260, 316–328 (1999) Article ID viro.1999.9832, available online at http://www.idealibrary.com on

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Extensive Sequence Divergence and Phylogenetic Relationships between the Fusogenic and Nonfusogenic : A Proposal

Roy Duncan1

Department of and Immunology, Faculty of Medicine, Dalhousie University, Halifax, Nova Scotia, Canada B3H 4H7 Recieved January 21, 1999; accepted June 3, 1999

The orthoreoviruses can be divided into subgroups based on either their restricted range or the unusual ability of certain members of this group of nonenveloped to induce –cell fusion from within. Phylogenetic relationships cannot be inferred based on these biological properties because fusogenic reoviruses are present in both the avian and mammalian subgroups. To address this issue, the complete nucleotide sequences of the three S- segments encoding the major ␴-class core, outer , and nonstructural proteins of four fusogenic reoviruses were determined and used to establish the phylogeny of the orthoreoviruses. The viruses analysed included two strains of avian reovirus and the only known fusogenic mammalian reoviruses, Nelson Bay and baboon reovirus. Comparative sequence analysis of these fusogenic reoviruses and the prototypical nonfusogenic mammalian reoviruses indicated a highly diverged with both conserved and unique sequence-predicted structural motifs in the major ␴-class proteins. Phylogenetic analysis provided the basis for the first taxonomic subdivision of the orthoreoviruses into species classes based on inferred evolutionary relationships. It is proposed that the orthoreoviruses consist of at least four species that separate into three clades. The nonfusogenic mammalian reovirus species represent a single clade, and the fusogenic reoviruses separate into two distinct clades. The first clade of fusogenic reoviruses contains the avian reovirus- and Nelson Bay virus-, with the second clade being occupied by the single baboon reovirus isolate that represents a fourth species. © 1999 Academic Press

INTRODUCTION lian reoviruses (MRVs) (Nibert et al., 1996; Robertson and Wilcox, 1986). There are, however, two atypical MRV iso- The genus Orthoreovirus is one of nine accepted genera lates, Nelson Bay virus (NBV) and baboon reovirus (BRV), in the (Murphy et al., 1995). Like most and two isolates from snakes that share the syncytium- double-stranded RNA (dsRNA) viruses, members of this inducing properties of the avian subgroup (Duncan et al., large and diverse family are all icosahedral, nonenveloped 1995, and references therein). The evolutionary relation- viruses (reviewed in Nibert et al., 1996; Pereira, 1991). There ships and the extent of divergence between the fusogenic is limited sequence conservation between members of different genera in the Reoviridae, which also display dis- and nonfusogenic reoviruses, or between the mammalian tinct capsid morphologies, host ranges, biological proper- and avian viruses, have not been established. ties, replication strategies, electropherotypes, and protein Phylogenetic analysis of the fusogenic and nonfuso- profiles (Murphy et al., 1995; Pereira, 1991). Members within genic reoviruses is an essential step in clarifying the the genus Orthoreovirus share similar capsid structures, species within the genus Orthoreovirus. The Inter- genome segment profiles, and protein compositions (Mur- national Committee on of Viruses (ICTV) re- phy et al., 1995; Nibert et al., 1996). Extensive sequence cently agreed to implement the species class as the analysis has clearly established the evolutionary relation- lowest taxonomic level (Mayo, 1996; Pringle, 1991; Van ships of the three serotypes of mammalian orthoreoviruses Regenmortel et al., 1997). Virus species are to be defined and revealed a fairly homogeneous genus (Weiner and using polythetic criteria to accommodate the inherent Joklik, 1987, 1989). variability of viruses (Mayo, 1996), and it is generally The orthoreoviruses are not, however, entirely homoge- accepted that the hierarchal taxonomic scheme should neous and can be subdivided based on distinct biological reflect both genealogical and biological similarities (Cal- properties, most notably, their host range and the unusual isher et al., 1995; Mayo and Pringle, 1997; Van Regen- ability of certain members to induce syncytium formation mortel et al., 1997). The particular criteria to be used in (Fig. 1). Cell–cell fusion, a property typical of avian reovi- determining species relationships will vary between the ruses (ARVs), is not associated with the prototype mamma- different families and genera and include such factors as sequence similarity, host range, tissue tropism, pathoge- nicity, physicochemical properties, and antigenic related- 1 To whom reprint requests should be addressed at Tupper Medical ness (Van Regenmortel, 1990; Van Regenmortel et al., Building. Fax: 902-494-5125. E-mail: [email protected]. 1997).

