key: cord-0744033-50k0fgqx
authors: Herrewegh, A.A.P.M.; Mähler, M.; Hedrich, H. J.; Haagmans, B. L.; Egberink, H. F.; Horzinek, M. C.; Rottier, P.J.M.; de Groot, R. J.
title: Persistence and Evolution of Feline Coronavirus in a Closed Cat-Breeding Colony
date: 1997-08-04
journal: Virology
DOI: 10.1006/viro.1997.8663
sha: d6c20ed949cb8f78cf2c780ec747a60be3ad2aed
doc_id: 744033
cord_uid: 50k0fgqx

Abstract Feline coronavirus (FCoV) persistence and evolution were studied in a closed cat-breeding facility with an endemic serotype I FCoV infection. Viral RNA was detected by reverse transcriptase polymerase chain reaction (RT-PCR) in the feces and/or plasma of 36 of 42 cats (86%) tested. Of 5 cats, identified as FCoV shedders during the initial survey, 4 had detectable viral RNA in the feces when tested 111 days later. To determine whether this was due to continuous reinfection or to viral persistence, 2 cats were placed in strict isolation and virus shedding in the feces was monitored every 2–4 days. In 1 of the cats, virus shedding continued for up to 7 months. The other animal was sacrificed after 124 days of continuous virus shedding in order to identify the sites of viral replication. Viral mRNA was detected only in the ileum, colon, and rectum. Also in these tissues, FCoV-infected cells were identified by immunohistochemistry. These findings provide the first formal evidence that FCoV causes chronic enteric infections. To assess FCoV heterogeneity in the breeding facility and to study viral evolution during chronic infection, FCoV quasispecies sampled from individual cats were characterized by RT-PCR amplification of selected regions of the viral genome followed by sequence analysis. Phylogenetic comparison of nucleotides 7–146 of ORF7b to corresponding sequences obtained for independent European and American isolates indicated that the viruses in the breeding facility form a clade and are likely to have originated from a single founder infection. Comparative consensus sequence analysis of the more variable region formed by residues 79–478 of the S gene revealed that each cat harbored a distinct FCoV quasispecies. Moreover, FCoV appeared to be subject to immune selection during chronic infection. The combined data support a model in which the endemic infection is maintained by chronically infected carriers. Virtually every cat born to the breeding facility becomes infected, indicating that FCoV is spread very efficiently. FCoV-infected cats, however, appear to resist superinfection by closely related FCoVs.

the culture (Stohlman et al., 1979; Hirano et al., 1981; Holmes and Behnke, 1981; Mizzen et al., 1983; Hingley Coronaviruses (genus Coronavirus, family Coronaviriet al., 1994) . An alternative mechanism for coronavirus dae, order Nidovirales), common pathogens of mammals persistence in vitro involves the selection of resistant and birds, are enveloped RNA viruses with an unseghost cells, with viral replication being supported by a mented genome 27-32 kb in size (for reviews see Sidsmall percentage of susceptible cells (MacIntyre et al., dell, 1995; de Vries et al., 1997) . The 5 two-thirds of the 1989; Hofmann et al., 1990; Sawicki et al., 1995) . Persisviral genome are taken up by the pol gene encoding tent coronavirus infections in vivo have mostly been studthe POL1a and POL1b polyproteins from which the viral ied using mouse hepatitis virus (MHV) as a model syspolymerase is derived by proteolytic cleavage. During tem. Suckling rodents intracranially inoculated with a replication, a 3-coterminal nested set of mRNAs that sublethal dose of neurotropic MHV variants develop codes for the structural proteins S, E, M, and N and chronic demyelination with viral replication in the central for a number of presumptive nonstructural proteins is nervous system (Sorensen et al., 1980; Knobler et al., produced. Each of these mRNAs contains a short non-1982; Jackson et al., 1984; Parham et al., 1986 ; Perlman translated 5 leader sequence derived from the 5 end et Morris et al., 1989; Fleming et al., 1994) . From of the genome. such animals, virus has been isolated as late as 1 year Although generally associated with acute, self-limiting after inoculation (Knobler et al., 1982) . Few studies have enteric and respiratory infections (McIntosh, 1990) , coroaddressed the role of viral persistence during natural naviruses can establish persistent infections both in vitro coronavirus infection. and in vivo. During persistent infection of tissue culture Feline coronaviruses (FCoVs) generally cause mild encells, replication-defective viruses often accumulate. Preteric infections but also cause a rare, fatal immune-medisumably, these moderate viral dissemination through ated disease called feline infectious peritonitis (FIP; for a review see de . The ''enteric'' FCoVs and the disease-causing FIP viruses are genetically that the latter are virulence variants arising spontaneously by genetic analysis of FCoV shed in their feces, we have obtained formal evidence for viral persistence. Further-in FCoV-infected hosts Poland et al., 1996) . FCoVs can be allocated to two serotypes on the more, we show that FCoV is subject to immune selection during chronic infection and that chronically infected cats basis of in vitro neutralization (Pedersen et al., 1984; Hohdatsu et al., 1991a,b) . The type II FCoVs are thought to have may shed virus for at least 7 months. originated from RNA recombination events during which the spike gene of canine coronavirus was incorporated into MATERIALS AND METHODS FCoV type I genomes Vennema et Animals and clinical specimens al., 1995; Motokawa et al., 1996) .

