|
Archives of Virology Isolation and Characterization of a Canine Rotavirus
Abstract
Canine rotavirus particles were visualized by direct electron microscopy in the feces from a clinically-normM dog. The virus was subsequently propagated in cell cnltures; it was chracterized and compared with rotaviruses from other species. ICeplieation of the virus in cell culture was found to be less dependent upon trypsin than that of human, bovine and porcine rotaviruses. Reproducible, sharp-edged plaques of various sizes were produced by the canine rotavirus in an established ceil line of fetal rhesus monkey kidney, MA104, under overlays of carboxymethyl cellulose or agarose, [ntraeytoplasmic inclusion bodies of different sizes and shapes were produced in infected MA104 cells. By plaque reduction neutralization assay, a two-way antigenic relationship was found between the canine (CU-1) and simian (rhesus MMU 18006 and SA-1 t ) rotaviruses. The canine rotavirus had a one-way antigenic relationship with feline (Taka), bovine (NCDV), and porcine (OSU) rotaviruses. 8 Arch. ViroL 72/1 -2
t~otaviruses are among the m.ost important causative agents of acute neonatal gastroenteritis in various mammalian and avian species. Rotaviruses have been detected in the feces of man, monkey, horse, cattle, bison, antelope, sheep, goat, deer, pig, cat, rabbit, guinea pig, mouse, chicken, turkey, duck and parrot (4, 10, 11, 13, 18, 20, 28, 29, 31) .
Recently the worldwide appearance of parvovirus enteritis in dogs has caused an intensified search for infectious agents in dog feces. Subsequently, reports of visualization of rotavirus particles in canine feces, and in two cases cultivation porcine rotavirus (OSU) by Dr. E. H. Bohl, simian rotavirus (SA-11) by Dr. H. MMherbe, and simian rotavirus (rhesus MMU 18006) by Dr. N. J. Schmidt. The DS-1 strain of human rotavirus is a reassortant virus between a temperature-sensitive mutant of a cultivatable bovine rotavirus (UK strain) and a noncultivatable human rotavirus (DS-1 strain) (t7). All rotaviruses were plaque-purified prior to use.
Hyperimmune rabbit sera to canine and feline rotaviruses were prepared by intramuscular (i.m.) inoculation of 0.5 ml of purified virus mixed 0.5 ml Freund's incomplete adjuvant followed by two i.m. re-inoculations of 1.0 ml purified virus without adjuvant on days 21 and 28. Rabbits were bled one week after the last injection. All other hyperimmune antisera used were prepared in guinea pigs (17) .
Plastic six-well plates (Costar, Cambridge, Mass.) with confluent MA 104 cell monolayers were washed three times with Leibovitz L-15 medium (M. A. Bioproducts, Walkersville, Md.) supplemented with 0.5 percent gelatin-glutamine and antibiotics (L-15/gel) and inoculated with 0.5 ml of virus samples (pretreated for 60 minutes at 37 ° C with trypsin at a final concentration of I0 fxg/ml). The virus inoeulum was allowed to adsorb for 60 minutes at 37 ° C. Excess inoculum was then removed, plates were washed once with L-tb/gel and 5 ml of the overlay medium was applied to each well. The overlay medium was either: a) Eagle's basal medium, 0.75 percent carboxymethyl cellulose (CMC, Hercules, Inc., Wilmington, Del.), antibiotics, 100 tzg DEAE-dextran/ ml (Pharmacia Fine Chemicals, Piscataway, N. J.) and 0.5 izg 2× crystalline trypsin/ ml, or b) Eagle's minimum essential medium (MEM) containing 0.9 percent agarose, glutamine, antibiotics, 100 ~g DEAE-dextran/ml, and 0.5 ~zg 2 × crystalline trypsin/ml. Vc'ith the use of CMC overlay, plates were stained with formalin/crystal violet solution after 5 days of incubation in a humidified CO2 incubator. With the use of agarose overlay, plates were applied a second overlay containing neutral red (0.067 mg/ml) after 3 to 4 days of incubation. Plaques were then counted 1--3 days later.
Plaque reduction neutralization tests were performed by mixing equal volumes of fourfold dilutions of sera (previously heat-inactivated at 56 ° C for 30 minutes) and the virus suspension diluted to contain approximately 30 plaque-forming units (PFU) 8* per welt. After an incubation of 60 minutes at, 37 ° C, 1 ml aliquots of the virus-serum mixtures were inoculated onto the cell monolayers of 6-well plates. Two wells were used for each dilution. After 60 minutes of adsportion at 37 ° C, the inoculum was removed and the cell monolayers were washed once with L-15/gel, an overlay medium was applied, the cultures were incubated, stained and read as described above. The neutralizing titer was expressed as the reciprocal of the calculated serum dilution causing a 60 percent reduction in plaque counts (57) .
