High titers of rotavirus antigen measured by an immunoassay were consistently detected in the stools of an immunodeficient pigtailed macaque, PFm-1, which was infected naturally with rotavirus and developed severe, chronic diarrhea. Rotavirus-like virus particles were detected by electron microscopy in stool extracts from this macaque (data not shown). Polyacrylamide gel electrophoresis of the viral RNA segments extracted from a stool specimen from PFm-1 revealed an electrophoretic pattern consistent with other rotaviruses, except for a lower intensity of segment 11 and an additional segment migrating slightly slower than segments 7, 8, and 9 (data not shown). This RNA electropherotype suggested that the stool extract was a mixture of subpopulations of rotavirus, as reported for other human rotavirus isolates with genome rearrangements [18,19]. The rotaviruses from PFm-1's stool extract were adapted to grow in MA-104 and plaque purified, which revealed two distinct viruses, named YK-1 and clone 311, that were distinguishable by plaque size and electropherotype (Figure 1a and 1b) [20,21]. YK-1 produced smaller plaques and had an electropherotype typical of group A rotaviruses and that was very similar to but distinct from RRV. Variant 311 produced larger plaques and had an RNA electropherotype identical to that for YK-1, except it had an additional segment that was migrating slightly slower than segments 7, 8, and 9 and did not have a typical migrating segment 11. Both YK-1 and 311 showed a cytopathic effect typical of rotavirus grown in MA-104 cells and readily grew to titers over 10₈ ffu per ml.

The nucleotide sequence of segment 11 from YK-1 was first determined as a reference. It consisted of 667 nt with a 594-bp ORF flanked by 5' and 3' UTRs of 21 and 52 nt, respectively. Segment 11 of YK-1 had 99% similarity to segment 11 of RRV. In the variant 311, segment 11 migrated slower than that of YK-1 and RRV, as determined by Northern blot analysis with a probe specific for segment 11 (Figure 1b). Sequence analysis of segment 11 from variant 311 identified a rearrangement consisting of a partial duplication of segment 11 from YK-1. The rearrangement occurred at nt 626 in the 3' UTR with the duplication of the ORF starting at nt 201 and included the entire 3' UTR (Figure 1c). The sequence of variant 311's segment 11 has 100% identity to YK-1's segment 11 both in the ORF and in the partial duplication.

The VP6 protein of rotavirus confers group specificity that is divided into seven groups (A to G). The commercial immunoassay Rotaclone utilizes a monoclonal antibody directed against the group A VP6 antigen and identified YK-1 as a group A rotavirus [22]. Group A rotavirus strains have been separated into four subgroups, and YK-1 is designated subgroup 1 as determined by reactivity with a subgroup 1 MAb 255/60 but not with subgroup 2 MAb 631/9.

To predict the G and P serotype specificities of the YK-1 strain, the sequences of the genes encoding both VP7 and VP4 were determined and compared with those of representatives of established G and P serotypes. The deduced VP7 amino acid sequence of YK-1 was closely related to other simian G3 rotavirus strains: 89% amino acid identity with RRV and 88% amino acid identity with SA11. A phylogenetic tree was constructed that included known VP7 amino acid sequences of G3 and other common G serotypes (Figure 2). The YK-1 strain clustered with other strains of G3 serotype, including RRV, SA11, AU1, and YO. YK-1 was also identified as a G3 serotype by reactivity with MAbs YO-1E2 (G3) and G3-159 (G3) in an immunoassay utilizing MAbs reactive to VP7-specific protein. The predicted VP4 amino acid sequence of YK-1 closely resembled that of GRV, a newly identified P[3]G3 caprine strain, and that of RRV P[3]G3 (Figure 3) [23].

The NSP4 gene segment of YK-1 was sequenced and its deduced gene product was compared with those of other known rotavirus strains. The structure of the NSP4 gene of YK-1 was similar to those of other rotavirus strains that were sequenced previously. The gene consists of a 528-bp ORF that encodes a protein with a predicted size of 175 amino acids with two conserved potential N-linked glycosylation sites. The deduced amino acid sequence of YK-1's NSP4 was 98% similar to that of the simian RRV strain. At least four genetic NSP4 groups are known, and the NSP4 of YK-1 can be classified as Group C by comparison of the amino acid sequence (aa 131–148) of the variable portion in the VP4 binding domain of various groups of NSP4 (Figure 4) [24,25].