0042-6822/99 $30.00 Copyright © 1999 by Academic Press 316 All rights of reproduction in any form reserved. ORTHOREOVIRUS PHYLOGENY 317

FIG. 1. Members of the genus Orthoreovirus and their genome segment profiles. (A) The various members of the genus Orthoreovirus are indicated, along with their abbreviated . The isolates obtained from mammalian sources represent three serotypes of MRVs and the single fusogenic isolates designated NBV and BRV. The ARVs include numerous isolates obtained from commercial poultry flocks that represent at least five different serotypes. Two additional isolates are also listed that were obtained from snakes. The bottom of the figure indicates the subdivision of the genus into the fusogenic and nonfusogenic subgroups. (B) The purified dsRNA genome segments of ARV-138 (38), ARV-176 (76), NBV (N), and BRV (B) were resolved by SDS–PAGE and visualized by silver staining. The locations of the large (L), medium (M), and small (S) size class genome segments are indicated along with the numbering scheme of the S-class genome segments. The S3 and S4 genome segments of ARV-138 comigrated on this particular gel.

Antigenic analysis of the fusogenic and nonfusogenic quences of ARV or MRV genome segments. The se- reoviruses indicated that ARV, MRV, NBV, and BRV pos- quences of the S3 and S4 genome segments of ARV- sess limited epitope conservation and may represent S1133 and the S3 genome segment of ARV-1733 have distinct orthoreovirus genogroups (Duncan et al., 1995). been previously reported (Yin et al., 1997; Chiu and Lee, The paucity of sequence information for the fusogenic 1997; Vakharia, 1997). A single-nucleotide difference ex- orthoreoviruses has prevented a more comprehensive ists between the ARV-176 S3 genome segment se- sequence-based analysis of the extent of divergence quenced in the present study and the previously reported and phylogenetic relationships of the fusogenic and non- ARV-1733 sequence. A maximum of five nucleotide dis- fusogenic orthoreoviruses. The current study, based on a crepancies exist between the S3 and S4 genome seg- multigenic sequence-based approach, indicated a re- ment sequences of ARV-176 and the reported ARV-S1133 markable degree of divergence between the fusogenic sequences (data not shown). These observations re- and nonfusogenic reoviruses and, most surprisingly, vealed the high fidelity of the reported ARV sequences even within the fusogenic mammalian reoviruses. The and suggested that ARV strains 176, S113, and 1733 are extensive sequence divergence facilitated a reevaluation most likely different isolates of the same strain. of sequence-predicted structural motifs in the major The clearly inferred homology with the nonfusogenic ␴-class proteins, and maximum parsimony analysis es- mammalian reoviruses established the genome segment tablished the phylogeny of these diverse reoviruses con- coding assignments of the fusogenic orthoreoviruses, tributing to the definition of four species classes in the which are presented in Table 1, along with a summary of genus Orthoreovirus. the predicted coding and noncoding regions of the genome segments. The ARV major ␴-class core, outer capsid, and RESULTS nonstructural proteins of ARV-176 and ARV-138 are en- coded by the S2, S3, and S4 genome segments, respec- Sequence analysis of the S-class genome segments tively, as previously demonstrated by in vitro translation of ␴ encoding the major -class core, outer capsid, and the ARV-S1133 genome segments (Schnitzer, 1985; Varela nonstructural proteins of the fusogenic reoviruses and Benavente, 1995). A similar situation exists for MRV The S-class genome segments of NBV, BRV, and two and NBV except for the reversed of the coding as- strains of ARV that were predicted to encode the major signments of the S3 and S4 genome segments (Table 1). ␴-class core, outer capsid, and nonstructural proteins of Sequence comparisons (see Figs. 4–6 and Table 2) clearly the viruses (Schnitzer, 1985; Varela and Benavente, 1994) revealed that the BRV S1, S2, and S3 genome segments were determined from cDNA clones. The cDNA se- encode the homologous ␴-class proteins. This surprising quences were used for BLAST searches of the updated result suggested that the smallest genome segment of BRV GenBank database, and in all cases, the only significant (S4 in Fig. 1B) represents a severely truncated equivalent of similarity detected was against previously published se- the S1 genome segment of the other viruses. 318 ROY DUNCAN

TABLE 1 S-Class Genome Segments Encoding the Major ␴-Class Proteins of the Orthoreoviruses

Major core protein Major outer capsid protein Nonstructural protein

Genome Base Genome Base Genome Base Virus segment pairs Codons segment pairs ORF Codons segment pairs ORF Codons

ARVa S2 1324 16–1263 416 S3 1202 31–1131 367 S4 1192 24–1124 367 NBV S2 1322 16–1263 416 S4 1185 31–1113 361 S3 1192 29–1129 367 BRV S1 1311 17–1255 413 S2 1253 31–1218 396 S3 1150 23–1081 353 MRVb S2 1331 19–1272 418 S4 1196 33–1127 365 S3 1198 28–1125 366

a Applies to ARV-176 and ARV-138. b Applies to MRV-La, MRV-Jo, and MRV-De.

