key: cord-270586-ohs8z91m
authors: Ballesteros, M. L.; Sánchez, C. M.; Enjuanes, L.
title: Two Amino Acid Changes at the N-Terminus of Transmissible Gastroenteritis Coronavirus Spike Protein Result in the Loss of Enteric Tropism
date: 1997-01-20
journal: Virology
DOI: 10.1006/viro.1996.8344
sha: 
doc_id: 270586
cord_uid: ohs8z91m

Abstract To study the molecular basis of TGEV tropism, a collection of recombinants between the PUR46-MAD strain of transmissible gastroenteritis coronavirus (TGEV) infecting the enteric and respiratory tracts and the PTV strain, which only infects the respiratory tract, was generated. The recombinant isolation frequency was about 10−9recombinants per nucleotide and was 3.7-fold higher at the 5′-end of the S gene than in other areas of the genome. Thirty recombinants were plaque purified and characterized phenotypically and genetically. All recombinant viruses had a single crossover and had inherited the 5′- and 3′-halves of their genome from the enteric and respiratory parents, respectively. Recombinant viruses were classified into three groups, named 1 to 3, according to the location of the crossover. Group 1 recombinants had the crossover in the S gene, while in Groups 2 and 3 the crossovers were located in ORF1b and ORF1a, respectively. The tropism of the recombinants was studied. Recombinants of Group 1 had enteric and respiratory tropism, while Group 2 recombinants infected the respiratory, but not the enteric, tract. Viruses of both groups differed by two nucleotide changes at positions 214 and 655. Both changes may be in principle responsible for the loss of enteric tropism but only the change in nucleotide 655 was specifically found in the respiratory isolates and most likely this single nucleotide change, which leads to a substitution in amino acid 219 of the S protein, was responsible for the loss of enteric tropism in the closely related PUR-46 isolates. The available data indicate that in order to infect enteric tract cells with TGEV, two different domains of the S protein, mapping between amino acids 522 and 744 and around amino acid 219, respectively, are involved. The first domain binds to porcine aminopeptidase N, the cellular receptor for TGEV. In the other domain maps a second factor of undefined nature but which may be the binding site for a coreceptor essential for the enteric tropism of TGEV.

to amino acids 522 to 744 of the spike protein were able to efficiently recognize the pAPN. Transmissible gastroenteritis virus (TGEV) is a mem-Since the S protein is responsible for the virus binding ber of the Coronaviridae family (Cavanagh et al., 1994, to the cell, it is expected that this protein would play an Enjuanes and Van der Zeijst, 1995; Siddell, 1995) . TGEV essential role in the control of the dominant tropism of replicates in both the villus epithelial cells of the small TGEV. Accordingly, there are data suggesting a correlaintestine and in lung cells of newborn piglets, resulting tion between tropism and S protein structure. Porcine in a mortality of nearly 100% (Saif and Wesley, 1992) . respiratory coronaviruses (PRCVs) have been originated Frequently these TGEV strains are referred to as enteric, independently in Europe (Callebaut et al., 1988 ; Pensaert as opposed to the respiratory strains which do not infect et al., 1986; Sá nchez et al., 1992) and in North America the enteric tract. Coronaviruses attach to host cells (Vaughn et al., 1994; Wesley et al., 1991 Wesley et al., , 1990b ) from through the spike (S) glycoprotein (Cavanagh et al., 1986;  enteric isolates (Enjuanes and Van der Zeijst, 1995; Sá n-Holmes et al., 1989; Sturman and Holmes, 1983; Suñé et chez et al., 1992) . PRCVs replicate to high titers only in al ., 1990) , and TGEV entry into swine testis (ST) cells the respiratory tract (Cox et al., 1990) and have a large is mediated through interactions between the virus S deletion at the 5 end of the spike gene, in positions glycoprotein and the porcine aminopeptidase N (pAPN) ranging from nucleotides (nt) 45 to 745 (Enjuanes and which serves as the cellular receptor (Delmas et al., Van der Zeijst, 1995; Sá nchez et al., 1992; Vaughn et al., 1992) . The S glycoprotein domain recognized by the cel-1994; Wesley et al., 1991) . Since this deletion is present lular receptor on ST cells is thought to be located spain all independently derived PRCVs it may be responsible tially close to the antigenic sites A and D (Suñé et al., for their loss of enteric tropism. 1990). In fact, recent studies (Godet et al., 1994) showed However, it can not be excluded that other viral genes, that baculovirus-expressed polypeptides corresponding apart from the S gene, could be involved in the determination of the tropism of TGEV. In fact, changes in ORF3a have been associated with the loss of enteric tropism.