Epidemiological studies suggest that an FCoV carrier Domestic short-hair cats (Felis silvestris felis catus) were state exists and that asymptomatic FCoV-infected cats bred and housed in the closed breeding colony of the may spread the infection to susceptible kittens, presum-Central Animal Facility at Medical School Hannover, Gerably via the fecal-oral route. Some of these kittens demany. This colony was free of ecto-and endoparasites, velop FIP subsequently (Addie and Jarrett, 1992) . Best feline leukemia virus, and feline immunodeficiency virus. evidence for a carrier state has come from an experiment Cats were vaccinated against infection with feline herpesviin which cats were infected with a sublethal dose of rus, feline calicivirus, and feline parvovirus. They were tissue culture grown FIPV and kept in isolation (Pedhoused in groups of 2-20 animals and could roam freely. ersen, 1987). To induce FIP, the cats were superinfected

The rooms were environmentally controlled and personnel with the immunosuppressive feline leukemia virus at varientering the cattery were required to wash their hands and ous times after isolation. From this work, it appeared that to wear overshoes and a gown. Cats were fed commercial FIPV could persist in the experimentally infected host for diets and water was provided ad libitum. The two cats that at least 4 months .

were placed in isolation were housed on different floors in Feline coronaviruses are notoriously difficult to isolate a separate building and tended by different animal caretakand to grow in tissue culture. To identify asymptomatic ers. The isolation regime included the wearing of gown, FCoV carriers and to monitor virus shedding, we therehead cover, face mask, overshoes and gloves. Plasma and fore developed a nested RT-PCR assay targeted to the fecal samples were collected of individual cats and stored highly conserved 3 nontranslated region (NTR) of the at 020Њ until analysis. FCoV genome. Using this assay, viral RNA was detected in the feces, tissues, and body fluids of cats with FIP Virus strains and sequence data Egberink et al., 1995; Addie et al., 1996; Fehr et al., 1996) . Interestingly, FCoV RNA was FCoV strains FIPV UCD1 and FECV 79-1683 were provided by N. Pedersen and J. Evermann, respectively, and also found in the feces, and occasionally in the serum, of asymptomatic cats, consistent with the notion that clin-grown in fcwf-D (felis catus whole fetus) cells as described previously (de Groot et al., 1987b) . Anti-FCoV ically healthy cats may shed FCoV. Here, we have studied the natural history and evolution of FCoV in a closed type I serum 701 and type II serum G73 were obtained from cats experimentally infected with FCoV strain FIPV cat-breeding facility. By placing animals in isolation and were detected by immunofluorescence, as described H360 3 M õ20 0 0

previously . sections (8 mm, cut at 020Њ) were fixed in acetone con-activity was detected using 0.003% H 2 O 2 and 0.5% 3,3-using a reverse transcriptase nested PCR (RT-nPCR) assay targeted to the 3 NTR of the viral genome, as diaminobenzidine in 0.05 M Tris-HCl buffer (pH 8.3). The described by Herrewegh et al. (1995) and outlined in Fig. preparations were counterstained with hematoxylin and 1. For RT-PCR detection of FCoV nucleocapsid mRNA, mounted. Organ sections of an SPF cat, tested FCoV-negatotal RNA was extracted from various organ samples as tive by serology and RT-PCR, were processed alongside to described . The RT reaction was serve as negative controls.

primed with p511 (Table 1) , followed by cDNA amplifica-Detection of FCoV RNA in feces, plasma, and tissues tion using p525 and p511. Subsequently, a seminested The presence of FCoV in fecal, plasma, and tissue or a nested PCR was performed with primer pairs p527/ p511 or p527/p510, respectively. The seminested RT-PCR samples was demonstrated by detection of viral RNA

