Journal of General Virology (2000), 81, 1237 1249. Printed in Great Britain... Simian rhesus rotavirus is a unique heterologous (non-lapine) rotavirus strain capable of productive replication and horizontal transmission in rabbits Max Ciarlet, 1 Mary K. Estes 1 and Margaret E. Conner 1, 2 1 Department of Molecular Virology & Microbiology, Baylor College of Medicine, One Baylor Plaza, Mailstop BCM-385, Houston, TX 77030, USA 2 Veterans Affairs Medical Center, Houston, TX 77030, USA Simian rhesus rotavirus (RRV) is the only identified heterologous (non-lapine) rotavirus strain capable of productive replication at a high inoculum dose of virus (S 10 8 p.f.u.) in rabbits. To evaluate whether lower doses of RRV would productively infect rabbits and to obtain an estimate of the 50% infectious dose, rotavirus antibody-free rabbits were inoculated orally with RRV at inoculum doses of 10 3,10 5 or 10 7 p.f.u. Based on faecal virus antigen or infectious virus shedding, RRV replication was observed with inoculum doses of 10 7 and 10 5 p.f.u., but not 10 3 p.f.u. Horizontal transmission of RRV to one of three mock-inoculated rabbits occurred 4 5 days after onset of virus antigen shedding in RRV-infected rabbits. Rabbits infected at 10 7 and 10 5, but not 10 3, p.f.u. of RRV developed rotavirus-specific immune responses and were completely (100%) protected from lapine ALA rotavirus challenge. These data confirm that RRV can replicate productively and spread horizontally in rabbits. In attempts to elucidate the genetic basis of the unusual replication efficacy of RRV in rabbits, the sequence of the gene encoding the lapine nonstructural protein NSP1 was determined. Sequence analysis of the NSP1 of three lapine rotaviruses revealed a high degree of amino acid identity (85 88%) with RRV. Since RRV and lapine strains also share similar VP7s (96 97%) and VP4s (69 70%), RRV might replicate efficiently in rabbits because of the high relatedness of these three gene products, each implicated in host range restriction. Introduction Group A rotaviruses are the major aetiological agents of acute neonatal and infantile gastroenteritis. To better understand rotavirus immunity and to develop and test vaccines, several animal models have been developed (Conner & Ramig, 1996). The rabbit (Ciarlet et al., 1998a, b, c; Conner et al., 1988, 1991, 1993, 1997; Thouless et al., 1988) and mouse models (Burns et al., 1995; Feng et al., 1994; Franco et al., 1996; O Neal et al., 1997, 1998; Ramig, 1988; Ward et al., 1990, 1992) have expanded our understanding of rotavirus pathogenesis. In both animal models, homologous virus strains (isolated from the same species) replicate efficiently and spread horizontally Author for correspondence: Margaret Conner at Baylor College of Medicine. Fax 1 713 798 3586. e-mail mconner bcm.tmc.edu Sequence data reported in this work have been deposited in GenBank under accession nos AF084549 (ALA), AF084550 (C-11) and AF084551 (BAP-2). to uninoculated control animals, while heterologous virus strains (isolated from a different species) do not (Burns et al., 1995; Ciarlet et al., 1998a; Conner et al., 1988; Feng et al., 1994; Franco et al., 1996; Hoshino et al., 1995; McNeal et al., 1994; Ramig, 1988). The piglet model is different from the rabbit and mouse models because piglets are susceptible to infection with human rotavirus strains, perhaps due to the close relationship of genes between rotaviruses infecting these two species (Bridger et al., 1975; Hoshino et al., 1995; Ward et al., 1996). Different constellations of rotavirus genes encoding both structural and non-structural proteins (VP3, VP4, VP7, NSP1, NSP2 and NSP4) are implicated in rotavirus host range restriction (Bridger et al., 1998; Broome et al., 1993; Ciarlet et al., 1998a; Gombold & Ramig, 1986; Hoshino et al., 1995). Among all rotavirus molecular or antigenic features examined, such as RNA electropherotype, subgroup specificity, G serotype or P serotype (genotype), genome segment 5 (encoding NSP1) shows one of the highest degrees of sequence 0001-6669 2000 SGM BCDH
M. Ciarlet, M. K. Estes and M. E. Conner variation among rotavirus strains isolated from humans, pigs, horses, mice, cows, monkeys and cats (Dunn et al., 1994a; Fujiwara & Nakagomi, 1997; Hua et al., 1994; Kojima et al., 1996; Mitchell & Both, 1990; Palombo & Bishop, 1994; Xu et al., 1994). Although not universal, NSP1 sequence homology of over 40 rotavirus strains shows the best association with species of origin and can best determine interspecies relatedness within the overall genomic RNA constellation (or rotavirus genogroup) (Dunn et al., 1994 a; Fujiwara & Nakagomi, 1997; Hua et al., 1994; Kojima et al., 1996; Nakagomi & Kaga, 1995; Nakagomi & Nakagomi; 1996; Palombo & Bishop, 1994; Wu et al., 1993). Sequence analyses have confirmed this relatedness or revealed exceptions based on (i) close interspecies relationships between human and animal strains as per RNA RNA hybridization, (ii) examples of interspecies transmission of rotavirus strains (human Ro1845, feline FRV-1, equine H-1), or (iii) naturally occurring rotavirus reassortant strains (feline Cat2) between different species or genogroups (Dunn et al., 1994 a; Is a et al., 1996; Fujiwara & Nakagomi, 1997; Kojima et al., 1996; Nakagomi & Kaga, 1995; Palombo & Bishop, 1994; Xu et al., 1994). A notable exception is that the level of identity of NSP1 between simian rhesus rotavirus (RRV) and SA11 strains is low (57%) (Dunn et al., 1994a). Because RRV and SA11 are the only two simian rotaviruses isolated to date, it is difficult to speculate why these simian NSP1s are so diverse. In mice, gene 5 segregates with transmission of virus to uninoculated littermates (Broome et al., 1993; Gombold & Ramig, 1986), but in piglets, gene 5 was not implicated in rotavirus host range restriction (Bridger et al., 1998; Hoshino et al., 1995). In an attempt to identify the host range restriction determinants in rabbits, we recently tested 20 heterologous (non-lapine) viruses or reassortants for replication in rabbits (Ciarlet et al., 1998 a). Five additional heterologous rotavirus strains, human S2 (P1B[4], G2), avian Ty-1 (P[17], G7), avian Ch-2 (P[17], G7), bovine B223 (P8[11], G10) and porcine Gottfried (P2B[6], G4), and two reassortant rotavirus strains D UK (P7[5], G1) and ST3 UK (P7[5], G4) did not replicate in rabbits (unpublished results). Among all 27 rotaviruses tested, only the simian RRV strain was efficient at replication in rabbits when compared to the other heterologous (nonlapine) rotavirus strains (Ciarlet et al., 1998a). Although an inoculum dose of 2 4 10 p.f.u. of