The majority of each of the fusogenic reovirus genome between the three genome segments of a particular segments is occupied by a single extended open reading virus. Between the different viruses, the 5Ј-termini dis- frame that is flanked by relatively short 5Ј-terminal and play a limited sequence conservation with a conserved longer 3Ј-terminal noncoding regions (15–18 nucleotides at GCT triplet present in the ARV, NBV, and MRV genome the 5Ј-termini versus 35–73 nucleotides at the 3Ј-termini) as segments that is absent from BRV (Fig. 2). The 3Ј-termini previously noted for MRV (Nibert et al., 1996). The open exhibit more extensive sequence conservation with a reading frames of homologous genome segments encode conserved cDNA pentanucleotide sequence (TCATC-3Ј) proteins of similar size with the notable exception of the that is present in all of the fusogenic and nonfusogenic BRV S2 genome segment. The BRV S2 genome segment is reoviruses. This pentanucleotide sequence is distinct approximately 50–70 nucleotides larger than the corre- from the terminal sequences of other genera in the family sponding genome segments of the other viruses and con- (e.g., Bigot et al., 1995) and may represent a signature tains the shortest 3Ј-terminal nontranslated region, result- sequence for the orthoreoviruses. The conserved termi- ing in a major outer capsid protein that is approximately 30 nal sequences indicated that the fusogenic reovirus amino acids larger than the homologous protein encoded cDNA clones represented full-length replicas of the by the other viruses (Table 1). genomic RNA segments and, in conjunction with the An examination of the terminal cDNA sequences (Fig. sequence similarities and previous biophysical compar- 2) revealed that the 5Ј-terminal 6–8 nucleotides and the isons (Duncan et al., 1995), supported the classification 3Ј-terminal 5–9 nucleotides are completely conserved of all of these viruses as orthoreoviruses.

TABLE 2 Percent Identities of the Homologous Genome Segments of the Orthoreoviruses and Their Encoded Proteins

Percent amino acid identitiesa

Strain ARV-176 ARV-138 NBV BRV MRV-La MRV-Jo MRV-De

ARV-176 98, 94, 95 60, 49, 38 30, 32, 22 31, 26, 24 31, 25, 25 31, 25, 23 ARV-138 94, 83, 89 60, 50, 38 30, 31, 24 31, 25, 23 31, 23, 24 31, 25, 25 Percent NBV 60, 53, 49 60, 53, 49 32, 19, 21 30, 24, 20 31, 21, 18 30, 24, 20 Nucleotide BRV 40, 41, 33 40, 40, 32 43, 40, 36 28, 27, 16 29, 18, 18 28, 27, 16 Identities MRV-La 42, 37, 37 41, 36, 38 41, 36, 35 40, 35, 30 94, 90, 90 94, 97, 97 MRV-Jo 40, 38, 38 41, 38, 39 43, 38, 33 39, 34, 29 86, 73, 80 94, 90, 90 MRV-De 41, 39, 36 42, 37, 38 44, 35, 34 39, 34, 30 86, 87, 94 77, 74, 77

a The percent identities were determined by pairwise comparisons of the homologous genome segments of the various orthoreoviruses, and their encoded proteins, using sequences aligned by the GAP algorithm of the GCG software. The protein sequences were aligned first, and the aligned nucleotide sequences were adjusted to account for small gaps inserted in the amino acid alignments. Numbers in the top right of the table indicate amino acid identities, whereas the bottom left numbers indicate nucleotide identities. Under each pairwise comparison, the first number indicates the percent identity of the major core proteins, the second number indicates the percent identity of the nonstructural protein, and the third number indicates the percent identity of the major outer capsid protein. The same situation applies to the percent nucleotide identities of the homologous genome segments. ORTHOREOVIRUS PHYLOGENY 319

genome segments of ARV and MRV revealed extensive divergence with percent nucleotide identities of 36–42% and amino acid identities of the encoded ␴-class pro- teins that ranged from 23% to 31% (Table 2 and Fig. 3A). The degree of sequence divergence indicated that the numerous ARV and MRV isolates represent two distinct orthoreovirus genogroups. The percent amino acid identities between the homol- ogous proteins of the fusogenic mammalian reovirus, NBV, and either the MRV or ARV genogroup ranged between 20% to 30% and 38% to 60%, respectively (Table 2 and Fig. 3A). The extent of divergence between NBV and MRV clearly established these isolates as two dis- FIG. 2. Conserved terminal sequences of the orthoreovirus S-class tinct genogroups. The two- to three-fold increase in the Ј Ј genome segments. The 5 - and 3 -terminal plus-strand cDNA se- percent amino acid identity of the NBV–ARV versus the quences of the cloned S-class genome segments of the fusogenic orthoreoviruses sequenced in this report are compared with the pre- NBV–MRV pairwise comparisons is in accord with pre- viously published terminal sequences of the homologous genome vious evidence obtained from an antigenic comparison segments of mammalian reovirus, strain Dearing (see Materials and Methods for the accession numbers of the sequences). The identities of the genome segments and their encoded proteins are indicated on the left. Sequences that are completely conserved between the differ- ent genome segments of a particular virus are underlined. The ARV-176 sequence is shown, which was identical to the ARV-138 sequence over this region. The sequences of mammalian reovirus strains Jones and Lang show slight variations from the presented Dearing sequence but not in the conserved underlined regions.