FIG. 2. Chronic shedding of FCoV as monitored by RT-PCR. (A) Detection of viral RNA in feces by RT-PCR. Cats H324 and H419 were placed in strict isolation at Day 0, and fecal samples were collected and processed for RT-PCR amplification of the 3 nontranslated region of the FCoV genome. The results are shown in a graph with the horizontal axis representing the number of days in isolation. Long and short bars above the horizontal axis indicate detection of FCoV RNA by single or nested RT-PCR, respectively. The instances in which viral RNA was detected neither by single nor by nested RT-PCR are represented by short bars below the x axis. Cat H419 was kept in isolation for 300 days. Cat H324 was sacrificed at Day 124 as indicated by an arrow. (B) Antibody titers in the plasma of cat H324 and H419 during isolation. FCoV-specific antibody titers of cat H324 (squares) and H419 (circles) were determined by immunofluorescence assay. products from colon and rectum were cloned in the FCoV was inadvertently introduced. During the next 12 years, 60 cats died of FIP, 31 (50%) of which were be-pGEM-T vector (Promega Corp., Madison, WI) and sequenced. The variable 5 regions of the spike and 7b tween 4 and 5 months of age. The yearly incidence of FIP was approximately 5%. In the past 3.5 years, no cases gene were reverse transcribed using primer p627 and p449, respectively. cDNA of the spike gene was amplified of FIP have occurred.

To study whether FCoV was still present in the colony, using primers p626 and p627 for the first rounds of amplification followed by a nested step using primers p628 a serologic survey was performed on 42 clinically healthy cats, ranging in age from 3 months to 9 years. Of these, and p629. The variable region of the 7b gene was amplified with primers p202 and p449 followed by a nested 29 (71%) were seropositive as determined by IFA (Table  2 ). Sera taken from cats H324 and H338 neutralized FCoV PCR using primers p287 and p267. The PCR products were directly sequenced using the AmpliCycle sequenc-strain UCD1 (serotype I) but not strain 79-1683 (serotype II), indicating that the cats had been infected by a type ing kit (Perkin-Elmer/Roche, Branchburg, NJ). I FCoV.

To identify cats with an ongoing FCoV infection, samples of the feces and the serum were screened by RT-Multiple alignments of nucleic acid and amino acid PCR targeted to the highly conserved 3 NTR of the FCoV sequences were performed using the PileUp program genome ( Fig. 1 ; Herrewegh et al., 1995) . Of the 42 se-(University of Wisconsin), which scores the similarity belected cats, 20 (48%) had FCoV RNA in the feces, 5 (12%) tween every possible pair using a method similar to the in the plasma, and 11 (26%) in the feces and plasma one described by Higgins and Sharp (1989) . Pairwise (Table 2 ). There was no apparent correlation between genetic distances between nucleic acid sequences were virus shedding and the antibody titer or the age of the estimated using the DnaDist program and the two-pacats. rameter model of Kimura (1980) . Pairwise genetic distances between amino acid sequences were estimated with the ProtDist program using maximum likehood esti-Persistence of FCoV infection mates based on the Dayhoff PAM matrix (Dayhoff, 1979) .

Interestingly, when cats H304, H324, H326, H330, and Unrooted phylogenetic trees (cladograms) were con-H337 were examined 3 months after the initial analysis, structed using the neighbor-joining algorithm (Saitou and all but H337 (i.e., 80%) again tested positive for viral RNA Nei, 1987) . Bootstrap resampling of the data was perin the feces (not shown). This finding could be explained formed using the SeqBoot program (Felsenstein, 1985) either by recurrent infections or by viral persistence. To with 100 iterations. Potential antigenic sites were calcudistinguish between these possibilities, cats H324 and lated using the method described by Jameson and Wolf H419 were each placed separately and kept in strict (1988). The antigenic index threshold was set at §1.3.