nonical sequence CUAAAC required for the leader auer et Jimé nez et al., 1986) . The clone obtained was named PTV-ts-dmar1C.C12-1D.E7. The PTV-ts and primed transcription by the introduction of deletions and point mutations. These mutations lead to the lack of PTV-ts-dmar virus strains were grown at the permissive temperature (34Њ). ORF3a expression (Britton et al., 1990; Enjuanes and Van der Zeijst, 1995; Laude et al., 1993; Rasschaert et al., Virus neutralization, temperature inactivation, and 1990; Wesley et al., 1990a) .

purification Interestingly, the Purdue-type virus (PTV), which displays respiratory tropism, has an S gene with an identical

The procedure for virus neutralization has been desize to that of the enteric isolates. The PTV S gene was scribed (Correa et al., 1988; Suñé et al., 1990) . The neusequenced and compared with the homologous setralization index (NI) was defined as the log of the ratio quence of several enteric isolates. Only three nucleotide of the plaque-forming units (PFU) after incubating the differences, not observed in enteric isolates, were noted virus in the presence of medium or the indicated MAb. To and all introduced amino acid substitutions. Two of these analyze virus inactivation by temperature, viruses were changes were located at nucleotides 214 and 655 within grown at both the permissive (34Њ) and nonpermissive the area deleted in PRCVs, while the other was outside, (39Њ) temperatures. The temperature inactivation index at nucleotide 2098 (Sá nchez et al., 1992) . The nucleotide (TII) was calculated as the log of the ratio of the PFU change at position 214 was also present in several enafter growing the virus at 34Њ or 39Њ. teric isolates. These data lead us to propose that alter-To partially purify TGEV, ST cells were grown in 500ations in the S gene around residue 655 could affect cm 2 roller bottles and infected with a multiplicity of infecenteric tropism (Sá nchez et al., 1992) . tion (m.o.i.) of 10 PFU/cell. Supernatant was harvested In order to analyze the role of different viral genes in 48 hr postinfection (h.p.i.) and clarified by centrifugation tropism, we have isolated recombinant TGEVs by crossin a Sorvall GSA rotor for 20 min at 6000 rpm. Virions ing the enteric PUR46 strain and the respiratory strain were concentrated by centrifugation at 25000 rpm at 4Њ PTV-ts-dmar, a temperature-sensitive mutant (ts) resisin a Kontron TST28.18 rotor for 2 hr through a 31% (w/v) tant to neutralization by monoclonal antibodies specific sucrose cushion. To clear the virus from the remaining for two different antigenic subsites (dmar). Analysis of sucrose, the pellet was resuspended in TEN (10 mM the tropism of the recombinant isolates demonstrates Tris-HCl, pH 7.4, 1 mM EDTA, 1 M NaCl) and sedimented that two changes at nucleotides 214 and 655 of the spike by centrifugation under the same conditions. The viral gene, leading to aspartic acid to asparagine and to alapellet was resuspended in 500 ml of TNE (10 mM Trisnine to serine amino acid changes, respectively, were HCl, pH 7.4, 100 mM NaCl, 1 mM EDTA). associated to the loss of enteric tropism in the PTV isolate.

The binding of a large panel of MAbs to purified virus Cells and viruses was performed by RIA as previously (Correa et al., 1988; Sá nchez et al., 1990) using optimum amounts of virus Viruses were grown in swine testis (ST) cells (McClur-(0.5 mg of protein per well). kin and Norman, 1966). The virus strain PUR46-MAD-CC120 has been described (Sá nchez et al., 1992) . The RNA isolation Purdue virus strain PTV was previously named NEB72 (Sá nchez et al., 1992) ; however, due to sequence similar-Genomic RNA was extracted from partially purified virus as described previously (Mendez et al., 1996) . Briefly, ity to the PUR46 strain, its name was changed to PTV (Purdue-type virus). A ts mutant derived from PTV was ST cells from 10 roller bottles (500 cm 2 ) were infected with a m.o.i. of 5. Medium was harvested at 22 h.p.i. kindly provided by M. Welter (Ambico). This ts mutant was obtained after 5-fluorouridine mutagenesis using a Virions were partially purified as described above. The viral pellet was dissociated by resuspending in 500 ml previously described procedure (Robb et al., 1979) and three cycles of plaque purification. The ts mutant growth of TNE containing 2% SDS and digested with 50 ng of proteinase K (Boehringer Mannheim) for 30 min at room was reduced ú10 3 -fold at the restrictive temperature (39Њ) and showed a reduced capacity for RNA synthesis.