RRV results in a productive infection in rabbits (as measured by the magnitude and duration of virus antigen shedding and the magnitude of the immune response), the dose of RRV tested was 684 times greater than our standard challenge dose of lapine ALA (3 5 10 p.f.u.). Therefore, it remained to be determined if a lower dose of RRV would result in a productive infection in rabbits. The replication kinetics of a particular rotavirus strain in a host species are likely to be determined by an overall combination of genes, making it difficult to dissect the individual gene(s) which control rotavirus host range restriction (Broome et al., 1993; Burke & Desselberger, 1996; Ciarlet et al., 1998 a; Hoshino et al., 1995). Simian rotavirus strains SA11 and RRV share similar VP7 (96%) and VP4 (88%), but the level of identity of NSP1 (57%) is low (Dunn et al., 1994 a). Unlike RRV, SA11 exhibits limited replication in rabbits at high doses ( 10 p.f.u.) (Ciarlet et al., 1998a). High levels of VP7 (96 97%) and VP4 (69 70 %) sequence identity exist between RRV and the lapine strains (Ciarlet et al., 1997), but levels of identity with other gene products implicated in rotavirus host range restriction, particularly NSP1, have not been previously assessed because the corresponding gene sequences of the lapine rotavirus strains were not known. To determine the basis of the replication efficiency of RRV in rabbits, (i) rotavirus antibody-free rabbits were inoculated orally with 10, 10 or 10 p.f.u. of RRV to evaluate whether lower doses of RRV would productively infect rabbits and to obtain an estimate of the 50% infective dose (ID ), and (ii) the genes encoding the NSP1s of three lapine strains (ALA, C-11 and BAP-2) were sequenced and the predicted amino acid sequences were compared to all known NSP1 sequences. Methods Viruses. Lapine rotavirus ALA and C-11 (P[14], G3) strains were provided by M. Thouless (University of Washington, Seattle, WA). The origin of lapine rotavirus strain BAP-2 (P[14], G3) was described previously (Ciarlet et al., 1998c). All lapine rotavirus strains were plaquepurified three times and were cultivated in MA104 cells as described (Conner et al., 1988). The simian RRV strain [RRV-2 (MMU 18006)] (P5B[3], G3) was originally isolated from the faeces of a 3 5-month-old rhesus monkey with diarrhoea (Stucker et al., 1980) and a stock from the original isolate (Greenberg et al., 1983) was supplied by H. Greenberg, Stanford University, Palo Alto, CA. The lapine ALA strain, used for rabbit live rotavirus challenge, was passaged 10 times in MA104 cells prior to inoculation of rabbits as described (Conner et al., 1988). Virus titres were determined by plaque-forming assay or focus fluorescent assay (FFA) and expressed as p.f.u. ml or f.f.u. ml, respectively (Ciarlet et al., 1994; Conner et al., 1988). In vitro transcription, PCR, cloning and sequencing. Singlestranded RNA transcripts were prepared from purified double-layered ALA, C-11 and BAP-2 particles as described (Ciarlet et al., 1997). The nucleotide sequences of gene 5 of ALA, C-11 and BAP-2 were determined by PCR which generated products of 1597 bp, corresponding to the fulllength NSP1 of ALA and C-11, and a product of 1350 bp, corresponding to the NSP1 of BAP-2 lacking the 3 end of the gene. Briefly, reverse transcriptase was used to generate gene 5 complementary (c)dna using a primer complementary to the 3 end of gene 5 of the simian rotavirus strain SA11 (5 GGGTTCACAGTATTTTGCCAGC 3 ). Following purification of the first cdna strand, PCR of gene 5 was achieved with the same 3 end primer and a primer complementary to the 5 end of gene 5 of the simian strain SA11 (5 GGGCTTTTTTTTGAAAAGTCTTG 3 ). Amplified DNA was cloned into TA3pCR2.1 vector (Invitrogen). For accuracy in sequence determination, two independent clones from each rotavirus strain from individual PCR reactions were sequenced by the dideoxynucleotide chain or dye termination method. Additional primers were synthesized to complete the nucleotide sequencing of the genes. Their nucleotide positions on gene 5 of ALA, polarities (plus or minus sense) and sequences (5 to 3 ) were as follows: Max 150 (nt 509 532, BCDI
Basis of RRV infection in rabbits plus sense), CCAATCACTCTAAACGCTGCACTG; Max 330 (nt 1037 1060, minus sense), CTTACACTTGGAAATTGTCGAAGC; and Max 440 (nt 1355 1378, minus sense), AAGAGCTTCAGTTTTT- AATATCAT. Confirmation of the DNA sequence was performed by sequencing both DNA strands of each of the different clones and by sequencing parts of the viral mrna directly (Ciarlet et al., 1994) using all three minus sense primers described herein. The 3 end sequences of the ALA and C-11 NSP1 genes are complementary to the PCR primer. The coding 3 end nucleotide sequence of the NSP1 gene of BAP-2 was determined from in vitro transcribed mrna as described previously (Gorziglia et al., 1986), using the primer complementary to the 3 end of gene 5 of the SA11 rotavirus strain. The non-coding 3 end nucleotide sequence of the NSP1 gene of BAP-2 was not determined. Animals. Three-month-old rotavirus- and specific-pathogen-free New Zealand White rabbits of either sex were obtained from Charles River Laboratories, Canada. To prevent complications due to Clostridium infections, 3 weeks prior to initiation of rotavirus studies, all rabbits were vaccinated once intramuscularly with a Clostridium spiroforme toxoid (kindly supplied by R. Carman, TechLabs, Blacksburg, VA) (C. O Neal, M. Ciarlet, R. J. Carman & M. E. Conner, unpublished results). Animal inoculations and procedures. Rabbits were inoculated orally with 2 4 10, 2 4 10 or 2 4 10 p.f.u. (corresponding to 4 10, 4 10 or 4 10 f.f.u., respectively) of simian RRV, with 1 10 or 3 5 10 p.f.u. (2 10 f.f.u.) of lapine ALA rotavirus, or with PBS as described (Ciarlet et al., 1998 a; Conner et al., 1988). All rabbits were individually housed and maintained in open cages for primary and challenge inoculations in a BL2 containment facility at Baylor College of Medicine. Rabbits inoculated with RRV or PBS were housed in a single room, while rabbits inoculated with ALA rotavirus were housed in a separate room. Experiments with ALA rotavirus and RRV were never performed concurrently. Cross-contamination of rabbits by different virus strains used in different experiments has never been encountered because each animal room and its entire contents are either heat- or chemically sterilized between experiments. Additionally, strict policies and procedures were followed for housing and handling the rabbits to prevent possible cross-contamination, and care and maintenance of the rabbits was performed by a single