isolation. Samples were taken from stools and plasma and analyzed for the presence of FCoV RNA. The results

are summarized schematically in Fig. 2A . Viral RNA was Occurrence of FCoV in a closed breeding colony readily detected in the feces during the first 4 months of isolation. Initially, a single PCR using primers p205 and The closed cat-breeding facility at the Medical School Hannover houses between 65 and 126 animals. In 1981, p211 ( Fig. 1 ) was sufficient to detect viral RNA in the (Fig. 1) , yielding a product of 177 bp, or to a nested and seminested RT-PCR targeted to the N mRNA (A), yielding products of 82 and 109 bp, respectively. Products were separated in 2% agarose gels, and Sau3AI-digested pUC 18 DNA was used as a molecular weight marker (lanes M). Sizes are given in basepairs. The origins of the tissue samples are indicated by the following abbreviations: D, duodenum; J, jejunum; I, ileum; C, colon; R, rectum; N, negative control. feces and all samples were FCoV positive. Later, the levels of viral RNA in their plasma at any time during the isolation period nor showed any sign of disease. more sensitive RT-nPCR was required and viral RNA was detected less frequently. For cat H419, the FCoV infection Viral persistence in the gastrointestinal tract appeared to wane after 4 months of isolation, but through RT-nPCR viral RNA was detected in the feces even after

Our findings indicate that FCoV can indeed establish asymptomatic chronic infections. To determine the site of 7 months of isolation. Neither of the cats had detectable Table 2 ). Fecal samples were collected and processed for RT-nPCR. Nucleotides 026-170 of ORF7b were amplified as indicated in Fig. 1 and the PCR products were sequenced directly by cycle sequence analysis. Only those nucleotides differing from the overall consensus sequence (Cons.) are depicted. Nucleotide changes leading to amino acid substitutions are boxed. a,b and c,d represent the consensus sequences of the FCoV quasispecies in fecal samples taken from cats H324 and 419, respectively. Animals were placed in isolation on Day 0. Fecal samples were collected on Days 0111, a /113, b 0, c and /94. d (B) Unrooted phylogenetic tree illustrating the evolutionary relationships of the FCoVs shed by the cats in the breeding facility to the FCoV laboratory isolates TN406, UCD1, UCD3, UCD4, Dahlberg, Wellcome, 79-1146, and 79-1683; the American field strain FECV RM; and the Dutch field strains C2461, C2490, C2494, CB02, CB07, and CB03. The tree was obtained using the neighbor-joining algorithm on the basis of nucleotide distances. H324 represents the overall consensus sequence from A. viral persistence, cat H324 was sacrificed after 124 days the colony. Another objective of this set of experiments was to obtain genetic evidence for FCoV persistence of isolation and tissue and organ samples were collected. Using the nested RT-PCR, FCoV RNA was detected in sev-and to exclude the possibility that the extended virus shedding during the isolation period was caused by acci-eral organs, including kidney, lungs, brain, tonsils, salivary glands, and bone marrow. However, viral RNA could be dental reinfections. Viral RNA was extracted from fecal samples, and sequences derived from the 5 ends of the detected by single PCR only in the samples of duodenum, jejunum, ileum, colon, and rectum, suggesting that the virus S and 7b genes ( Fig. 1) were amplified by RT-nPCR. The PCR products were directly sequenced using cycle was most abundant in these tissues.

Evidently, the RT-nPCR assay targeted to the 3 NTR sequence analysis, thus yielding the consensus sedoes not differentiate between the viral genome and the quence of the FCoV quasispecies shed by each individ-mRNAs (Fig. 1) . To test for FCoV replication, we perual animal. The nucleotide sequences and deduced formed a semi-nested RT-PCR specific for the mRNA amino acid sequences were compared by multiple seencoding the nucleocapsid protein (N), using oligonuclequence alignment and subjected to phylogenetic analyotide primers p525, p527, and p511, which were desis (Figs. 5 and 6 ). signed after the FCoV 5 common leader sequence and A previous genetic comparison of 11 FCoV isolates from the N gene (Table 1 various origins revealed 83-94% overall nucleotide sede Groot, unpublished). An RT-PCR product of the prequence identity in ORF7b, with most sequence variation dicted size of 109 bp was obtained only for samples occurring in the 5-most 150 residues (Herrewegh et al., taken from the colon and rectum (Fig. 3B) . Sequence 1995). Ten of eleven FCoVs sampled from cats in the breedanalysis of this product yielded the sequence of the ing facility shared 99-100% sequence identity in this region. leader-body fusion region of the mRNA for N (not shown),

The FCoV shed by cat H039 appeared to be somewhat confirming the specificity of the RT-PCR assay. Using a more distant, displaying 96% sequence identity to the conmore sensitive RT-nPCR (Fig. 3A ), we could detect N sensus sequence (Fig. 5A) . A comparison to various FCoV mRNA also in the ileum but not in other tissues (Fig. 3B) .