temperature. RNA was extracted twice with phenol-chloroform and precipitated with ethanol. Two neutralizing monoclonal antibodies (MAbs), 1C.C12 and 1D.E7, that are specific for the antigenic subsites RNA from TGEV-infected ST cells was obtained as described previously (McKittrick et al., 1993) . Briefly, ST Aa and Ab, respectively (Correa et al., 1988) , were used to select the neutralization escape mutant (double MAb-cells, grown in 8-cm 2 wells, were infected with TGEV at a m.o.i. of 5. At 22 h.p.i., the cells were resuspended resistant mutant, dmar) using the PTV-ts strain. The procedure used to obtain the dmar mutant was identical to in 400 ml of phosphate-buffered saline (PBS) and were incubated for 10 min on ice with 40 ml of 2 mM Vanadyl the one previously described (Correa et al., 1988; Geb- absence of neutralizing MAbs. A significant proportion of virions resulting from this infection should have only the spike protein encoded by one virus genome. The super-M9. The origin of the nucleotides in the recombinant natant was harvested at 12 h.p.i., and neutralization with viruses, in the positions indicated by these molecular MAbs 1C.C12 and 1D.E7 was performed to select recommarkers, were determined by sequencing RT-PCR-debinant viruses. Potential recombinants were plaque isorived cDNA fragments using the fmol DNA Sequencing lated at restrictive temperature in the presence of the two System (Promega). The primers used to sequence the neutralizing MAbs used in the selection. The surviving genetic markers M1-M9 are described in Table 1 . isolates were phenotypically characterized by calculating their neutralization and temperature inactivation indices.

Virus tropism analysis In parallel, independent ST cell cultures were infected with each of the two parental viruses and the same re-Viral tropism was determined in NIH miniswine (Luncombinant selection procedure was attempted. ney et al., 1986; Sachs et al., 1976) or in piglets derived from crossing Belgium Landrace and Large White swine.

Two-to three-day-old conventional (i.e., non-colostrumdeprived) piglets were used. Piglets were obtained from To identify nucleotide differences between the two pasows that were seronegative for TGEV neutralizing antirental viruses, cDNA fragments covering different rebodies by RIA. Inbred and outbred animals were oronagions of the genome were synthesized by RT-PCR.

sally and intragastrically inoculated with doses of 5 1 These regions included the first 1 kb from the 5-end of 10 7 and 5 1 10 8 PFU, respectively, in a final volume of the genome, ORF1 nucleotides 12208 to 20363, the first 2 ml of PBS supplemented with 2% of fetal calf serum. 5 2.3 kb of the S gene, and the most 4.3 kb 3 end of PTV-Groups of piglets inoculated with the same virus were ts-dmar. These cDNAs were subcloned into pBluescript grouped and housed in isolation chambers located in a (SK 0 ) (Stratagene) or pGEM-T (Promega) vectors. Plas-P3 level containment facility at 18Њ to 20Њ. Animals were mid DNA was purified using the FlexiPrep kit (Pharmacia) fed three times per day with 30 ml of milk formula for and sequenced with Sequenase 2.0 (USB). Sequence newborns (Nidina 1-Nestlé ). Virus titers at 1, 2, 3, and 4 data were compiled using the University of Wisconsin days postinoculation were determined in tissue extracts Genetic Computer Group (UWGCG) sequence analysis from jejunum and ileum and lungs. Tissues were homogsoftware package and compared to previously published enized at 4Њ using a tissue homogenizer Pro-200 (Pro-PUR46 virus strains (Eleouet et al., 1995; Kapke and Scientific) . Lungs and jejunum and ileum extracts were Brian, 1986; Mendez et al., 1996; Rasschaert et al., 1987;  obtained by homogenizing the whole organs in order to Sá nchez et al., 1992) . Mutations were confirmed by seobtain representative samples. quencing the viral RNA of the two parental viruses PUR46-MAD and PTV-ts-dmar.