caretaker not exposed to other laboratory animals. Transmission efficiency of RRV in rabbits was also monitored by its ability to spread horizontally to mock-inoculated control animals housed in the same room and in the same rack, but in different cages from RRV-inoculated rabbits. In the previous study, transmission of RRV could not be assessed because the rabbits were housed in individual isolators under negative pressure (Ciarlet et al., 1998a). All rabbits were challenged orally with 10 infectious doses (ID ) (3 5 10 p.f.u. ml) of lapine ALA rotavirus 28 days postinoculation (p.i.) as described (Ciarlet et al., 1998a; Conner et al., 1988). Collection of samples. For detection of rotavirus-specific antibody responses, serum and faecal samples were collected at 0 and 28 days p.i. and 28 days post-challenge (p.c.) as described (Ciarlet et al., 1998a; Conner et al., 1988). Processing of faecal samples for antibody detection was performed as described (Ciarlet et al., 1998a). To detect virus shedding, faecal samples were collected 0 10 days p.i. and 0 14 days p.c. and were processed as described (Ciarlet et al., 1998a; Conner et al., 1988). Gastrointestinal transit time in rabbits. To measure gastrointestinal transit time in rabbits, 3-month-old fasted rabbits (n 4) were inoculated orally with 10 ml of sterile H O containing 30 small (1 5 mm) radiopaque markers using a blunt-ended feeding needle (Popper & Sons, New Hyde Park, NY). Faecal pellets were collected twice a day for 5 days, and transit time through the gastrointestinal tract was assessed by counting excreted markers at each time-point following radiography of the faecal material. ELISA to measure rotavirus excretion and total antibody responses. The ELISAs to measure rotavirus excretion and to measure total (IgA, IgM, IgG) antibody to rotavirus responses were performed as described (Ciarlet et al., 1998a). A positive reaction was defined as an absorbance reading at 450 nm (A ) 0 1 for the antigen ELISA or a value 0 1 after subtracting A values of the antigen-negative well (mock) from the antigen-positive well for antibody ELISA (Conner et al., 1988). Analysis of rotavirus infection in rabbits. Detection of virus antigen shedding after inoculation of heterologous viruses was defined as productive if faecal virus antigen shedding was comparable in magnitude and duration to homologous lapine ALA rotavirus infection, whereas limited replication was defined as when the amount and days of shedding was approximately two- to fourfold lower than the values obtained with homologous strains (productive infection) and antibody conversion occurred (Ciarlet et al., 1998a). Minimal replication of a virus in rabbits was defined as antibody conversion after inoculation of the heterologous strain, but undetectable virus antigen shedding. Total protection from challenge was defined as no faecal shedding of virus antigen as detected by ELISA. Analysis of RNA electropherotype by PAGE. To confirm that rabbits shed only the virus with which they were inoculated, 10% faecal suspensions from virus-inoculated rabbits which were positive for rotavirus by ELISA or FFA were tested by PAGE. Nucleic acids of representative input and recovered virus from faecal material were extracted and subjected to electrophoresis in a 7% polyacrylamide gel and genome segments were visualized by silver staining (Ciarlet & Liprandi, 1994). FFA to determine infectious virus titres and FFNA to measure neutralizing antibodies. FFAs were performed as described (Ciarlet et al., 1998b) on faecal samples collected following heterologous or homologous rotavirus inoculation of rabbits to compare infectious virus shedding relative to ELISA virus antigen shedding. Total infectious rotavirus shedding in faecal samples was calculated by the addition of individual titres of infectious virus in positive faecal samples and expressed in f.f.u. ml. An estimate of the net yield of virus output was calculated by multiplying the total infectious virus shed in faecal samples by the minimal volume (50 ml) of the entire (small, large and caecal) intestinal contents as determined previously (Conner et al., 1988). Focus fluorescent neutralization assays (FFNAs) were performed as described (Ciarlet et al., 1994), with the endpoint determined as the serum dilution producing a 66% reduction in the number of fluorescent foci. Statistical analyses. Statistical analyses of antibody titres between groups and correlation coefficients were performed by the Mann Whitney U-test and by Pearson s correlation coefficient using SPSS version 7.5 for Windows (SPSS, Chicago, IL). Results Infection of rabbits inoculated with different doses of RRV Although two of two rabbits inoculated with 2 4 10 p.f.u. of RRV were previously shown to be productively infected with RRV (Ciarlet et al., 1998a) (Fig. 1A), only one of four rabbits inoculated with 2 4 10 p.f.u. of RRV was productively infected and shed virus antigen for 5 days from 5 to 9 days p.i. (peak of virus shedding, 7 days p.i.) (Fig. 1B). BCDJ
M. Ciarlet, M. K. Estes and M. E. Conner Fig. 1. Faecal virus antigen shedding curves of rabbits inoculated with various doses of heterologous simian RRV or homologous lapine ALA rotavirus. Each line represents one rabbit. Rabbits were inoculated with (A) 2 4 10 8 p.f.u. simian RRV (n 2) (Ciarlet et al., 1988 a), (B) 2 4 10 7 p.f.u. simian RRV (n 4), (C) 2 4 10 5 p.f.u. simian RRV (n 4), (D) 2 4 10 3 p.f.u. simian strain RRV (n 4), (E) 1 10 7 p.f.u. lapine ALA rotavirus (n 2), and (F) 3 5 10 5 p.f.u. lapine ALA rotavirus (n 2). Faecal rotavirus antigen shedding was assessed by ELISA from 0 to 10 days p.i. and expressed as net A 450 readings. Readings 0 1 (above horizontal line) are considered positive. Limited replication, in comparison to lapine ALA productively infected rabbits (Fig. 1E, F), was observed in the other three rabbits inoculated with 2 4 10 p.f.u. of RRV (Fig. 1B); the level and the duration of faecal virus antigen shedding was reduced to 3 or 4 days, with a mean duration of faecal shedding of 3 3 days (Fig. 1B). In the group of rabbits (n 4) inoculated with 2 4 10 p.f.u. of RRV, productive, limited or no infection of rabbits was observed (Fig. 1C). One rabbit shed faecal virus antigen from 4 to 8 days p.i. (peak of virus shedding, 7 days p.i.), two rabbits had detectable faecal virus antigen which lasted 2 or 5 days, whereas one rabbit did not have detectable virus antigen shedding. There was a 1- to 2-day delay in onset and peak of RRV antigen shedding (Fig. 1B, C) compared to results from ALA-infected rabbits (Fig. 1E, F). None of the rabbits (n 4) inoculated with 2 4 10 p.f.u. of RRV had detectable faecal virus antigen shedding (Fig. 1 D). To facilitate comparisons of RRV virus antigen shedding to homologous lapine ALA virus antigen shedding, inoculation of rabbits with ALA at analogous doses were also performed. Following inoculation of two rabbits each inoculated with 1 10 or 3 5 10 p.f.u. of ALA rotavirus, rabbits shed virus antigen for 5 6 days from 2 to 8 days p.i., with a peak occurring at 3 or 4 days p.i. (Fig. 1E, F). Previously published results with three rabbits inoculated with 1 10 p.f.u. of ALA showed that productive infection was observed in two rabbits (faecal virus antigen shedding for 5 or 6 days), while one rabbit BCEA