Identification of four genogroups of orthoreoviruses The predicted amino acid sequences of homologous genome segments from the fusogenic reoviruses were aligned with published sequences of the prototype non- fusogenic mammalian reoviruses. The sequences of the major ␴-class core and nonstructural proteins were es- sentially contiguous, requiring only a few small gaps to maintain the alignments (see Figs. 4 and 5). The ␴-class major outer capsid proteins, which have diverged exten- sively and differ in size (see Table 1), required the inser- tion of numerous gaps and manual adjustment to im- prove the alignments (see Fig. 6). The aligned sequences were used in pairwise com- parisons, and the percent amino acid sequence identi- ties were determined. A similar approach was used to determine the percent nucleotide identities of the homol- ogous genome segments using cDNA sequences with inserted nucleotide triplets in regions corresponding to amino acid insertions used to maintain the protein align- ments. The results of the pairwise comparisons are sum- marized in Table 2, and for ease in comparison, the percent amino acid identities are presented graphically FIG. 3. Amino acid identities in the three homologous major ␴-class in Fig. 3. proteins of the fusogenic and nonfusogenic orthoreoviruses. The The percent nucleotide identity of the homologous cDNA-predicted sequences of the ␴-class major core (black), nonstruc- genome segments of the two ARV isolates exceeded tural (lined), and major outer capsid (gray) proteins of the fusogenic and 83%, resulting in amino acid identities of the encoded nonfusogenic orthoreoviruses were aligned in pairwise comparisons using the GAP algorithm of the GCG software. The percent amino acid proteins in the range of 94–98%, similar to the situation identities were determined from the aligned sequences. (A) Pairwise between the various MRV sequences (Table 2). In con- comparisons of MRV strain Dearing, ARV-176, and NBV. (B) Pairwise trast, sequence comparisons between the homologous comparisons of BRV with MRV strain Dearing, ARV-176, and NBV. 320 ROY DUNCAN