laboratory isolates and European The presence of N mRNA strongly suggested that viral field strains (A. A. P. M. Herrewegh and R. J. de Groot, unreplication occurs in cells of the lower gastrointestinal published) showed that the FCoVs circulating in the breedtract. To corroborate our findings, cryostat tissue secing facility form one clade (Fig. 5B ). FCoVs sampled from tions of organ samples were tested for the presence another, commercial, breeding facility also clustered and of FCoV-infected cells by immunohistochemistry using a showed a similar degree of sequence variation (CB02, horseradish peroxidase-conjugated FCoV-specific anti-CB03, and CB07; Fig. 5B ). serum. Tissue sections taken from FCoV-negative spe-Among the FCoVs in the breeding facility, the secific pathogen-free cats served as negative controls.

quence variation in ORF7b was limited. We therefore Cells containing FCoV antigens were found only in the examined another region of the FCoV genome, i.e., the ileum, colon, and rectum of cat H324. In sections of the 5 end of the S gene (Fig. 1) . Comparative sequence ileum, these cells were located at the periphery of the analysis of various coronaviruses has shown that this Peyer's patches (Fig. 4A) , whereas in sections of the gene segment is highly variable (de Groot et al., 1987c ; large intestine, cells containing FCoV antigens were ob- Cavanagh, 1995) . Oligonucleotide primers designed after served facing the luminal side of the crypts of Lieberkühn the S sequences of the type I FCoV strain KU2 (Motokawa (Fig. 4B) . The combined findings of RT-PCR and histoet al., 1995) (Fig. 6A) . The average To further our understanding of FCoV epidemiology, percentage nucleotide substitutions is 2.3, and 91% of we performed a genetic analysis of the viruses present these substitutions results in an amino acid change ( Table 2 ). FCoV RNA extracted from fecal samples was subjected to RT-nPCR to amplify residues 51-531 of the S gene as outlined in Fig. 1 . The PCR products were sequenced directly and the nucleotides 79-478 were aligned. a,b,c,d and e,f represent the consensus sequences of the FCoV quasispecies in fecal samples taken from cats 324 and 419, respectively. The animals were placed in isolation on Day 0. In the case of cat 324, samples were collected on Days 0111, a /2, b /8, c and /113. d Those from cat 419 were collected on Days 0 e and /94. f Only those nucleotides differing from the overall consensus sequence (Cons.) are shown. (B) Unrooted phylogenetic tree based on comparison of the S nucleotide sequences illustrating the evolutionary relationship of the FCoVs isolated from the cats in the breeding facility. The tree was obtained using the neighbor-joining algorithm on the basis of nucleotide distances.

Thr and Ser, respectively. In a phylogenetic analysis, the FCoVs sampled from cat H324 cluster and, with the exception of FCoV sampled from cat H326, are more closely related to each other than to the FCoVs sampled from other cats. Given the overall genetic diversity in this region of the S gene among the FCoVs in the cattery, these findings support our conclusion that cats H419 and H324 carried an asymptomatic chronic FCoV infection and argue against accidental reinfections during the isolation period.

Coronavirus epidemiology, persistence, and evolution were studied in a closed cat-breeding facility with an endemic FCoV infection. Serological and genetic analysis (Fig. 8 ) revealed that the virus involved was a serotype I strain, closely related to the FCoV isolate Dahlberg (Fig.  5B ). An initial survey showed that 86% of the cats had an ongoing FCoV infection, as demonstrated by RT-PCR detection of viral RNA in feces and/or plasma. When tested 100 days later, four of five cats that had previously been identified as virus shedders still had detectable viral RNA in the feces. One explanation was that the cats in the breeding facility were subject to frequent reinfections. Alternatively, as previously speculated by us and others Addie and Jarrett, 1992; Herrewegh et al., 1995; de Groot and Horzinek, 1995) , FCoV may cause chronic infections resulting in prolonged virus shedding. We now provide the first direct evidence for FCoV persistence. By placing cats in strict isolation and by comparative sequence analysis of excreted virus, it was shown that naturally infected asymptomatic cats may remain infected and shed FCoV in their feces for at least 7 months.