Neutralization test criteria of significance were arbitrarily established on the basis of enterovirus standard system (7, 23, 35) . If the difference between homologous and heterotogous neutralizing titers was tess than 20-fold, the serological relationship was considered to be significant; if the difference between titers was 20-fold or greater, viruses were considered to be significantly different.
Monolayers of MA 104 cells grown either on glass slides (Lab-Tek chamber slides, Lab-Tek Products, Napperville, Ill.) Or on coverslips in Leighton tubes infected with viruses, washed with PBS, fixed with acetone at --20°C for 15 minutes, airdried at room temperature, and then stored at --20°C until used. The fixed celt cultures were stained by the indirect immunofluorescent methods using rabbit or canine sera and the corresponding fluorescein isothiocyanate (FITC)-conjugated antigammaglobulin sofa or protein A-FITC conjugate (Pharmacia Fine Chemicals, Piscataway, N. J.). The slides were then eounterstained with a 1 : 500 dilution of 1 percent Evans Blue in PBS for t0 minutes and washed twice with PBS. After air -drying, slides were mounted with 50 percent glycerol in distilled water, and examined for specific flourescence with an epifluorescence UV microscope (Ernst Leitz, Ltd., Midland, Ontario, Canada).
In negatively stained preparations of the fecal sample, both double-shelled and single-shelled particles were observed (Fig. 1 ). The canine rotavirus particles were indistinguishable in morphology from those of known rotaviruses of other species. Double-shelled particles had sharply defined margins forming a continuous covering over the main capsid, which is characteristic for rotaviruses (26, 38) .
Smooth particles were approximately 70 nm in diameter and rough paI~icles were approximately 57 nm in diameter. Complete particles showed a characteristic "spoke-like" arrangement of inner capsomeres surrounded by an outer layer, which gave a honeycomb-like appearance on the surface of the virion.
The first cythopathie effect (CPE) was observed i day after inoculation (DPI).
Many cells adhered to the glass wall by only a single process ("flagging" CPE), which is a characteristic cytopathic effect produced in vitro by rotaviruses (33) and reoviruses (43) . As incubation progressed, these cells detached from the glass wall and floated free in the medium. The rotavirus isolate (CU-i) could be passaged in MA 104 cells without the aid of trypsin. Viruses were serially passaged more than ten times in MA 104 cells. Virus particles were demonstrated by EM in cell culture lysates in all passages.
Canine rotavirus (CUd) induced intracytoplasmic inclusion bodies in MAt04 cells. Inclusions were first noticed 24 hours post-infection. As incubation progressed, infected cells detached from the glass wall as noted above, however, inclusions were detectable in cells attached to the glass wall 5 DPI. Inclusions were of different sizes and shapes (Fig. 2) .
Reproducible and clear-cut plaques were formed 5 D P I under the overlays of CIVIC or agarose in the wesenee of trypsin (Fig. 3) . Although canine rotavirus readily replicated and produced CPE in monolayer cultures of MA104 cells in the absence of trypsin, detectable plaques were not formed even 8 D P I under the overlays of CMC or agarose without trypsin.
The antigenic relationship between canine and other mammalian rotaviruses was studied by plaque reduction neutralzation assay (Table 2) . A two-way relationship between canine and simian (rhesus MMU 18006 and SA-11) rotaviruses was found, indicating that these viruses were similar, if not identical, by this assay. In contrast, canine rotavirus (CU-t) was distinct from two human rotavirus strains (Wa and DS-1). There was a one-way antigenic relationship between the canine and the feline, bovine and porcine rotaviruses; in each instance, serum. to the non-canine strain demonstrated the broader reactivity. Virus-specific immunofluoreseence was demonstrated with rabbit immune serum but not with preimmunization serum in the cytoplasm of MA104 cells infected with canine rotavirus (CU-1). In some infected cells, the areas of immunofluorescence were small and scattered in the cytoplasm (Fig. 4A) , while in other cases, the areas coaleseed to from large diffuse fluorescent foei (Fig. 4B) . A crossreactive antigen(s) was demonstrated in the celts infected with canine rotavirus by indirect immunofluorescenee test (IFT) using hyperimmune anti-bovine rotavirus serum. Canine rotavirus antigen was not stained by hyperimmune sera against reovirus type I, II, and III by IFT; neither did hyperimmune serum to rotavirus stain reovirus antigens. Fluorescence was not detected in the uninfected control MA 104 cells.
Infected cell lysates were treated with trypsin (final concentration of trypsin at 10 fxg/ml), incubated for 1 hour at 37 ° C and then directly assayed by plaquing. The PFU infectivity was enhanced approximately 100-fold.