RESULTS RNA was directly sequenced by oligodeoxynucleotide primer extension and dideoxynucleotide chain termina-Generation and characterization of TGEV mutant PTVtion procedures (Sanger et al., 1977) , using a modified ts-dmar1C.C12,1D.E7 protocol previously described (Fichot and Girard, 1990 ). Nucleotide differences between the parental viruses

To generate a panel of recombinants between enteric and respiratory strains of TGEV, parental viruses that were used as genetic markers and were named M1 to facilitate the selection of recombinants were generated markers, the first 1 kb from the 5 end of the genome, and characterized. A double mar mutant virus (PTV-ts-ORF1 nucleotides 12208 to 20363, the first 5 end 2.3 kb dmar1C.C12-1D.E7) derived from the respiratory PTV-ts of the S gene, and the most 3 end 4.3 kb were sestrain of TGEV was isolated. The growth of PTV-ts and quenced in the PTV-ts-dmar isolate. These sequences PTV-ts-dmar viruses in ST cells at 39Њ was at least 10 3were compared to the PUR46-PAR strain (Eleouet et al., fold lower than at 34Њ, while the parental PUR46 virus 1995; Rasschaert et al., 1987) and to PUR46-MAD (Menstrain replicated similarly well at both temperatures (data dez et al., 1996; Sá nchez et al., 1992) . Nine nucleotide not shown). The antigenic characterization of the mutant differences between the two parental viruses were iden-PTV-ts-dmar and its ancestor PTV-ts by RIA and neutraltified (Fig. 3A) . These markers were named M1 to M9 in ization using MAbs showed (Fig. 1) that the PTV-ts-dmar order from the 5-end of the genome. The PTV-ts-dmar mutant lacked the S protein antigenic subsite Aa comvirus was derived from a PTV-ts isolate which originated pletely, and subsite Ab partially, since some MAbs spefrom PTV. These three related isolates all display respiracific for these subsites did not bind or neutralized this tory tropism. All of them differ with the enteric PUR46isolate. In addition, the S protein of the escaping mutant MAD by three nucleotide substitutions in the S gene (M3, was not bound by MAbs 8B.F3 and 9F.C11, suggesting M4, and M6). In addition, PTV-ts-dmar differs by a fourth that the epitopes recognized by these MAbs are located nucleotide change (M5), which is responsible for the in subsites Aa or Ab or in close association with these dmar mutation. Two of the nine nucleotide differences subsites. Sequencing of the S gene in the PTV-ts-dmar (M8 and M9) did not result in an amino acid change. M8 mutant showed that the loss of antigenic subsites Aa was located in the intergenic region between ORF3a and and Ab was due to a single point mutation at nucleotide ORF3b, and M9 in ORF3b. Neither of these two areas 1756, resulting in a change of aspartic acid to tyrosine are expressed in PTV or PUR46-MAD strains. at position 586.

Analysis of the recombinant genome sequences

showed that the recombinants originated by fusing 5 sequences of the enteric parental virus to 3 sequences In order to study the molecular basis of TGEV tropism of the respiratory parental virus. Using these molecular recombinant viruses were obtained by coinfecting ST markers, the 30 recombinants were classified into three cells with the enteric PUR46-MAD and respiratory PTVdifferent groups according to the position of the crossts-dmar strains (Fig. 2) . ST cell monolayers were infected over (Fig. 3) . Group 1 recombinants contained 8 clones in parallel, with either PUR46-MAD or PTV-ts-dmar and their crossover was located within the 1102 nucleostrains. To isolate recombinants, selective pressure tides spanning the genetic markers M4 and M5 (Fig. 3 ). based on virus inactivation at high temperature and MAb Group 2 recombinants comprised 15 isolates that had neutralization was used. The virus titer in the ST cell culture coinfected with the two parental viruses was 1.9 the crossover located between M2 and M3 markers. The 1 10 3 PFU/ml, while cells infected with the respiratory viruses included in Groups 1 and 2 had the same seor enteric parental viruses contained 35 and less than 10 quence except for nucleotides 214 and 655 of the S pro-PFU/ml, respectively, indicating that recombinant viruses tein gene (genetic markers M3 and M4) that were derived resistant to the selective pressure were likely generated.