Through the use of a highly sensitive RT-nPCR targeted to the 3 NTR , FCoV RNA was detected in several tissues of the chronically inscattered throughout the gene segment but rather ocfected cat H324. The lower part of the gastrointestinal curred at seven sites, six of which coincided with potentract was identified as a major site of viral replication, tial antigenic sites as predicted by a Jameson-Wolf analas indicated by RT-PCR detection of viral mRNA and by ysis (Jameson and Wolf, 1988 ; Fig. 7B ).

immunohistochemical detection of FCoV-infected cells. As illustrated by the cladogram in Fig. 6B , several clus-The failure to detect viral mRNAs or infected cells in ters of more closely related FCoVs can be distinguished.

organs other than the intestinal tract does not exclude In most cases, these viruses were sampled from litviral replication in these tissues. Rather, the number of termates, such as cats H349 and H350, cats H319 and infected cells may have been below detection level. In H314, and cats H324 and H326 ( Fig. 6B; Table 2 ). As the large intestine, infected cells were found lining the shown in Fig. 6A , the consensus sequences of the FCoV luminal side of the crypts of Lieberkühn. Conceivably, quasispecies shed by cat 419 at Days 0 and 94 postisolathese infected enterocytes could represent a main tion were identical. In the case of cat H324, the consensource of FCoV present in the feces. In the ileum, FCoVsus sequence of the virus shed at 111 days before isolainfected cells were located at the periphery of the Peyer's tion was identical to that of the quasispecies sampled patches, a mucosa-associated lymphoid tissue. FCoV at Day 2 postisolation. At Day 8 postisolation, a single may have gained access to these sites via the so-called point mutation was found resulting in an Asn 52 r His M cells, which are in close contact with intestinal substitution at site D (Fig. 7A) , whereas by Day 113 postlymphoid tissue and mediate the transepithelial transport isolation, two additional nucleotide substitutions were of macromolecules, particles, and microorganisms (for a review see Neutra et al., 1996) . Presumably, the infected detected, resulting in substitutions of Ala 90 and Asn 115 to cells seen in the Peyer's patches were monocytes. En-each time replacing the preexisting population, apparently as a result of immune selection. This is reminiscent terocytes and cells of the monocyte lineage are generally considered the main host cells of FCoV. Aminopeptidase of the genetic variation occurring in the HVR I of hepatitis C virus (Kurosaki et al., 1994; van Doorn et al., 1995) and N (APN), a protein expressed on the apical membrane of enterocytes, was recently identified as a receptor for in the variable regions of the envelope protein (Env) of the human immunodeficiency virus (Holmes et al., 1992; FCoV (Tresnan et al., 1996) . Monocytes also express APN (Griffin et al., 1981; Look et al., 1989; Delmas et al., 1992 ) Strunnikova et al., 1995 and feline immunodeficiency virus (FIV) (Rigby et al., 1993; Sodora et al., 1994) during and support replication of both pathogenic and nonpathogenic FCoVs in vitro (Pedersen, 1976b ; Jacobse-chronic infection. The endemic infection in the facility seems to be main-Geels and Horzinek, 1983; Stoddart and Scott, 1989) .

To determine the relationship between the FCoV types tained by chronically infected asymptomatic carriers, rather than through the repetitive occurrence of novel in the breeding facility and to study coronavirus evolution during chronic infection, a genetic analysis was per-FCoV variants that escape herd immunity. In fact, it would appear that infected animals develop some resistance formed on FCoV samples taken from individual cats. According to the quasispecies concept, (Eigen and Schus-against superinfection. Prior to isolation, cat H324 was exposed to other infected, virus-shedding cats over a ter, 1979; Eigen, 1971; Eigen and Biebricher, 1988; reviewed in Holland et al., 1992; Duarte et al., 1994;  period of at least 111 days. Yet, the H324 quasispecies was maintained, indicating that effective superinfection Domingo et al., 1996) , each FCoV sample does not constitute a genetically homogeneous virus population but by one of the other, genetically distinct, FCoVs in the cattery had not occurred. Similarly, the 6-year-old tom rather a cloud of variants with related, yet nonidentical genomes. Direct sequence analysis of RT-PCR-amplified cat H039 has been exposed repeatedly to different FCoVshedding queens. Nevertheless, the FCoV shed by this cDNA yields the consensus sequence of the quasispecies, which often but not always coincides with the pre-cat is quite unique and genetically farthest removed from the FCoVs sampled from the other cats, both in the dominant viral genome, the master sequence. This approach has the added advantage that sequence errors ORF7b and in the S regions. Conceivably, the considerable sequence divergence observed for H039 may have generated during reverse transcription or amplification will not be detected. Two regions of the genome were resulted from antigenic drift during prolonged chronic infection. Resistance to superinfection would also ex-selected for analysis: residues 7-146 of ORF7b, which appeared highly conserved among the FCoVs in the plain why littermates, which most likely were infected as kittens, still harbor genetically related FCoV populations breeding facility, and residues 79-478 of the S gene, which represent a hypervariable region (HVR). A phyloge-long after weaning. The data would fit a model in which the FCoV-specific immune response, though inadequate netic comparison to ORF7b sequences of independent FCoV isolates and field strains demonstrated that the to clear the infection, is vigorous enough to enforce immune selection and to prevent effective infection by FCoVs in the breeding facility form a separate clade, consistent with the notion that these viruses originated other, antigenically related, FCoVs. Antibodies crossneutralizing type I FCoV strain UCD1 were detected in from a single founder infection. Formally, however, multiple introduction of genetically related viruses cannot the sera of cats H324 and H338. It is of note, however, that FCoV-specific antibody titers were modest and re-completely be excluded.