Basis of RRV infection in rabbits Fig. 2. Faecal infectious virus shedding curves of rabbits inoculated with homologous lapine ALA rotavirus or with varying doses of heterologous simian RRV. Each line represents infectious virus shedding for one rabbit. Rabbits were inoculated with (A) 3 5 10 5 p.f.u. (2 10 5 f.f.u.) lapine ALA rotavirus (n 2), (B) 2 4 10 7 p.f.u. (4 10 6 f.f.u.) simian RRV (n 4), (C) 2 4 10 5 p.f.u. (4 10 4 f.f.u.) simian RRV (n 4), and (D) 2 4 10 3 p.f.u. (4 10 2 f.f.u.) simian strain RRV (n 4). Faecal infectious rotavirus shedding was assessed by FFA from 0 to 10 days p.i. and expressed as f.f.u./ml. When fluorescent foci in 1 : 10 dilutions could not be visualized by fluorescence microscopy, the samples were considered negative, and a value of 50 f.f.u./ml was assigned. had no detectable virus antigen shedding (Conner et al., 1997) (data not shown). To compare infectious virus shedding relative to ELISA virus antigen shedding, faecal samples collected following ALA or RRV inoculation of a subset of rabbits were tested by FFA. The titres of homologous ALA at the peak of virus shedding ranged from 1 75 10 to 6 10 f.f.u. ml in rabbits inoculated with 3 5 10 p.f.u. (2 10 f.f.u.) of ALA (Fig. 2A). Peak titres of infectious heterologous RRV shed by the rabbits productively infected with either 2 4 10 or 2 4 10 p.f.u. (4 10 or 4 10 f.f.u., respectively) ranged from 5 10 to 4 10 f.f.u. ml, while those of the rabbits that exhibited limited infection ranged from 2 10 to 6 10 f.f.u. ml (Fig. 2B, C). In the subset of rabbits that exhibited productive RRV infection, RRV antigen shedding was similar in magnitude and duration to that obtained with lapine ALA. In fact, infectious virus titres at the peak of virus shedding were similar. For example, a rabbit inoculated with 2 4 10 p.f.u. of RRV that was productively infected shed 3 10 f.f.u. of RRV at 7 days p.i. (Fig. 2B), while a rabbit inoculated with 3 5 10 p.f.u. of ALA shed 4 5 10 f.f.u. of ALA at 4 days p.i. (Fig. 2A). This was less than a twofold difference in titre. More limited replication or virus shedding was observed in another subset of RRV-infected rabbits; there was at least a two- to fourfold decrease in magnitude and duration of virus antigen shedding and up to a several log decrease in the peak infectious virus titre compared to titres obtained with homologous lapine ALA. One of the four rabbits inoculated with 2 4 10 p.f.u. of RRV had detectable faecal infectious virus shedding at 7 days p.i. (Fig. 2 D). As seen previously with homologous ALA rotavirus infection (Ciarlet et al., 1998b), detection of virus antigen shedding by ELISA and detection of infectious virus by FFA correlated (P 0 001, r 0 988, Pearson s correlation coefficient). These data further support our definition of productive versus limited infections in rabbits (Ciarlet et al., 1998a, b). To confirm that RRV replication, and not degradation or excretion of virus input, occurred in rabbits, the transit time through the rabbit gastrointestinal tract and an estimate of the net yield of virus output were determined. Radiopaque markers administered orally to 3-month-old fasted rabbits (n 4) were excreted in faeces between 10 and 26 h p.i. (data not shown). Therefore, since gastrointestinal transit time in rabbits is approximately 24 h and RRV was shed 4 6 days p.i., RRV antigen shedding must reflect virus replication, not degradation or excretion of virus input. The net yield of infectious virus was estimated knowing that the minimal volume of the entire BCEB
BCEC Table 1. Rotavirus infection of rabbits inoculated orally with homologous (lapine) ALA (P[14], G3) rotavirus and different doses of heterologous (nonlapine) RRV (P5B[3], G3) Inoculum Species of origin Dose of inoculum Rabbit Number of days of virus antigen shed (ELISA) Number of days of infectious virus shed (FFA) Range of titres* of excreted virus in faecal samples (f.f.u./ml) Total titres of excreted virus in faecal samples (f.f.u./ml) Net yield of virus (f.f.u.) Seroconversion No. of responders / inoculated rabbits Type of rotavirus infection M. Ciarlet, M. K. Estes and M. E. Conner ALA Lapine 3 5 10 p.f.u. 1 5 6 4 5 10 5 5 10 1 2 10 6 10 Yes 2 2 Productive (2 10 f.f.u.) 2 6 8 3 10 1 8 10 2 3 10 1 2 10 Yes Productive RRV Simian 2 4 10 p.f.u. 1 5 6 5 8 10 7 10 8 8 10 4 4 10 Yes 2 2 Productive (4 10 f.f.u.) 2 5 6 1 6 10 7 10 1 4 10 6 8 10 Yes Productive RRV Simian 2 4 10 p.f.u. 1 5 6 3 10 4 10 6 7 10 3 3 10 Yes 4 4 Productive (4 10 f.f.u.) 2 3 3 1 5 10 6 10 1 7 10 8 5 10 Yes Limited 3 3 4 1 5 10 6 10 6 5 10 3 3 10 Yes Limited 4 4 4 1 10 1 4 10 6 4 10 3 2 10 Yes Limited RRV Simian 2 4 10 p.f.u. 1 3 6 4 5 10 2 4 10 7 8 10 3 9 10 Yes 4 4 Limited (4 10 f.f.u.) 2 2 3 3 10 2 5 10 4 1 10 2 1 10 Yes Limited 3 5 7 1 10 2 10 2 9 10 1 5 10 Yes Productive 4 0 0 50 50 50 Yes Minimal RRV Simian 2 4 10 p.f.u. 1 0 0 50 50 50 No 1 4 None (4 10 f.f.u.) 2 0 1 50 1 10 1 10 5 10 Yes Minimal 3 0 0 50 50 50 No None 4 0 0 50 50 50 No None * Faecal infectious rotavirus shedding was assessed by FFA from 0 to 10 days p.i. and expressed in f.f.u. ml. When fluorescent foci in 1 10 dilutions could not be visualized by fluorescence microscopy, the samples were considered negative, and a value of 50 f.f.u. ml was assigned. Total infectious rotavirus shed in faecal samples was calculated by addition of titres of infectious virus in positive faecal samples and expressed in f.f.u. ml. Net yield of virus output was calculated by multiplying the total infectious rotavirus shed in faecal samples by the minimal volume (50 ml) of the entire (small, large and caecal) intestinal contents as determined in previous studies (Conner et al., 1988). As determined by ELISA at 28 days p.i. (see Table 2). Rabbits which seroconverted or shed virus antigen (as measured by ELISA) following ALA or RRV rotavirus inoculation. Data for rabbits inoculated with 2 4 10 p.f.u. of RRV were extracted from Ciarletet al. (1998a) for comparison.