of the viruses (Duncan et al., 1995), and in conjunction whereas the C-terminal domain may participate in inter- with the more extensive conservation of the ARV and actions with the internal genomic RNA (Dermody et al., NBV terminal genome segment sequences (Fig. 2), sug- 1991). A region of similarity between MRV ␴2 and Esch- gests that ARV and NBV share a more recent evolution- erichia coli DNA-dependent RNA polymerase was also ary ancestor. The ARV–NBV similarity is, however, con- previously noted in the N-terminal helical domain (Der- siderably lower than the identities observed between mody et al., 1991). This region (indicated in Fig. 4) con- members within the ARV or MRV genogroups, suggest- tains a preponderance of acid and amide residues that ing that NBV represents a third orthoreovirus genogroup. can be modeled as an amphipathic ␣ helix and was Most surprising was the demonstration that the sec- suggested as a possible site for RNA binding. The same ond fusogenic mammalian reovirus, BRV, represents a region in the ␴-class major core proteins of ARV and NBV separate genogroup that is apparently no more closely does not contain conserved primary sequence identity related to the fusogenic NBV mammalian reovirus than it with the E. coli RNA polymerase, but it can be modeled is to either the MRV or ARV genogroup. Percent amino as an amphipathic helix, and it is somewhat enriched for acid identities between the BRV proteins and the homol- acid and amide residues (Fig. 4). However, this motif is ogous proteins of the other three orthoreovirus geno- not conserved in either primary sequence or predicted groups ranged from 16% to 32% depending on the par- secondary structure in the sequence of the BRV ␴-class ticular pairwise comparison (Table 2 and Fig. 3B). There major core protein, suggesting either that the BRV ␴-core was no significant increase in the percent amino acid protein may not bind dsRNA or that this motif is not identities between NBV and BRV versus comparisons of directly required for RNA binding. BRV with ARV or MRV, providing the first direct evidence The ␴-class major nonstructural protein of MRV, ␴NS, that the only known fusogenic mammalian reoviruses localizes with the cytoskeletal fraction of cells, binds single- represent two distinct genogroups. stranded RNA (ssRNA) in a sequence-independent manner, and has been implicated in the earliest stages of RNA Comparison of the major ␴-class core, nonstructural, packaging and progeny virion assembly (Antczak and Jok- and outer capsid proteins of the fusogenic and lik, 1992; Huismans and Joklik, 1976; Mora et al., 1987). The ␴ nonfusogenic orthoreoviruses N-terminus of the MRV NS protein has recently been shown to be important in influencing ssRNA binding (Gillian Previous reports delineated conserved sequence-pre- and Nibert, 1998) and was previously recognized as one of dicted structural motifs in the MRV ␴-class proteins, the more highly conserved regions of the MRV ␴NS pro- based on a comparative analysis of different MRV strains teins (Weiner and Joklik, 1987). The ARV ␴NS protein also (Dermody et al., 1991; Kedl et al., 1995; Weiner and Joklik, binds ssRNA (Yin et al., 1998), and the N-terminal 11 resi- 1987). The highly conserved nature of the primary se- dues are relatively well conserved between the fusogenic quences within the MRV genogroup makes interpretation and nonfusogenic reoviruses (Fig. 5). However, the am- of the significance of these conserved sequence-pre- phipathic-helical nature of the N-terminal sequence in the dicted structural motifs difficult. However, the conserva- MRV ␴NS proteins (Gillian and Nibert, 1998) is not con- tion of these motifs between the four highly diverged served in the fusogenic orthoreovirus ␴NS proteins (data genogroups of reoviruses would lend added support to not shown), suggesting that an amphipathic helix at the the speculation that these motifs are functionally signif- N-terminus of ␴NS is likely not required for ssRNA binding. icant. Accordingly, the homologous proteins of the fuso- The greatest sequence conservation in the ␴NS proteins genic reoviruses were examined for the presence of occurs over the middle third of the protein (alignment po- these conserved structural motifs. sitions 137–276 in Fig. 5). This region exhibits highly con- The ␴-class major core protein of MRV, ␴2, represents served hydropathy profiles in all of the ␴NS proteins and a core structural protein that binds dsRNA and is impli- consists primarily of predicted ␤ sheets, turns, and loops in cated in virion assembly and possibly RNA synthesis the N-terminal portion with a highly conserved extended (Nibert et al., 1996; Ramig et al., 1978; Schiff et al., 1988). helical domain at the C-terminus (data not shown). The Sequence analysis suggested a two-domain model for conserved primary and predicted secondary structures of ␴2 with a predominantly ␤ sheet structure over the N- this central region in all four genogroups are suggestive of proximal three fourths of the protein and a smaller hy- a common structural domain in the reovirus ␴NS proteins. drophilic C-proximal domain with high helix propensity The ␴3 major outer capsid protein of MRV forms the (Dermody et al., 1991). Despite the extensive sequence outermost surface of the virus particle, associates with the divergence, all of the fusogenic reovirus ␴-class major ␮-class major outer capsid protein, binds dsRNA and zinc, core proteins exhibit the same two-domain structure and is involved in the assembly of replication complexes (data not shown). The conserved two-domain model is and in the regulation of translation in virus-infected cells consistent with the hypothesis that the N-terminal do- (Antczak and Joklik, 1992; Huismans and Joklik, 1976; Ja- main may mediate the participation of ␴2 as a compo- cobs and Langland, 1998; Nibert, 1998; Schiff, 1998; Schiff et nent of the icosahedral shell of the reovirus core, al., 1988; Schmechel et al., 1997). The ␴-class major outer ORTHOREOVIRUS PHYLOGENY 321

FIG. 4. The aligned ␴-class major core proteins of the fusogenic and nonfusogenic reoviruses. The cDNA-predicted amino acid sequences of the ␴-class major core proteins of the fusogenic and nonfusogenic reoviruses were aligned using PILEUP and shaded using BOXSHADE. Amino acid residues that are identical in four of the seven proteins are indicated by black background shading. Gray shading indicates conservative amino acid substitutions in four of the seven sequences. A region of similarity previously noted between the MRV sequences and the RNA polymerase of E. coli is overlined. The sequences are identified using a that indicates the particular virus [MRV strains Lang, Dearing, and Jones (La, De, Jo), ARV-176 and -138 (76, 38), NBV (nb), and BRV (bv)] and genome segment (S1, S2). capsid proteins displayed the highest levels of sequence the C-terminus of the MRV ␴3 protein (alignment positions divergence (Fig. 3), presumably a reflection of evolutionary 257–263 and 316–322 in Fig. 6) have been implicated in the pressures exerted by the host immune response against dsRNA-binding activity of the protein (Mabrouk et al., 1995; this external capsid protein. The majority of the conserved Miller and Samuel, 1992; Wang et al., 1996). Only the sec- amino acid positions are located in the N-terminal 80 res- ond of these two basic regions displayed some limited idues (Fig. 6). This region includes the previously identified conservation in the ARV and NBV sequences (Fig. 6), which zinc-finger motif (Schiff et al., 1988; Mabrouk and Lemay, was completely absent in the BRV sequence. The complete 1994), which is conserved as a CCHC-finger in all of the absence of the first basic region and the limited conserva- orthoreovirus outer capsid proteins. Two basic regions near tion of basic residues in the second domain, which is 322 ROY DUNCAN