In contrast to ORF7b, the 5 HVR of the S gene showed mained low throughout the chronic infection as monitored by immunofluorescence assay (Fig. 2B) . considerable sequence heterogeneity, and each cat in the facility appeared to carry a genetically distinct quasi-

The relationship between the asymptomatic carrier state and the development of feline infectious peritonitis remains species. The vast majority of the nucleotide differences (91%) were nonsynonymous. Sequence variation was not to be resolved. Genetic evidence implies that the avirulent FCoVs, which have been designated ''feline enteric coro-randomly distributed but confined to seven sites, six of which coincided with regions identified as potential anti-naviruses,'' and the FIP viruses are not separate species but merely virulence variants of the same virus (Herrewegh genic sites by computer-assisted analysis of the polypeptide sequence. The combined data suggest that the seet al., 1995; de Groot and Horzinek, 1995; Poland et al., 1996) . Actually, apathogenic FCoVs quence heterogeneity in the 5 HVR of the S gene results from antigenic drift. In cat H419, no changes occurred in readily give rise to disease-causing variants in hosts, immunosupressed by an FELV or FIV superinfection (Pedersen, the quasispecies over a 94-day period of chronic shedding. Similarly, the FCoV quasispecies of cat H324 re-1987; Poland et al., 1996) . In the field, FIP is a rare disease, predominantly occurring in cats younger than 1 year (Ped-mained invariant in this region of the genome over a period of 111 days prior to isolation. However, during ersen, 1976a; Addie and Jarrett, 1992; Kass and Dent, 1995) . Of the cats that died of FIP in the Hannover breeding facility, the subsequent isolation period, three nonsynonymous nucleotide changes occurred. The consecutive accumu-50% were 4-5 months of age. Apparently, these animals were unable to mount a protective immune response. Fail-lation of nonsynonymous nucleotide substitutions in the 5 HVR of S suggests a sequential emergence of variants, ure to control FCoV replication may increase viral load, thus Fig. 6A . Only those residues differing from the overall consensus sequence (Cons.) are presented. a,b,c,d,e,f are as in Fig. 6 . Potential N-glycosylation sites are underlined. (B) Jameson-Wolf prediction of potential antigenic sites in the N-terminal hypervariable region of S. The overall consensus sequence of the region formed by residues 27-160 of the S protein was analyzed using the Jameson-Wolf algorithm (Jameson and Wolf, 1988) . The results are plotted in a two-dimensional representation of the predicted secondary structure of this region with a helices indicated by a sine wave, b sheets by a sharp sawtooth wave, turns with 180Њ turns, and coils with a dull sawtooth wave. Regions with an antigenic index exceeding 1.3 are indicated by circles. The size of each circle is proportional to the value of the antigenic index.

FCoV strain UCD1 and serum 701, and H. Vennema for providing serum raising the odds that a pathogenic mutant is generated. In G73 and the FCoV RM sequence data, for helpful discussions, and for this view, both host and viral factors determine the outcome critically reviewing the manuscript. We are indebted to I. van der Gaag of an FCoV infection. and J. Mouwen of the department of Pathology for their assistance in interpreting the immunohistological data. R. J. de Groot was supported ACKNOWLEDGMENTS by a fellowship from the Royal Netherlands Academy of Arts and Sciences. We thank M. Reinacher for providing the horseradish peroxidasecoupled polyclonal anti-FCoV antibodies, N. Pedersen for providing

A study of naturally occurring feline coronavirus infections in kittens

Feline coronavirus in the intestinal contents of cats with feline infectious peritonitis