In 1978 MCNULTY et al. (30) reported that antibody to rotavirus was detected by indirect immunofluorescent test (IFT) in 49 of 62 dogs (79 percent) from the Belfast area in Northern Ireland. TAKAHASm et al. (48) in 1979 reported that 10 of 18 dogs (55.6 percent) in Aomori prefecture of Japan had antibody against calf rotavirus as determined by complement fixation test (CFT). Since both IFT and CFT detect only group-specific antibodies, and rotaviruses from one species can infect another species (5, 8, 34, 36, 46, 50, 51, 52, 55) , it was not clear from these studies whether the antibodies in the dogs resulted from infection with distinctly canine rotavirus or with human or some other rotavirus.
EVGSTER and SIDWA (11) reported in 1979 that rotavirus was visualized by EM in feces from a 12-week-old puppy with diarrhea. Virus isolation and characterization were not reported. In 1980, E~-GLA~D and PosTo~ (9) reported visualization by EM and subsequent isolation in MDCK cells of a rotavirus from a dog with fatal neonatal diarrhea. They also reported that experimental inoculation of two 6-month-old beagles with purified intestinal contents did not result in clinical signs of illness or virus shedding. In 1981, FuLTO~ et al. (15) isolated a rotavirus from a newborn dog that died after having clinical signs of diarrhea. They reported that their rotavirus isolate, "designated LSU 79--36, may be a specific canine rotavirus or a rotavirus from another species".
The results of the present study confirm the existence of a canine rotavirus. This canine rotavirus (isolated in New York State) was found to be similar, if not identical, serologically by plaque reduction neutralization assay to simian rotaviruses isolated in California (47) and South Africa (Table 2 ). However, they are different in RNA eleetropherotype, in hemagglutination pattern and in plaque size (HostIINO, Y., KALICA, A. t~. et al., unpublished results). In addition, the relatedness of two rotaviruses based on serotype does not always correlate with the relatedness based on genotype (14, I-IosHINO, Y. et al., unpublished results). Further biological and biochemical studies, including hybridization analysis, on both canine and rhesus rotaviruses, are in progress. Whether these findings are unique to this specific isolate (CUd st~rMn) of canine rotavirus or this is a common characteristic of any canine rotavirus isolate is being investigated in our laboratory. Recently we isolated by cloning two small variants of feline rotavirus which had a two-way antigenic relationship with the CU-1 strain of canine rotavirus. Further studies on plaque variation of rotaviruses are underway. No antigenic relationship was found between this canine rotavirus and two human rotaviruses by plaque reduction neutralization assay (Table 2) , although we have observed that there is a one-way antigenic relationship between this isolate of canine rotavirus and another strain of human rotavirus (Hosm•o, Y. et al., unpublished results). Since dogs, as pets, have close contact with humans, it will be interesting to examine further the serological relationships between canine and human rotaviruses, and to look for evidence of interspecies transmission.
The negatively stained canine rotavirus particles were indistinguishable in morphology from those of known rotaviruses of other species. The spoke-like arrangement of inner capsomeres of double-shelled particles, with a sharply defined outer capsid margin, differs from reovirus which has a featureless layer located either over or among capsomeres of the main capsid (21) , and from orbivirus which has a fuzzy indistinct layer covering the main eapsid (53) .
Rotaviruses are, in general, difficult to isolate and to grow in cell cultures using conventional techniques. However, certain strains may be serially propagated in cell cultures if the virus inuculum is treated with proteolytic enzymes or if proteolytic enzymes are incorporated into the medium after infection (1, 2, 3, 6, 16, 40, 41, 44, 45, 49, 56) . ENGI, A~D and POSTON (9) reported that rotavirus observed in feces by E?¢[ from a diarrheal dog grew in cell cultures with the aid of trypsin, but did not grow in cell cultures without trypsin treatment. Our isolate, canine rotavirus CU-1, readily replicated in cell cultures without trypsin treatment. Both double-shelled and single-shelled virus particles were demonstrated by EM in cell culture lysates in all passages.
It has been reported that both simian and bovine rotaviruses induced intracytoplasmic inclusoin bodies which were small with clearly defined round or oval outlines, and they did not coalesce to from large bodies such as those seen in cells infected with reovirus (27, 54) . Canine rotavirus, CUd, produced intracytoplasmic inclusion bodies in different sizes and shapes. Some were small and multiple as have been reported for those induced by bovine and simian rotaviruses and some were large and coalesced like those induced by reoviruses. The size and shape of inclusion bodies may depend upon each virus-cell culture system. We observed large inclusions in bovine rotavirus/MA104 cell and simian rotavirus/MA 104 cell systems (unpublished observation).
Further serological studies will determine antibody prevalence to canine rotavirus and establish whether or not infection with this strain is associated with disease. If canine rotavirus (CU-1) is found to be of low virulence, it may be evaluated as a potential vaccine strain for use in homologous or even hetero]ogous hosts. The potential use of a heterologous bovine rotavirus for possible use as a vaccine for humans has been suggested previously (25, 55) .
|