Basis of RRV infection in rabbits (small, large and caecal) intestinal contents of age-matched ALA-infected rabbits is 50 ml (Conner et al., 1988). The net yield of ALA infectious virus in the rabbit intestine was at least 1 2 logs (1 10 1 2 10 f.f.u.) over the inoculum (Table 1). As with productively ALA-infected rabbits, RRV replication in productively RRV-infected rabbits exceeded input levels of RRV (Table 1). The net virus yield shed in the faecal samples alone of the rabbit inoculated with 2 4 10 p.f.u. (4 10 f.f.u.) of RRV was 2 9 10 f.f.u. ml, surpassing the virus inoculum titre by more than 1 log, and the net virus yield would be 1 5 10 f.f.u., almost 3 logs in titre over input virus (Table 1). Also, the net virus yield of the rabbit inoculated with 2 4 10 p.f.u. (4 10 f.f.u.) of RRV was estimated at 3 3 10 f.f.u. (1 log above virus input) (Table 1). Hence, productive infection of rabbits by both ALA and RRV corresponds to net yield of virus titre over inoculum titre, indicating that replication of both viruses occurred. Limited replication corresponds to lower infectious virus titres compared to those obtained in a productive infection. In fact, net virus yield of rabbits that exhibited limited RRV replication either barely rose or did not rise above virus input titres, suggesting limited, and not productive, replication (Table 1). Transmission of RRV to control rabbits To determine if horizontal transmission of RRV occurred, we monitored rotavirus antigen shedding in mock-inoculated rabbits housed in the same room and cage rack, but in different cages from RRV-inoculated rabbits. No rotavirus antigen shedding was detected in rabbits inoculated with PBS (n 3) from 0 to 10 days p.i. (data not shown). In studies on horizontal transmission with homologous ALA lapine rotavirus, onset of virus antigen shedding usually occurs by 2 3 days p.i., but ALA virus antigen shedding due to transmission to control rabbits, housed in separate cages in the same room, is not detected until 6 8 days p.i. (Ciarlet et al., 1998c; Conner et al., 1988). RRV-inoculated rabbits had a 1- to 3-day delay in onset of virus antigen shedding (4 7 days p.i.) compared to those for ALA-inoculated rabbits (3 and 4 days p.i.) (Fig. 1 E, F); therefore, it was possible that RRV transmission to PBS controls was not detected by 10 days p.i. because of the delay in RRV antigen shedding of RRV-inoculated rabbits. So, at 12 days p.i. additional faecal samples from the mock-inoculated rabbits were collected and tested for virus antigen. No faecal samples were collected beyond 12 days p.i. One of three control rabbits excreted high levels (A 2 0) of virus antigen at 12 days p.i. (data not shown); this rabbit was housed individually, but directly below a rabbit inoculated with 2 4 10 p.f.u. of RRV that was productively infected with RRV. The two PBS-inoculated rabbits that did not shed virus antigen were individually housed below a rabbit inoculated with 2 4 10 or 2 4 10 p.f.u. of RRV that exhibited limited infection or no infection with RRV, respectively. Therefore, horizontal transmission of heterologous RRV did occur in rabbits, although it was not observed as readily as with Fig. 3. Electrophoresis of genome RNA electropherotype of lapine ALA rotavirus and simian RRV inocula and output virus in faecal samples from rabbits inoculated with RRV or PBS. Lanes: 1, lapine ALA rotavirus; 2, simian RRV; 3, empty; 4, faecal sample collected at 6 days p.i. from a productively infected rabbit inoculated with 2 4 10 7 p.f.u. of RRV; 5, faecal sample collected at 7 days p.i. from a productively infected rabbit inoculated with 2 4 10 5 p.f.u. of RRV; 6, faecal sample collected 7 days p.i. from a minimally infected rabbit inoculated with 2 4 10 3 p.f.u. of RRV; 7, faecal sample collected at 12 days p.i. from the rabbit inoculated with PBS that became infected with RRV by horizontal transmission. Samples were subjected to electrophoresis in a 7% polyacrylamide gel and genome segments were visualized by silver staining. homologous ALA lapine rotavirus, which spreads throughout the animal room regardless of the distance (Ciarlet et al., 1998a, c; Conner et al., 1988). To confirm that RRV was transmitted horizontally to the PBS-inoculated rabbit, we analysed the RNA electropherotype of the virus recovered in selected faecal samples of rabbits inoculated with ALA, RRV or PBS by PAGE. No experiments with lapine ALA rotavirus were concurrently performed with those of RRV or PBS. In faecal samples from separate experiments in which rabbits were inoculated with ALA or RRV, the recovered virus was always identical to the inoculum virus (Fig. 3 and data not shown). The RNA electropherotype of ALA is easily distinguishable from that of RRV since genome segment 11 of ALA is re-arranged (Fig. 3). The RNA electropherotype observed in the faecal sample from the PBScontrol rabbit that shed virus at 12 days p.i. was identical to that of RRV (Fig. 3). Thus, RRV unequivocally spread to one of the control rabbits, confirming horizontal transmission of RRV in rabbits. Immune response of rabbits inoculated with different doses of RRV Infection of rabbits with different doses of the heterologous RRV strain was also monitored by the ability of the infecting virus to induce a primary serological and intestinal antibody response. All preinoculation serum and faecal samples were rotavirus-antibody-negative at a dilution of 1 :50 and 1 :5, respectively (data not shown). All RRV-infected rabbits that exhibited either productive or limited replication based on faecal virus antigen shedding (inoculated with 2 4 10 or 2 4 10 p.f.u.) developed rotavirus-specific serum and intestinal antibody responses (Table 2). The one rabbit inoculated BCED
BCEE Table 2. Rotavirus antibody responses and protection of rabbits inoculated orally with different doses of heterologous (non-lapine) RRV (P5B[3], G3) strain and challenged with 10 3 ID 50 of homologous (lapine) ALA (P[14], G3) rotavirus strain Total (IgM, IgG, IgA) antibody GMT (range) measured by ELISA at indicated times Serum neutralizing antibody GMT (range) against indicated virus following primary and challenge inoculation Mean % No. of No. of protection Species Dose of responders*/ 28 days p.i. 28 days p.c. 28 days p.i. 28 days p.c. protected / from of inoculum inoculated challenged challenge Inoculum origin (p.f.u.) rabbits Serum Intestinal Serum Intestinal RRV ALA RRV ALA rabbits (range) M. Ciarlet, M. K. Estes and M. E. Conner PBS 1 3 504 8 102400 320 159 126 1600 3200 1 3 33 ( 50 51200) ( 5 80) (25600 (40 1280) ( 100 ( 100 (1600) (800 6400) (0 100) 204800) 1600) 800) ALA Lapine 1 10 2 2 102400 640 ND ND 600 2400 ND ND ND ND (51200 204800) (640) (400 800) (1600 3200) ALA Lapine 3 5 10 2 2 144815 160 102400 160 1131 4526 800 3200 2 2 100 (102400 204800) (160) (102400) (160) (800 1600) (3200 6400) (800) (3200) (100) ALA Lapine 1 10 3 3 12800 202 36200 101 ND 318 ND 800 3 3 100 (1600 51200) (80 640) (25600 51200) (40 320) ( 100 800) (200 3200) (100) RRV Simian 2 4 10 2 2 204800 640 102400 160 3200 1131 3200 1131 2 2 100 (102400 409600) (640) (51200 204800) (80 320) (1600 6400) (800 1600) (1600 6400) (800 1600) (100) RRV simian 2 4 10 4 4 30444 80 30444 67 1600 673 2263 1600 4 4 100 (12800 51200) (40 160) (12800 51200) (40 320) (800 3200) (400 800) (1600 3200) (800 1600) (100) RRV Simian 2 4 10 4 4 43054 48 72408 48 1903 951 2691 1345 4 4 100 (25600 51200) (20 80) (51200 102400) (40 80) (800 3200) (800 1600) (1600 6400) (800 1600) (100) RRV Simian 2 4 10 1 4 100 2 5 172216 269 50 50 1131 2691 0 4 3 ( 50 800) (102400 819200) (160 320) (400 3200) (1600 6400) (0 12) * Rabbits which seroconverted or shed virus antigen (as measured by ELISA) following primary rotavirus inoculation. For titres 50 and 5, 25 and 2 5, respectively, were used to calculate the genometric mean titre (GMT). A GMT of 25 or 2 5, for serum or intestinal antibodies, respectively, was considered negative. Serum neutralizing antibody was measured by FFNA. For titres 100, 50 was used to calculate the GMT. A GMT of 50 was considered negative. Rabbits which exhibited 100% reduction in virus antigen shedding following challenge by comparing the area under the curve for each individual animal to the mean of the areas under the curves of negative control animals. One rabbit was infected through horizontal spread of RRV. Data for rabbits inoculated with 1 10 p.f.u. of ALA were originally reported elsewhere (Conner et al., 1997), but are shown here for comparison. Data for rabbits inoculated with 2 4 10 p.f.u. of RRV were originally reported elsewhere (Ciarlet et al., 1998a), but are shown here for comparison. ND, Not done.
