo in a common precursor. In the case of the two isolates, Erica and Slagharen, template switching between the polymerase and S gene of the TGEV genome could account for the insertion, since there is a region of secondary structure predicted in front of the insertion site, and the same three nucleotides are present before the insertion and at its end. The greatest sequence dissimilarity was between isolates 96±1933 and Slagharen with 35 nucleotide sequence differences (91% identity), discounting the 15 nucleotide insertion in Slagharen. All of the isolates were more closely related to one another than to the related coronaviruses from dogs and cats. If insertions and deletions are excluded, approximately 60% of nucleotide positions where sequence differences are observed give rise to amino acid changes. Alignment of the deduced amino acid sequences showed greatest variability around amino acid positions 97 and 144, previously shown to be part of mAb binding site B [9] . Interestingly, mAb 1D.B12, previously shown to bind to site B, showed a different reactivity with two isolates (64±216 and 84±3 658) that had no sequence differences, suggesting that some other region may in¯uence this antigenic site. Figure 2 shows the similarity of the isolates to one another deduced by comparison of the nucleotide sequences. A number of clusters of similar The phylogenetic tree shown in Fig. 2 reveals a number of clusters of similar viruses, but in several cases there is no obvious correlation between clustering and the date or place of isolation. Some viruses are identical or nearly identical over the region examined despite being apparently isolated many years apart, and even from different continents. For instance the large group of viruses closely related to the prototype Purdue isolate. Other examples were IA-139 and Miller viruses and the 64±216, 70±772 and 84±3 658 viruses isolated in US and England respectively, in each case over a time period of approximately 20 years. To be certain that these identities did not arise from crosscontamination during RT-PCR procedures, we repeated the sequencing on all viruses appearing genetically identical, using freshly obtained amplicons derived from an earlier passage of the virus. Where possible, sequencing was repeated after direct ampli®cation from virus stocks as received from abroad. Further support that the similarities identi®ed were genuine, is provided by the mAb typing, which generally suggested the same virus groupings, and which was performed on the same virus stocks used for the RT-PCR. All of the above-mentioned Purdue-like group of viruses reacted identically with the mAbs, whilst the similarity of the three English viruses, 64±261, 70±772 and 84±3 658 was demonstrated by their unique reactivity with the mAb 7A6.

Other than extreme evolutionary stability, the reintroduction of laboratory isolates into the ®eld or the contamination of laboratory isolation cultures are possible explanations for the observed sequence identity of some isolates. In the case of the Purdue-like group of viruses, use in live vaccines could have caused reintroduction to the ®eld, whilst in vitro contamination could have been associated with use as laboratory reference strains, for instance in virus-serum neutralisation tests.

In contrast to these examples of apparent evolutionary stasis, the European viruses isolated in the last ten years are much more distinct, implying either a different origin or else a more rapid accumulation of mutations. It is tempting to speculate that a more rapid rate of evolution could somehow be caused by the spread of PRCV. However, since the sequenced region is deleted from PRCV, immunological cross-reactivity between PRCV and TGEV seems an unlikely cause. Since few outbreaks of TGE have been con®rmed in Britain in recent years, it was uncertain where the 96±1 933 isolate responsible for a recent episode of TGE [12] had come from. Interestingly, there is a unique nucleotide at position 426, present in all of the English isolates as well as in the sequence of TF1 from Taiwan, an isolate that appears phylogenetically closest to English isolate 83±3 289. This also suggests a common origin, but in the absence of a representative set of sequence data from viruses circulating elsewhere outside Britain, no ®rm conclusions on the immediate origin of any of these viruses is possible. As well as obtaining S gene sequence data from more viruses, there is a need to compare other parts of the TGEV genome using phylogenetic analysis, in order to check for evidence of recombination between strains, a known feature of coronaviruses [13] .

Sequence of the S gene from a virulent British ®eld isolate of transmissible gastroenteritis virus

Cloning and sequencing of a 8.4-kb region from the 3'-end of a Taiwanese virulent isolate of the coronavirus transmissible gastroenteritis virus

Antigenic structure of the E2 glycoprotein from transmissible gastroenteritis coronavirus

A comprehensive set of sequence analysis programs for the VAX

Complete sequence (20 kilobases) of the polyprotein-encoding gene of transmissible gastroenteritis virus

PHYLIP: phylogenetic inference package (version 3.2)

Identi®cation of epitopes of immunological importance on the peplomer of procine transmissible gastroenteritis virus

Differentiation of porcine coronavirus from transmissible gastroenteritis virus

Residues involved in the formation of the antigenic sites of the S protein of transmissible gastroenteritis coronavirus

Detection of antibodies against hog cholera virus and bovine viral diarrhoea virus in procine serum. A comparative examination using CF, PLA and NPLA assays

Critical epitopes in transmissible gastroenteritis virus neutralization

Classical transmissible gastroenteritis returns

Genetic recombination in RNA viruses

Procine respiratory coronavirus: molecular features and virus-host interactions

Sequence analysis reveals extensive polymorphism and evidence of deletions within the E2 glycoprotein gene of several strains of murine hepatitis virus

Detection of transmissible gastroenteritis virus by RT-PCR and differentiation from porcine respiratory coronavirus

Genetic evolution and tropism of transmissible gastroenteritis coronavirus

Transmissible gastroenteritis

Region-speci®c initiation of mouse mammary tumour virus RNA synthesis by endogenous RNA polymerase II in preparations of cell nuclei

Comparison of the antibody response to transmissible gastroenteritis virus and porcine respiratory coronavirus, using monoclonal antibodies to antigenic sites A and X of the S glycoprotein

We are very grateful to all those who supplied us with viruses and mAbs for use in this study. Viruses were given by L. Valicek, Brno, Czech Republic; A. van Nieuwstadt, Lelystad, The Netherlands; K. van Reeth, Gent, Belgium and M. Frey, Ames, USA. MAbs were given by L. Enjuanes, Madrid, Spain; D. Garwes, Compton, England and A. van Nieuwstadt. This work has bene®ted from use of the Seqnet facility.

Authors' address: Dr. D. Paton, Central Veterinary Laboratory, Woodham Lane, Addlestone, Surrey KT15 3NB, U.K.Received December 23, 1996