from the enteric parent in Group 1 isolates, and from the The supernatant from the ST cell culture coinfected respiratory parent in Group 2 isolates. Group 3 recombiwith the enteric and respiratory strains was used to nants included 7 isolates that had recombined between plaque purify 34 putative recombinant clones and the genetic markers M1 and M2. progeny of the coinfection was phenotypically character-Molecular marker M1 was sequenced in all recombiized under restrictive conditions. Thirty of the 34 clones nant viruses, since it was the nucleotide difference loanalyzed showed the recombinant phenotype (Table 2) . cated closest to the 5 end. All recombinants inherited Most of the progeny isolates (53%) showed the expected this marker from the PUR46-MAD strain, indicating that selectable recombinant phenotype (SR), being resistant there was most likely only one crossover at the 5-half of to both MAb neutralization and temperature inactivation. the genome. Three nucleotide differences were observed Among the recombinant viruses, 35% were dmar mutants from gene 3 up to the 3 end of the genome (Fig. 3A) . Only which were partially sensitive to the restrictive temperaone of these differences (genetic marker M7, located ture (ts intermediate phenotype recombinants, IPR). No in ORF3) led to an amino acid change. However, this virus was isolated with the phenotypic characteristics nucleotide difference was not present in the respiratory against which the selection had been performed (nonse-PTV strain, strongly suggesting that it was not involved lectable recombinants, NSR) ( Table 2 ).

in the control of TGEV tropism. The possibility of a second Genotypic characterization of the recombinant crossover at the 3 half of the genome was not analyzed, isolates because if a second crossover had taken place, it would have replaced a fragment by another one with an equiva-The identification of genetic markers was required to map the recombination sites. In order to identify such lent sequence. 

lar for the two viruses from each group. In addition, one isolate from Group 3 recombinants was studied. Tropism The tropism of the three groups of recombinants was was studied in parallel in the parental viruses PUR46, next studied. Two isolates from Group 1 and two from Group 2 were evaluated. The results obtained were simi-PTV-ts-dmar, and the PTV strain. Each isolate was tested at least three times. All recombinants and the parental shown) indicated that the ts mutation mapped at the ORF1a of the respiratory isolate, between genetic markviruses were isolated from the lungs, but only the PUR46 ers M1 and M2 (i.e., from nt 955 to nt 13272 of ORF1). strain and the recombinants of Group 1 could be isolated

The dmar mutation was localized at position 1756 of S from the small intestine (Fig. 4) . Maximum virus producgene (genetic marker M5). Thus, the minimum distance tion in the lungs varied from 10 4 to 10 6 PFU/g tissue (Fig. for recombination with the selective pressure used (high 4). PTV-ts-dmar produced less infectious virus than the temperature and neutralization by MAbs) was the interval other strains. Maximum PTV-ts-dmar production was between markers M2 and M5. The recombinant isolation about 10 4 PFU/g of lung tissue, while its ancestor PTVfrequency was calculated as the ratio between the progwt could replicate to higher titers (10 6 PFU/g tissue).

eny virus titer with a recombinant genotype (1.8 1 10 3 PFU) and the titer (8 1 10 7 PFU) of the parental viruses Recombinant isolation frequency grown in parallel in the absence of selective pressure, The procedure used to isolate recombinant viruses divided by the distance between M2 and M5. This frefavored the selection of viruses which had recombined quency was £2.3 1 10 09 recombinants per nucleotide between the two markers used in the selection, the ts for this interval. The recombinant frequency was also calculated for recombinants of Groups 1, 2, and 3 and and the dmar mutations. Preliminary results (data not Below this bar, the location of nine nucleotide differences (genetic markers M1 to M9) between the two parental viruses is indicated. Recombinants were classified into three groups (named 1 to 3), and the origin of their genomes, whether derived from the enteric parental (dark bars) or from the respiratory parental (white bars) is indicated. In the bars corresponding to the parental viruses PUR46-MAD and PTV-ts-dmar, the individual nucleotide differences are indicated. (B) Summary of the genetic characterization of the groups of recombinant viruses. The two markers flanking each crossover and the distance expressed in nucleotides between the two markers are indicated in columns 2 and 3, respectively. Since the exact location of the ts mutation is not known, the crossover in Group 3 recombinants is indicated as a maximum distance. The number of recombinants included in each group and the percentage in relationship to the total recombinant virus population are shown in column 4. The last column shows the frequency of recombinants isolated in each group, as the ratio of the number of isolates in the group relative to the number of nucleotides between the molecular markers flanking the crossover site.

were 5.5 1 10 09 , 1.5 1 10 09 , 4.1 1 10 010 , respectively residues 214 and 655 of the spike gene were responsible (Fig. 3B ). Since the exact location of the ts mutation is not for the loss of enteric tropism. known, the interval in which the crossover takes place in Group 3 recombinants can not be precisely defined and

Recombination among isolates of the TGEV cluster the frequency provided for this group is the minimum.