Basis of RRV infection in rabbits Fig. 4. Comparison of the deduced amino acid sequence of the NSP1s of lapine rotavirus strains ALA, C-11 and BAP-2 with simian rotaviruses RRV and SA11. Dashes indicate amino acid residues identical to those of ALA. The number of amino acids is based on ALA NSP1 and gaps (indicated by dots) have been inserted in ALA, C-11, BAP-2 and SA11 NSP1s to obtain the best alignment. The NSP1 sequences of RRV (U08433; Dunn et al., 1994 a) and SA11 (X14914; Hua et al., 1993) were reported previously. with 2 4 10 p.f.u. of RRV that had no detectable faecal virus antigen shedding (Fig. 1 C) also developed both serological and intestinal antibodies to rotavirus, suggesting that low level replication occurred without detectable virus shedding. The magnitude of the serological or intestinal immune responses of rabbits inoculated with 2 4 10 or 2 4 10 p.f.u. of heterologous RRV was equivalent (P 0 278, Mann Whitney U), but was significantly lower (P 0 017, Mann Whitney U) than that obtained with 2 4 10 p.f.u. of heterologous RRV (Ciarlet et al., 1998 a) or with homologous ALA infection. Virus antigen shedding detected by ELISA was not observed in any of the four rabbits inoculated with 2 4 10 p.f.u. of RRV (Fig. 1D). However, one of these four rabbits shed low titres of infectious virus for 1 day (Fig. 2 D) and developed low-titred serological, but not intestinal, antibodies to rotavirus (Table 2). The occurrence of minimal replication that results in a low serological, but not intestinal, antibody response occurs also with other heterologous viruses in rabbits (Ciarlet et al., 1998 a). No rotavirus-specific antibodies were detected in rabbits inoculated with PBS (n 3), except in the one that was found to be shedding virus antigen at 12 days p.i. as a result of horizontal spread of RRV (Table 2). RRV did not spread to the other two control animals inoculated with PBS, based on both lack of virus antigen shedding and seroconversion. Neutralizing antibody titres were determined using sera from all inoculated rabbits against both ALA and RRV. Prior to inoculation, none of the sera had neutralizing antibodies to either virus at a dilution of 1 :50 (data not shown). All rabbits infected with ALA or RRV had equivalent (P 0 291, Mann Whitney U) serum neutralizing antibodies (1 : 800 1 : 6400) to the corresponding immunizing rotavirus strain (Table 1). However, although both RRV and ALA rotaviruses share the same VP7 type (G3), serum neutralizing antibody titres of RRV-infected rabbits were two- to fourfold lower to ALA rotavirus than to RRV (Table 2). The opposite was observed in ALA-infected rabbits (Table 2), suggesting that neutralizing antibodies were preferentially made to the immunizing strain. Similar results have been noted in rabbits BCEF
M. Ciarlet, M. K. Estes and M. E. Conner Table 3. Comparison of NSP1 amino acid identities (%) of lapine and representative group A rotavirus strains of different species of origin Virus strain Species of origin ALA % Identity with C-11 BAP-2 ALA Lapine 100 0 C-11 Lapine 92 9 100 0 BAP-2 Lapine 91 5 91 5 100 0 RRV Simian 87 3 85 3 87 5 SA11 Simian 56 1 56 7 57 3 FRV64 Feline 85 0 83 0 84 8 Cat2 Feline 37 4 38 0 37 4 L338 Equine 71 4 71 6 71 2 H-2 Equine 71 4 71 7 71 6 FI-23 Equine 71 6 72 2 72 1 H-1 Equine 37 2 35 9 37 0 OSU Porcine 37 4 35 8 37 1 YM Porcine 37 2 36 0 37 0 Gottfried Porcine 37 5 36 0 37 0 Wa Human 40 0 39 2 40 0 K8 Human 39 2 37 9 39 0 DS-1 Human 36 5 36 1 36 7 69M Human 36 1 36 2 36 3 St-3 Human 37 5 36 9 38 1 EHP Murine 49 6 49 6 49 4 EW Murine 50 4 50 4 51 1 UK Bovine 36 2 36 6 36 6 B223 Bovine 35 4 36 6 35 6 infected orally with different heterologous rotavirus strains (Ciarlet et al., 1998 a) or immunized parenterally with virus-like particles (Ciarlet et al., 1998b; Crawford et al., 1999). Estimation of ID 50 of RRV in rabbits Based on virus antigen shedding or induction of an immune response, all eight rabbits inoculated with 2 4 10 (n 4) or 2 4 10 (n 4) p.f.u. of RRV were infected with RRV, whereas only one of three rabbits inoculated with 2 4 10 p.f.u. of RRV, became infected following inoculation. Based on these data, we estimated using the Karber equation that the ID of RRV in rabbits is 2 4 10 p.f.u. A more precise estimate of the ID of RRV in rabbits might be obtained if additional rabbits were inoculated with doses of 10 and 10 p.f.u. of RRV. Protection induced by RRV against virus antigen shedding from challenge Following challenge with the homologous ALA rotavirus strain, all the rabbits inoculated with ALA or with 10 and 10 p.f.u. of RRV were completely protected against infection based on lack of virus antigen shedding (data not shown) and lack of anamnestic responses ( fourfold increases in titre) of either total or neutralizing antibody (Table 2). The PBSinoculated rabbit that was infected with RRV as a result of horizontal transmission of RRV was also 100% protected against challenge (Table 2). The other two PBS-inoculated rabbits were not protected from challenge and each shed virus antigen for 5 days (data not shown) and consequently developed serum and intestinal rotavirus-specific antibody responses (Table 2). None of the rabbits (n 4) inoculated with 10 p.f.u. of RRV were protected from challenge; all rabbits shed virus antigen, with a mean duration of 4 75 days (data not shown), and developed serum and intestinal rotavirus-specific antibody responses following challenge (Table 2). As with previous results (Ciarlet et al., 1998a), protection from ALA challenge correlated with the presence of a local immune response (P 0 001, r 0 765, Pearson s correlation coefficient) or a serological immune response (P 0 001, r 0 873, Pearson s correlation coefficient) (Table 2). However, a low level systemic immune response in the absence of a local response was not enough to confer protection from challenge. Sequencing of NSP1 from lapine rotaviruses The fundamental structure of the NSP1 genes from the three lapine strains sequenced was similar to those of other BCEG