The YK-1 strain of simian rotavirus was isolated from the diarrheal stool of a 2-year-old pigtailed macaque (Macaca nemestrina) housed at the Yerkes National Primate Research Center, Emory University (Atlanta, GA). This immunodeficent macaque, PFm-1, had chronic diarrhea associated with high titers of fecal rotavirus antigen detected by Rotaclone immunoassay (Meridian Diagnostics, Cincinnati, OH). The virus was isolated by previously described methods with modifications [20,21]. An extract from an antigen positive stool was prepared as a 20% (wt/vol) suspension in phosphate buffered saline (PBS, pH 7.4) and centrifuged twice at 8500 g for 10 min for clarification. The supernatant was extracted with 1,1,2-trichlorotrifluoroethane (Sigma, St. Louis, MO), and centrifuged at 4000 × g for 5 minutes. The extract was treated with tryspin (15 _g/ml) for 45 min at 37°C and inoculated onto a confluent monolayer of MA104 cells (African green monkey kidney cells) for 1 h. After being washed, the monolayer was maintained in serum-free minimal essential medium (MEM) (Gibco, Grand Island, NY) supplemented with 2 _g/ml tryspin and 50 _g/ml neomycin for 3 days. A viral lysate, obtained by freeze-thawing three times and clarification at 8500 g for 30 min, was inoculated into MA104 cells and plaque purified three times. Two distinct plaques distinguished on size were obtained and further passed in MA104 cells.

Virus stocks were activated with 15 _g/ml tryspin in MEM for 45 min at 37°C. The activated virus was 10-fold serially diluted in MEM and 500 _l/well was inoculated onto 6-well tissue culture plates (Corning, Corning, NY) with a confluent monolayer of MA104. After a 1 h incubation at 37°C, the inoculum was aspirated and 4 ml of a 3.5% agarose (Seakem, Biowhittaker, Rockland, ME) in MEM was overlayed on the monolayer. The agar was allowed to solidify at room temperature (RT), after which the plates were incubated at 37°C. Plaques were visualized by adding 1 ml of MEM with 2% neutral red and 0.3% agarose 6 h prior to reading.

Rotavirus RNA was extracted from stools and infected cell cultures by a modification of a previously described method [33]. In brief, a 10% stool extract or 30% cell-culture suspension was prepared with Tris-buffered saline supplemented with 1% sodium dodecyl sulfate (SDS), vortexed, and incubated at RT for 10 min. Equal volumes of virus and 1,1,2-trichlorotrifluoroethane were mixed for 1 minute, and centrifuged for at 8000 g for 10 min. The supernatant was added to 2 volumes of 6 M guanidine isothyiocyanate and incubated at 56°C for 10 min. Silica beads were added to each sample, vortexed, and incubated at RT for 10 min. The beads were washed once with a 2:1 solution of 6 M guanidine isothyiocyanate with 50 mM Tris-HCl (pH 7.5) and then three times with 70% ethanol. After the final wash, the beads were air dried, incubated with H₂O for 10 minutes at 65°C, and centrifuged for 2 min at 10,000 g. The extract was saved and stored at -70°C until use.

Rotavirus double-stranded genomic RNA extracted from fecal samples and cell-culture lysates were analyzed by SDS-polyacrylamide gel electrophoresis as described previously [34].

The procedures employed for Northern hybridization and chemiluminescent detection of bound digoxigenin-labeled probe using a commercial reagent (ECL, Amersham, Piccataway, NJ) have been described [35]. Two digoxigenin-labeled probes, ggcttttaaagcgctacagtgatgt and ggtcacaaaacgggagtggggagctcc, were used to identify genomic segment 11 of rotavirus.

Subgroup and VP7 serotyping were determined by use of a panel of monoclonal antibodies: 225/60 (Subgroup I), 631/9 (Subgroup II), KU-4 (G1), 5E8 (G1), S2-SG10 (G2), IC10 (G2), YO-1E2 (G3), G3-159 (G3), and ST-2G7 (G4) [36-38]. In brief, Immulon II plates (Nagle Nunc, Rochester, NY) were coated overnight with serum from a rabbit hyperimmunized with purified RRV rotavirus particles for positive wells and normal rabbit sera for negative wells. After the plate was washed with wash buffer (PBS plus 0.1% Tween 20), cell-culture lysates were added to duplicate positive and negative wells. The plates were incubated for 2 h at RT and washed. Specific monoclonal antibodies were added to wells, and the plates were incubated for 1 h at RT and then washed. Biotinylated goat anti-mouse IgG (Southern Biotechnology, Birmingham, Al) was added, incubated 30 min at RT and followed by washing and the addition of strepavidin-horseradish peroxidase (Southern Biotchnology). Wells were developed by adding tetramethylbenzidine (Sigma) and stopped after 10 min with 1 N HCl. A sample was considered positive if the OD value of the positive coated well was >2 times and 0.100 greater than the negative coated well.