Initial attempts to isolate TGEV recombinant viruses The calculated data indicate that the frequency of recomusing selective pressure with neutralizing MAbs specific binant isolation at the 5 of the S gene was 3.7-fold higher for a single epitope lead to the isolation of neutralization than that of Group 2 recombinants. escape mutants instead of recombinants (data not shown). To diminish the frequency of escape mutants, DISCUSSION selective pressure with two MAbs specific for different epitopes of antigenic subsites Aa and Ab (Gebauer et In order to determine the role of different viral genes in TGEV tropism, a collection of 30 recombinants was al., 1991) were used. This strategy, in fact, decreased the frequency of neutralization escape mutants, although the generated by coinfecting ST cells with enteric (PUR46-MAD) and respiratory (PTV-ts-dmar) strains of TGEV. Phe-selected dmar mutant used as a parental virus in the recombination (PTV-ts-dmar) had only a single nucleotide notypic, genotypic, and biological characterizations of the recombinants showed that two nucleotide changes at change, instead of two nucleotide changes that might To study TGEV tropism, 2to 3-day-old non-colostrum-deprived piglets were individually inoculated with two isolates from Group 1, with two isolates from Group 2, and with 1 isolate from Group 3 recombinants. Recombinants were isolated by crossing the enteric strain PUR46 and the respiratory strain PTV-ts-dmar. Top horizontal thick bars indicate the genome and the origin of each recombinant, whether enteric (dark bar) or respiratory (white bar). The thin horizontal bar indicates the S gene. Triangles indicate the positions of nucleotides 655 and 2098 of the S gene and the origin of these nucleotides, whether enteric (dark triangle) or respiratory (white triangle). Diagram is not at scale and the size of the S gene has been magnified. The recovery of infectious virus was determined in PFU per gram of tissue at the indicated time in h.p.i. Each virus was tested at least three times. Vertical thin bars indicate standard error of the mean.

have been expected for a mar mutant escaping to the 3.7-fold higher than that of Group 2 recombinants. This could be due to a selective advantage in their growth on simultaneous neutralization by two different MAbs.

Isolation of TGEV recombinants required the use of cell cultures or to a higher recombination frequency at the 5-end of S gene, between nucleotides 655 and 1756. selection pressure. Using this procedure the frequency of recombination was estimated at £2.3 1 10 09 for TGEV.

In fact, extensive sequence variability has been observed in this region. During the isolation of TGEV-defective in-In contrast, the isolation of the MHV recombinant does not require the use of selection pressure (Makino et al., terfering viruses, a deletion is introduced at the beginning of the S gene, starting from nucleotides 6 to 74 and 1986). The recombination frequencies estimated for both coronaviruses are not directly comparable since the se-ending at ORF7 (Mendez et al., 1996) . In addition, during the generation of both European and American PRCVs lection strategies were very different; nevertheless, from reported data it seems that the recombinant isolation in field conditions, four different deletions at the beginning of the S gene have been identified in positions rang-frequency is higher for MHV than for TGEV. Both recombination and the generation of defective interfering (DI) ing from nucleotides 45 to 745 (Sá nchez et al., 1992; Vaughn et al., 1994; Wesley et al., 1991) . These data genomes occurs at a lower frequency in TGEV (Mendez et al., 1996) than in MHV (Lai, 1990) , possibly due to a suggest that the 5-half of the S gene is an area with an intrinsically high recombination frequency. Although a higher accuracy in the replication of TGEV RNA.

Group 1 recombinants were isolated at a frequency selective advantage for the recombinants could not be excluded, it seems unlikely because Group 1 and 2 re-tine until the fourth day postinoculation, virus was never detected in the enteric tract in any of the more than 60 combinants differ only in two nucleotide positions located at the 5-half of the S gene (nucleotides 214 and piglets inoculated with a respiratory isolate. This indicates that the virus detected in the enteric tract was not 655), and recombinants which had the same 5-half S gene as Group 2 recombinants grew as efficiently as due to residual virus from the inoculation, nor swallowed virus originating in the respiratory tract, but was the result PUR46. An increased recombination frequency in the S gene of MHV has also been described (Fu and Baric, of local virus replication in the intestine. All the isolated recombinants, including the ones lack-1994).