Basis of RRV infection in rabbits rotavirus strains (Dunn et al., 1994a; Fujiwara & Nakagomi, 1997; Kojima et al., 1996; Xu et al., 1994). The gene of all three lapine NSP1s consisted of 1597 bp with one open reading frame beginning at nucleotide 32 and a single TAA stop codon at nucleotide 1515. The predicted size of the lapine NSP1 proteins was 492 amino acids, two amino acids less than the NSP1 of RRV (1601 bp, 494 aa) (Dunn et al., 1994a). As with most rotavirus strains, the eight prolines (potentially important for conformation) and the cysteine-rich region (putative RNA binding domain) were conserved in all three rabbit strains. A comparison of the complete deduced amino acid sequence for the genes encoding the NSP1 of the lapine strains with that of simian rotavirus strains is shown in Fig. 4. A high degree of amino acid identity (91 5 95 1%) was found among the NSP1 proteins of all three lapine rotavirus strains. The deduced amino acid sequences of the NSP1 genes of the lapine strains were also compared with the NSP1 sequences of representative rotavirus strains (Table 3). The highest degree of identity was found with the simian RRV strain (85 3 87 5%) and with the feline FRV64 strain (83 85%). With the remaining rotavirus strains examined, the amino acid identity with any lapine NSP1 ranged from 35 4% (bovine strain B223) to 71 6% (equine strains H-2, L338 and FI-23). Discussion Among 27 different heterologous (non-lapine) or reassortant rotavirus strains, only simian RRV replicated with any efficiency in rabbits (Ciarlet et al., 1998a; unpublished results). Although a high dose (10 p.f.u.) of RRV resulted in a productive infection in rabbits (Ciarlet et al., 1998 a), the replication capabilities of RRV doses 10 p.f.u. was unknown. In the current study, we confirm that (i) RRV can replicate productively in rabbits when given at lower doses, and (ii) RRV can be horizontally transmitted between rabbits. However, the efficiencies of replication and horizontal transmission of RRV in rabbits are reduced compared to those of the lapine ALA rotavirus. First, the estimated ID of RRV (2 4 10 p.f.u.) in rabbits is approximately 100-fold greater than that of the homologous lapine ALA strain (1 7 10 p.f.u.) (Conner et al., 1997). Second, there was a 1- to 2-day delay in onset and peak of RRV virus antigen shedding compared to those for ALA-inoculated rabbits. Third, RRV does not spread as efficiently to control rabbits (one out of three) as homologous lapine rotavirus strains ALA, C-11, R-2 and BAP-2 (Ciarlet et al., 1998a, c; Conner et al., 1988). Finally, the immune response following heterologous RRV infection of rabbits is dose-dependent, unlike that of homologous ALA infection of rabbits, which is not dose-dependent (Conner et al., 1997). However, at RRV inoculation doses of 10 p.f.u., RRV induces a protective (100%) immune response against infection from homologous ALA lapine rotavirus challenge. When compared to 27 other heterologous rotavirus strains tested, the replication kinetics of RRV in rabbits are unique. Other heterologous rotavirus strains exhibited only limited replication in rabbits even at inoculum doses of 10 p.f.u. (Ciarlet et al., 1998 a). Additionally, RRV is the only heterologous rotavirus strain tested that is capable of horizontal transmission in rabbits (Ciarlet et al., 1998a; Conner et al., 1988). Together with our previous report (Ciarlet et al., 1998 a), these data provide the first indication that a heterologous virus can replicate productively and spread horizontally in rabbits. Studies by others have indicated that gene 5 sequences generally segregate according to the species of origin and represent the best association with species of origin and can best determine interspecies relatedness within the overall genomic RNA constellation (or genogroup) (Dunn et al., 1994a; Fujiwara & Nakagomi, 1997; Hua et al., 1994; Kojima et al., 1996; Nakagomi & Kaga, 1995; Palombo & Bishop, 1994). NSP1 has been implicated in rotavirus host range restriction in some studies (Broome et al., 1993; Dunn et al., 1994 a; Fujiwara & Nakagomi, 1997; Gombold & Ramig, 1986; Kojima et al., 1996), but not in others (Bridger et al., 1998; Ciarlet et al., 1998a; Gombold & Ramig, 1986; Hoshino et al., 1995). The role of NSP1 in rotavirus host range restriction may depend on the genetic background of a particular virus strain or the constellation of rotavirus genes, or alternatively, the host (Burke & Desselberger, 1996; Ciarlet et al., 1998a; Hoshino et al., 1995). The genes encoding the NSP1s of three lapine rotavirus strains (ALA, C-11 and BAP- 2) were sequenced and our results clearly indicate that there is a close relationship between the NSP1 of the lapine strains and the simian RRV strain (85 88%). The amino acid identity of NSP1 of RRV and the lapine rotavirus strains is the highest reported yet for this protein from viruses from different species. Therefore, rabbits are possibly more susceptible to infection by RRV than by other heterologous rotavirus strains because of the similarity of the NSP1 sequences between RRV and the lapine rotaviruses. A close identity between the NSP1 of human and porcine strains has also been noted, albeit slightly lower (range 69 86%) (Dunn et al., 1994a; Palombo & Bishop, 1994; Xu et al., 1994). Recently, dot blot hybridization assays revealed that the NSP1s of canine (RS15, K9 and CU-1) and feline (FRV64 and Cat97, but not Cat2) rotaviruses share a high level of similarity (Fujiwara & Nakagomi, 1997). In fact, the amino acid sequence of the NSP1 of FRV64 strain is very similar to that of RRV (85% amino acid identity) (Fujiwara & Nakagomi, 1997). Therefore, we predict that FRV64 might be capable of replication in rabbits, but FRV64 was not available for testing. The fact that the lapine strains and RRV share an unusually high amino acid identity (85 88%) of gene 5, does not directly implicate gene 5 in host range restriction of rotaviruses. We showed previously that a reassortant rotavirus with ALA gene 5 and all other genes from simian SA11 replicated with less efficiency than either parental rotavirus strain (Ciarlet et al., 1998a). In mice, gene 5 is implicated in rotavirus host range BCEH