FIG. 5. The aligned ␴-class major nonstructural proteins of the fusogenic and nonfusogenic reoviruses. The cDNA-predicted amino acid sequences of the ␴-class nonstructural proteins of the fusogenic and nonfusogenic reoviruses were aligned using PILEUP and shaded using BOXSHADE. The binomial nomenclature for the individual sequences and the shading scheme are the same as in Fig. 4. A conserved region implicated in ssRNA binding (thin overline) and a central region with conserved secondary structure (thick overline) are indicated. intolerant of amino acid substitutions to basic residues in pologies of the phylogenetic trees generated for each of the MRV ␴3 protein (Wang et al., 1996), are consistent with three ␴-class proteins are presented in Fig. 7. In all the recent proposal by Yin et al. (1997) that the outer capsid cases, of the 945 possible trees for seven taxa, only one proteins of ARV may not bind dsRNA. The absence of tree was preferred (i.e., required the fewest number of dsRNA binding by the fusogenic reovirus ␴-class major changes). outer capsid proteins, a property implicated in the interac- This phylogenetic analysis supported the subdivision tion of MRV with the dsRNA-activated translational regula- of the orthoreoviruses into at least four distinct geno- tory pathways in cells (Imani and Jacobs, 1988; Jacobs and groups that cluster into three clades. The prototype MRV Langland, 1998; Lloyd and Shatkin, 1992; Schmechel et al., isolates represent a distinct clade, whereas the ARV 1997), may profoundly influence host cell interactions by isolates are grouped in a second clade. The greater these viruses. extent of amino acid conservation between ARV and NBV in all three proteins resulted in the grouping of the Phylogenetic relationships of the fusogenic and fusogenic mammalian reovirus NBV as a separate geno- nonfusogenic orthoreoviruses group within the ARV clade (Fig. 7). The second fuso- The evolutionary relationships between the four or- genic mammalian reovirus, BRV, clearly segregates into thoreovirus genogroups was determined by maximum a third clade that is more distantly related to both the parsimony analysis using exhaustive searches; the to- MRV and ARV clades. The fact that the tree topologies ORTHOREOVIRUS PHYLOGENY 323

FIG. 6. The aligned ␴-class major outer capsid proteins of the fusogenic and nonfusogenic reoviruses. The cDNA-predicted amino acid sequences of the ␴-class major outer capsid proteins of the fusogenic and nonfusogenic reoviruses were aligned using PILEUP and colored using BOXSHADE. The binomial nomenclature for the individual sequences and the shading scheme are the same as in Fig. 4. The location of the conserved CCHC zinc finger is indicated. Open circles over the aligned sequences indicate the location of conserved basic residues in the two RNA-binding motifs present in the MRV sequences. were identical for all three S-segment-encoded proteins protein, suggesting that the orthoreoviruses represent a indicates that natural genome segment did relatively homogeneous genus (e.g., Dermody et al., not contribute to the recent evolution of these orthoreo- 1991; Duncan et al., 1990; Kedl et al., 1995; Weiner and viruses, at least for the three S-class genome segments Joklik, 1987, 1989). Contrary to this perception, the current examined. These results provide the first basis for sub- study revealed that the orthoreoviruses constitute a division of the orthoreoviruses into distinct taxonomic highly diverse genus with pairwise identities between units based on their inferred evolutionary relationships. the homologous proteins of the four genogroups that range from 16% to 32%. DISCUSSION To place these percent identities in a broader context, the human, bovine, simian, and avian members of the Extensive divergence of the orthoreoviruses group A possess amino acid identities in the The orthoreoviruses have been characterized on the major core protein VP2 that range from approximately basis of extensive sequence analysis of the prototype 75% to 98% (Murphy et al., 1995). However, percent iden- mammalian reoviruses that display amino acid se- tities between the VP2 proteins of the group A rotavi- quence homologies exceeding 85% in all except the ␴1 ruses with the group B and C rotaviruses decrease to 324 ROY DUNCAN

FIG. 7. Phylogenetic trees of the fusogenic and nonfusogenic orthoreoviruses. Phylogenetic trees were generated using the aligned amino acid sequences of the major ␴-class core, nonstructural, and outer capsid proteins of the fusogenic and nonfusogenic orthoreoviruses. Maximum parsimony analysis with exhaustive searches was performed using PAUP (Swofford, 1993). The resulting phylograms were rooted at the midpoint of the longest branch and the single most parsimonius tree is presented for each protein. The overall tree lengths were 695, 703, and 785 steps for the core, nonstructural, and outer capsid proteins, respectively. The horizontal branch lengths are proportional to number of inferred changes (bar indicates the scale for 100 steps); vertical distances are arbitrary. The sequences are identified using the same binomial nomenclature scheme described in Fig. 4.