ing enteric tropism, were temperature resistant, indicating that the ts mutation was not responsible for the loss Molecular basis of TGEV tropism.

of enteric tropism. Studies on PUR46-PAR mar mutants also showed a Only nine nucleotide differences were found between the enteric PUR46-MAD and the respiratory PTV-ts-dmar correlation between the N-terminal half of the S protein and viral pathogenesis (Bernard and Laude, 1995) . Nev-strains of TGEV. Four of them mapped in the S protein gene at nucleotides 214, 655, 1756, and 2098. The nucle-ertheless, these results did not differentiate between virus tropism and virulence, since only parameters such otide change at position 1756 of S protein gene, which is responsible for the neutralization escape phenotype, as death, or weight loss, caused by the virus mutants were studied, but not virus replication in enteric or respi-is not responsible of the loss of enteric tropism since it was not present in the respiratory isolate PTV which ratory tissues. Coronavirus spike protein is involved in virus attach-lacks enteric tropism.

In order to analyze which of the other three nucleotide ment to cells (Cavanagh et al., 1986; Holmes et al., 1989; Sturman and Holmes, 1983; Suñé et al., 1990) . Studies changes located in the S protein gene, at positions 214, 655, and 2098, were involved in the control of the enteric on the inhibition of virus binding to cells indicated that the receptor binding site for TGEV had to be located tropism, recombinant viruses containing one or the three nucleotide differences from the respiratory isolate were between antigenic sites D and A of the spike protein (Suñé et al., 1990) , mapping between amino acids 385 selected. These recombinants belong to Groups 1 and 2, respectively. Group 2 recombinants only infected lungs, and 631. In agreement with these data, it was shown that porcine APN, the receptor for TGEV (Delmas et al., while Group 1 replicated in the epithelial cells of both the enteric and respiratory tracts. The two nucleotide 1992), binds to S protein residues between aminoacids 522 and 744 (Godet et al., 1994) . These sequences map changes between the enteric recombinants (Group 1) and the respiratory ones (Group 2) were at nucleotides to a distal area in relationship to amino acid 219 of S protein, which, as shown in this paper, influences TGEV 214 and 655 of the S protein gene, which caused amino acid changes from aspartic acid to asparagine at residue enteric tropism. Since pAPN is a protein present in lung epithelium and in enterocytes (Kenny and Maroux, 1982; 72 and from alanine to serine at residue 219. These results demonstrate that two amino acid changes at the Noré n et al., 1986; Semenza, 1986) , and the respiratory PTV isolate conserves the pAPN binding site previously N-terminus of the viral spike protein were associated to the loss of enteric tropism in the TGEV cluster of viruses.

described, the loss of enteric tropism in the PTV isolate should not be due to a failure in pAPN attachment. Fur-The possibility that the loss of enteric tropism was a consequence of the addition of a nucleotide change at thermore, it has been demonstrated that PRCV isolates attach to pAPN (Delmas et al., 1992) , although they can-position 655 of S gene to a preexisting change at nt 214 cannot be completely ruled out. Nevertheless, this not infect the enteric tract. This apparent discrepancy could be explained if an interaction between pAPN and possibility seems unlikely because most enteric viruses have the same nucleotide as PTV at position 214 (Sá n-two domains of S protein located at both areas (amino acids near residue 219 and amino acids 522 to 744) are chez et al., 1992), indicating that most likely a single nucleotide change at position 655 was responsible for required to infect the enteric tract. Alternatively, a putative second factor, such as coreceptor, mapping around the loss of enteric tropism. Nucleotides 214 and 655 are located within the area of the S gene which is deleted amino acid 219 of the spike protein could be specifically required to infect the enteric tract and responsible for in PRCVs, strongly suggesting that this deletion was responsible for the loss of enteric tropism in PRCVs. In the loss of enteric tropism in PTV and PRCV isolates. Other explanations are also possible and the loss on human immunodeficiency and other virus systems it has also been shown that a single point mutation can alter enteric tropism could also be due to: (i) a decrease in the pH stability required to allow the virus passage tropism (Takeuchi et al., 1991) .