M. Ciarlet, M. K. Estes and M. E. Conner restriction (Broome et al., 1993) and analysis of NSP1 amino acid sequences revealed that murine rotavirus strains (EW and EHP) are more closely related to RRV (50 51%) than to other rotavirus strains (37 40%) that do not replicate in mice (Broome et al., 1993; Dunn et al., 1994a; Kojima et al., 1996). Since RRV and the murine strains and RRV and the lapine strains also share similar VP4s (77 82% and 69 70%, respectively) and VP7s (88 90% and 96 97%, respectively), which are both gene products associated with rotavirus host range restriction (Ciarlet et al., 1998a; Dunn et al., 1994b; Hoshino et al., 1995), the overall similarity of their genes encoding NSP1, VP4 and VP7 may explain the replication competence of RRV in mice and in rabbits (Ciarlet et al., 1998 a). If our hypothesis is correct, then RRV may also replicate efficiently in cats and dogs due to the similarities of the NSP1s between RRV and canine and feline rotavirus strains, which also share the same VP7 serotype (G3), VP4 type (P[3]) as well as similar NSP4 type (Fujiwara & Nakagomi, 1997; Horie et al., 1997; Kirkwood & Palombo, 1997; Ciarlet et al., 2000). Other genes not yet analysed for the lapine strains may also share high similarities with RRV, which might contribute to the replication efficiency of RRV in rabbits. Sequence comparison between RRV and lapine rotaviruses of the other rotavirus genes whose products have been implicated in rotavirus host range restriction (VP3, NSP2 and NSP4) cannot be performed because of lack of sequence availability; only the sequences of NSP2 and NSP4 of RRV have been reported. However, recent sequencing data of gene 10 (encoding NSP4) of lapine rotavirus strains indicate that the lapine NSP4s are closely related to each other, and they are divergent from that of RRV (Ciarlet et al., 2000). The low amino acid identity (56 57%) shared between the NSP1 of SA11 rotavirus and the NSP1s of lapine rotavirus strains and RRV (Dunn et al., 1994a) may explain why SA11 rotavirus, a strain that shares similar VP7 (95 97%) and VP4 (68 70 %) with the lapine rotavirus strains (Ciarlet et al., 1997), exhibits limited replication in rabbits (Ciarlet et al., 1998a). Thus, whether susceptibility of rabbits to infection by RRV is due to NSP1 remains to be determined. Reassortants containing gene 5 from RRV and the other genes from a heterologous strain with limited replication in rabbits, and vice versa, could provide evidence for or against NSP1 as an important determinant for rotavirus host range in rabbits. It will be interesting to determine what unusual properties RRV possesses that allow it to replicate in a broad range of species (monkey, mouse, rabbit and human) and whether RRV will replicate efficiently in other species. We are extremely grateful to Reginald Semiens for excellent work in the maintenance of rabbits and Sharon Krater for expert tissue culture assistance. We express our gratitude to R. Frank Ramig for critically revising the manuscript. This work was supported by Public Health Service grant AI 24998 from National Institutes of Allergy and Infectious Diseases, by World Health Organization grant MIMV 2718130, and by National Institute of Diabetes and Digestive and Kidney Diseases grant DK30144. References Bridger, J. C., Woode, G. N., Jones, J., Flewett, T. H., Bryden, A. & Davies, H. (1975). Transmission of human rotaviruses to gnotobiotic piglets. Journal of Medical Microbiology 8, 565 569. Bridger, J. C., Dhaliwal, W., Adamson, M. J. & Howard, C. R. (1998). Determinants of rotavirus host range restriction a heterologous bovine NSP1 gene does not affect replication kinetics in the pig. Virology 245, 47 52. Broome, R. L., Vo, P. T., Ward, R. L., Clark, H. F. & Greenberg, H. B. (1993). Murine rotavirus genes encoding outer capsid proteins VP4 and VP7 are not major determinants of host range restriction and virulence. Journal of Virology 67, 2448 2455. Burke, B. & Desselberger, U. (1996). Rotavirus pathogenicity. Virology 218, 299 305. Burns, J. W., Krishnaney, A., Vo, P. T., Rouse, R., Anderson, L. & Greenberg, H. B. (1995). Analyses of homologous rotavirus infection in the mouse model. Virology 207, 143 153. Ciarlet, M. & Liprandi, F. (1994). Serological and genomic characterization of two porcine rotaviruses with serotype G1 specificity. Journal of Clinical Microbiology 32, 269 272. Ciarlet, M., Hidalgo, M., Gorziglia, M. & Liprandi, F. (1994). Characterization of neutralization epitopes on the VP7 surface protein of serotype G11 porcine rotaviruses. Journal of General Virology 75, 1867 1873. Ciarlet, M., Estes, M. K. & Conner, M. E. (1997). Comparative amino acid sequence analysis of the outer capsid protein VP4 from four lapine rotavirus strains reveals identity with genotype P[14] human rotavirus. Archives of Virology 142, 1059 1069. Ciarlet, M., Estes, M. K., Barone, C., Ramig, R. F. & Conner, M. E. (1998a). Analysis of host range restriction determinants in the rabbit model: comparison of homologous and heterologous rotavirus infections. Journal of Virology 72, 2341 2351. Ciarlet, M., Crawford, S. E., Barone, C., Bertolotti-Ciarlet, A., Estes, M. K. & Conner, M. E. (1998b). Subunit rotavirus vaccine administered parenterally to rabbits induces active protective immunity. Journal of Virology 72, 9233 9246. Ciarlet, M., Gilger, M. A., Barone, C., McArthur, M., Estes, M. K. & Conner, M. E. (1998c). Rotavirus disease, but not infection and development of intestinal histopathological lesions, is age restricted in rabbits. Virology 251, 343 360. Ciarlet, M., Liprandi, F., Conner, M. E. & Estes, M. K. (2000). Species specificity and interspecies relatedness in NSP4 genetic groups by comparative NSP4 sequence analyses of animal rotaviruses. Archives of Virology 145, 371 383. Conner, M. E. & Ramig, R. F. (1996). Enteric diseases. In Viral Pathogenesis, chapter 30, pp. 713 743. Edited by N. Nathanson, R. Ahmed, F. Gonza lez-scarano, D. E. Griffin, K. V. Homes, F. A. Murphy & H. L. Robinson. New York: Lippincott-Raven. Conner, M. E., Estes, M. K. & Graham, D. Y. (1988). Rabbit model of rotavirus infection. Journal of Virology 62, 1625 1633. Conner, M. E., Gilger, M. A., Estes, M. K. & Graham, D. Y. (1991). Serologic and mucosal immune response to rotavirus infection in the rabbit model. Journal of Virology 65, 2562 2571. BCEI