16–47%, in the same range as percent identities between Phylogeny of the orthoreoviruses the VP12 proteins of different groups of (Mur- The degree of dissimilarity in the four genogroups of phy et al., 1995) and similar to the extent of divergence of orthoreoviruses indicates that these viruses have been the orthoreoviruses demonstrated in this report. This following separate evolutionary trajectories and served extensive divergence between, and even within, individ- to clarify the origins of the fusogenic mammalian reovi- ual genera in the family Reoviridae contrasts with per- ruses. Before this study, the origins of the atypical fuso- cent amino acid identities observed in some other RNA virus families. For example, identities between different genic mammalian reoviruses, NBV and BRV, were un- subgroups of , a large and diverse genus in clear. It was reasonable to assume that NBV and BRV the family Picornaviridae, range from 53% to 84% (Poyry may have originated recently, either from a fusogenic et al., 1996), whereas percent nucleotide identities be- ARV that crossed the species barrier or from a nonfuso- tween the 14 clades of exceed 53% (Kuno et genic MRV that acquired the fusion-inducing genome al., 1998). Clearly, members of the various genera in the segment from ARV after natural genome segment reas- family Reoviridae are among the most evolutionarily di- sortment. The present results, however, clearly indicate vergent of RNA viruses, and as shown in this report, the that NBV and BRV represent distinct replicating lineages orthoreoviruses are no exception. that have been evolving independently for an extended Despite the level of primary sequence divergence, the period of time. four genogroups of orthoreoviruses display similar bio- Aside from the ARV–NBV comparison, almost all of the chemical and biophysical properties, suggesting that es- other pairwise comparisons revealed amino acid identi- sential structural domains remain conserved. This con- ties of approximately 20–30% (Table 2), suggesting that tention is supported by the present comparative se- ARV, MRV, and BRV all diverged from each other at about quence-based structural predictions of the homologous the same point in time, followed by the more recent proteins encoded by these diverged viruses that sug- divergence of the ARV and NBV lineages (Fig. 7). Assum- gested the presence of conserved structural domains. In ing a nucleotide substitution rate of approximately 2.2 ϫ Ϫ3 addition, recent cryoelectron microscopy image recon- 10 nucleotide substitutions/site/year, the predicted rate structions of ARV particles indicate that ARV and MRV of error for the bluetongue virus RNA polymerase (Kow- share similar, although not identical, capsid morpholo- alik and Li, 1991), the fusogenic and nonfusogenic or- gies (S. Walker, M. L. Nibert, T. Baker, and R. Duncan, thoreoviruses are estimated to have diverged from each unpublished observations). However, sequence analysis other at least 300 years ago, followed by the divergence also indicated that some motifs that are conserved in of the ARV and NBV lineages approximately 100 years MRVs are absent from the fusogenic reoviruses, sug- later. However, these are likely minimum estimates that gesting that additional comparative structural and bio- almost certainly underestimate the time of divergence of chemical studies are likely to reveal distinct biological the orthoreovirus genogroups if we accept the hypothe- differences between the four genogroups of orthoreovi- sis that the rate of amino acid substitution of a protein ruses. may significantly decrease as some minimum level of ORTHOREOVIRUS PHYLOGENY 325 conservation of essential structural and functional motifs dence of a particular protein or genome segment with a is approached (reviewed in Li, 1997). significantly higher percent of identity between discor- Assuming that a particular protein has a relatively dant virus pairs, suggesting that reassortment has not constant initial rate of amino acid substitution after di- influenced the recent evolution of the four orthoreovirus vergence, the phylograms and the extent of amino acid genotypes. These results are consistent with the pro- substitution suggest that the three clades (BRV, ARV/ posal that these genogroups represent independently NBV, and MRV) have been evolving separately for ap- evolving lineages and, hence, warrant a species desig- proximately the same length of time. Although the early nation. divisions that lead to the current extant species cannot In the absence of any evidence indicating genome be determined from the currently available data, the segment reassortment between two reovirus isolates, levels of sequence identity imply that a primary bifurca- the species designation must be inferred from other tion in the tree, for example, to generate the fusogenic properties of the viruses such as the extent of antigenic and nonfusogenic lineages, was followed almost imme- conservation, genome organization, biological proper- diately by a second bifurcation to generate the lineage ties, and nucleotide and amino acid conservation (Van for the third clade. In addition to the uncertain rooting of Regenmortel et al., 1997). The proposed species classi- the phylogenetic trees, it is not clear from the present fication of the orthoreoviruses is supported by the exten- parsimony analysis whether the unusual syncytium-in- sive divergence of the amino acid sequences of homol- ducing property of the fusogenic reoviruses represents ogous proteins (Table 2 and Fig. 3), by their distinct an ancestral state or the acquirement of a new property. electropherotypes (the characteristic migration of the For example, depending on how the trees are rooted, a ARV and NBV S1 genome segments and the BRV S4 single gain or loss of the fusogenic property could ex- genome segment) (Fig. 1B), by their host ranges and plain the appearance of the extant species. Ascertaining syncytium-inducing properties, and by previous studies the precise sequence of events that lead to the ARV, that demonstrated limited antigenic similarity in the virus MRV, and BRV lineages (i.e., rooting the phylogenetic structural proteins as determined by immunoprecipita- tree) will require additional sequence or biological infor- tion (Duncan et al., 1995). The cumulative weight of these mation from a defined outlier, possibly the fusogenic experimental data supports the designation of ARV, MRV, reptilian reovirus isolates (Ahne et al., 1987; Vieler et al., and BRV as distinct species isolates. 1994). In the case of NBV, the results demonstrate that NBV and ARV share a characteristic retarded gel mobility of A species proposal for the genus Orthoreovirus the S1 genome segment (Fig. 1B), more extensive con- servation in their 5Ј-terminal genome segment nucleo- Based on the available data, it is proposed that the tide sequences (Fig. 2), and higher overall sequence four reovirus genogroups be assigned as separate virus identity (Fig. 3A), indicating that NBV is more closely species. The species designation in virus classification related to the ARV species than to either of the other two generally follows the concept of independently evolving species. However, the 40–60% amino acid identity be- lineages and should reflect phylogenetic relationships tween NBV and ARV is well below that observed be- (Mayo, 1996; Van Regenmortel et al., 1997). In the case of tween different ARV isolates (Ͼ95% identity). Although it viruses with segmented such as the Reoviri- is not possible to define a single level of percent identity dae, the basis for assigning a virus to the species class, that delineates virus species in different genera (Van following the concept of replication in genetic isolation, Regenmortel et al., 1997), it is proposed that for the should reflect the ability of different isolates to exchange orthoreoviruses a species definition should include genetic material via reassortment of genome segments. amino acid identities that exceed 85% in the majority of By this criterion, the various nonfusogenic MRV isolates the encoded products, particularly in the more are clearly representatives of a single species as dem- conserved core proteins of the virus. Similar levels of onstrated by extensive reassortment studies (reviewed identity are observed between the members of particular in Joklik and Roner, 1995; Ramig and Fields, 1983; Ramig subgroups of other genera in the family Reoviridae (Mur- and Ward, 1991). The species classification also applies phy et al., 1995). Based on the percent identity between to the ARV isolates that generate reassortants after coin- ARV and NBV, in conjunction with the avian versus mam- fection in cell culture (Duncan and Sullivan, 1998; Ni and malian host range preferences, it is proposed that NBV Kemp, 1990, 1992). be given a separate species classification in the genus The present sequence analysis of three of the S-class Orthoreovirus. genome segments provided no evidence of recent ge- Although the four genogroups of orthoreoviruses have nome segment reassortment between the four geno- been assigned to the species taxon, the extent of diver- groups, as evidenced by the similar topologies of all gence of the orthoreoviruses and the Reoviridae in gen- three trees (Fig. 7), suggesting that these isolates repre- eral, coupled with the numerous unique biological prop- sent separate orthoreovirus species. There was no evi- erties observed between different genera in the Reoviri- 326 ROY DUNCAN