An intragastric inoculation route was employed to as-through the stomach, (ii) a decrease in virion resistance to bile salts and proteolytic enzymes in gut, and (iii) an sure that the inoculum of each isolate was introduced into the stomach, independently of their tropism. While alteration in the strength or affinity of the S/receptor interaction. Recent studies also located the receptor binding viruses with enteric tropism have been found in the intes-

Cell Biology of Virus Entry, ble gastroenteritis virus spike protein results in markedly reduced Replication, and Pathogenesis'' (R. W. Compans, A. Helenius, and pathogenicity

Critical epitopes in transmissible gastroenteritis virus neutralrespiratory coronaries: Comparison with transmissible gastroenteriization

Sequence analysis of the porcine transmissible gastroenteritis coronavirus nucleocapsid protein gene

Genetic evolution and tropism of transmissible hydrolases of kidney and intestine

Coronavirus-organization, replication and expresgastroenteritis coronaviruses

Bullido, sion of genome

Antigenic homology among coronaviruses recoronavirus-molecular features and virus host interactions

The swine major histocompatibility complex: its structure and function. In ''Swine in chain-terminating inhibitors

Anchoring and biosynthesis of stalked brush border membrane proteins: glycosidases and peptidases of enterocytes

Highfrequency RNA recombination of murine coronaviruses

A monoclonal antibody

gastroenteritis of swine. II. Selected characteristics of a cytopathogenic virus common to five isolates from transmissible gastroenteri-Siddell

The molecular biology of corotis

Efficient naviruses

Mechanisms of transmissible gastroenteritis coronavirus neutralization

Analysis of the receptor binding site of murine coronavirus spike protein

Elsemodification of cell tropism by a single point mutation at the neutralization epitope in the env gene

Three new isolates of porcine respiratory coronavirus with various pathogenicities and gastroenteritis virus in swine: Old and news. 9th Int

Porcine respiratory coronavirus differs from transmissible gastroenteritis virus by a few for the pathogenesis of transmissible gastroenteritis virus

Genetic analysis of porcine respiratory coronavirus, an attenuated variant of transmispartial sequence of the genomic RNA, its organization and expression

Pathogenic murine coronaviruses. III. Biological and biochemical characterization of for a porcine respiratory coronavirus, antigenically similar to transmissible gastroenteritis virus, in the United States

spike protein-dependent cellular factor other than the viral receptor is required for mouse hepatitis virus entry

Transmissible Gastroenteritis

comparison of porcine transmissible gastroenteritis virus (TGEV) with site of the JHM strain of MHV in rats on the S1 subunit porcine respiratory coronavirus. VIIIth International Congress of Virolof the spike protein (Suzuki and Taguchi, 1996) , and in ogy, pp. P6-018. IUMS, Berlin.MHV it was suggested that a second cellular factor, apart Callebaut, P., Correa, I., Pensaert, M., Jimé nez, G., and Enjuanes, L. from the cellular receptor which interacts with the S pro- (1988) . Antigenic differentiation between transmissible gastroenteritis virus of swine and a related porcine respiratory coronavirus. J. tein, is involved in virus entry (Yokomori et al., 1993) . Gen. Virol. 69, 1725 -1730 The requirement of a coreceptor to infect cells has been Cavanagh, D., Brian, D. A., Enjuanes, L., Holmes, K. V., Lai, M. M. C., described in human immunodeficiency virus and in polio- Laude, H., Siddell, S. G., Spaan, W., Taguchi, F., and Talbot, P. (1994). virus (Deng et al., 1996; Dragic et al., 1996; Feng et al., Revision of the taxonomy of the Coronavirus, Torovirus, and Arteri-1996; Shepley and Racaniello, 1994) . ORF3a is not expressed in PRCVs isolates, while it is Coronavirus IBV: virus retaining spike glycopolypeptide S2 but not S1 expressed in enteric strains. Thus, it has been proposed is unable to induce virus-neutralizing or haemagglutination-inhibiting that ORF3a plays an essential role in the control of virus antibody, or induce chicken tracheal protection. J. Gen. Virol. 67, enteropathogenecity (Britton et al., 1991; Laude et al., 1435 Laude et al., -1442 Laude et al., . 1993 Wesley et al., 1991) . RNA sequence comparison of Correa, I., Jimé nez, G., Suñé , C., Bullido, M. J., and Enjuanes, L. (1988) .Antigenic structure of the E2 glycoprotein from transmissible gastro-the 3-half of the respiratory virus PTV-ts-dmar, from enteritis coronavirus. Virus Res. 10, 77-94.ORF3a to 3-UTR, with that of the enteric isolate PUR46-