The coronavirus surface glycoprotein

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

Atlas of protein sequence and structure

Sequence analysis of the 3-end of the feline coronavirus FIPV 79-1146 genome: Comparison with the genome of porcine coronavirus TGEV reveals large insertions

Feline infectious peritonitis

cDNA cloning and sequence FIG. 8. Unrooted phylogenetic tree illustrating the relationship of the FCoVs in the breeding facility to type I and type II FCoVs and to CCV and analysis of the gene encoding the peplomer protein of feline infectious peritonitis virus

Intracellular RNAs of the feline infectious peritonitis of the FCoVs present in the Hannover breeding colony (see Fig. 7A) to that of the FCoV type I strains FECV RM and FIPV Ku2, the type II coronavirus strain 79-1146

strains FIPV 79-1146 and FECV 79-1683, the CCV strains Insavc-1 and K378, and the TGEV strains TFI and Purdue

Preparation of MAbs which discriminate between FIPV strain 79-1146 and FECV strain 79-1683

differences among arteri-, toro-and coronaviruses

Convergent and divergent sequence evolution in

Aminopeptidase N is a major receptor for the type 1 within a single infected patient

Basic concepts in RNA virus persistent infection in vitro

Analysis of a 9

RNA virus quasispecies: Significance for viral disease kb sequence from the 3 end of canine coronavirus genomic RNA. and epidemiology

FIP, of murine hepatitis virus (JHM) RNA from rats with experimental easy to diagnose?

Self-organization of matter and the evolution of biologi-Jacobse-Geels

RNA Genetics'' (E. Domingo, macrophage-like cells

The antigenic index: A novel algo-Raton, FL. rithm for predicting antigenic determinants

transmissible gastroenteritis coronavirus nucleocapsid protein gene. Detection of feline coronavirus using RT-PCR: Basis for the study of

The epidemiology of feline infectious Tierheilkd

A simple model for estimating evolutionary rates of using the bootstrap

Aiquences

Virus perin the central nervous system of mice inoculated with MHV-4. In sistence and recurring demyelination produced by a temperature

Expresand selection of hepatitis C virus variants in patients with chronic sion of myeloid differentiation antigens on normal and malignant hepatitis C

University of Bristol, Bristol, UK. GenBank Accession Herrewegh

Hu-RNA in feces, tissue, and body fluids of naturally infected cats by man myeloid plasma membrane glycoprotein CD13 (gp150) is identireverse transcriptase PCR

The molecular genetics of feline coronavimurine coronavirus infection involving maintenance via cytopathiruses: Comparative sequence analysis of the ORF7a/7b transcription cally infected cell centres

Coronaviruses. In ''Virology'' (B. N. Fields and D. M

Fast and sensitive multiple Knipe

MHV-A59 of coronavirus persistence

infection with mouse hepatitis virus JHM strain in DBT cell culture

Characterization of of the peplomer, integral membrane and nucleocapsid proteins of monoclonal antibodies against feline infectious peritonitis virus type feline, canine and porcine coronaviruses

Epithelial M cells: Gateways for mucosal infection and immunization

Analysis of JHM central nervous system infections in rats. human immunodeficiency virus type I evolutionary patterns

Serologic studies of naturally occurring feline of demyelinating disease. III. JHM virus infection of rats

Intrinsic resistance of feline feline infectious peritonitis virus and its growth in autochthonous peritoneal macrophages to coronavirus infection correlates with in peritoneal cell cultures

529-tion of the cold-sensitive hepatitis virus mutants rescued from latently 550. infected cells by fusion

Spaan, and man immunodeficiency virus type I in association with disease

Infection studies in tological demonstration of feline infectious peritonitis virus antigen kittens utilizing feline infectious peritonitis virus propagated in cell in paraffin-embedded tissues using feline ascites or murine monoculture

Feline aminopeptiof virus in the central nervous system of mice persistently infected dase N serves as a receptor for feline, canine, porcine, and human with murine coronavirus JHM

Sequence evolution of the immunocompromised cats infected with a feline enteric coronavirus. hypervariable region in the putative envelope region E2/NS1 of hepa

after feline infectious peritonitis virus challenge due to recombinant vaccinia virus immunization

406-A comparison of the genomes of FECVs and FIPVs and what they 425. tell us about the relationships between feline coronaviruses and their

Genomic organization and expression of the 3 end of

the canine and feline enteric coronaviruses

Identification of three feline immunodefigene of canine coronavirus and comparison with the S protein of ciency virus (FIV) env gene subtypes and comparison of the FIV and feline and porcine coronavirus