dae, raises the question of whether the existing genera min. To extend truncated cDNA products, the reverse might be better considered as subfamilies or possibly transcriptase reaction was repeated a second time. An families within a new order Reovirales. In such a case, aliquot (6 ␮l) of the final cDNA mixture was added to a the three clades of orthoreoviruses would be elevated to 100-␮l PCR containing Thermopol reaction buffer (New

the genus level, one of which contains two species England Biolabs), 2 mM MgSO4, 0.2 mM deoxynucleoti- ␮ groups: ARV and NBV. Such a reclassification will obvi- des, 1 g of oligo(dT)18–NotI primer, and2UofVent ously require considerable discussion among members polymerase (New England Biolabs). The PCR protocol of the Reoviridae study group of the ICTV. included an incubation at 94°C for 5 min, followed by 35 cycles at 94°C for 1 min, 59°C for 2 min, and 72°C for 4 MATERIALS AND METHODS min and a final 10-min incubation at 72°C. The amplified cDNA was isolated by phenol–chloro- Viruses and cells form extraction and ethanol precipitation, digested with ARV strain 176 (ARV-176) has been previously de- NotI, and gel-purified from 1% agarose gels using Gene- scribed (Duncan et al., 1996). Strain SK138a (ARV-138) Clean (BIO 101). The digested cDNA was ligated to was isolated from the hock joint of an infected chicken in NotI-digested, dephosphorylated pBlueScript II SK (Strat- ␣ New Brunswick (Drastini et al., 1992) and was obtained agene) and transformed into E. coli DH5 , and clones from Frederick Kibenge (Atlantic Veterinary College, Uni- were isolated by blue-white selection in the presence of versity of Prince Edward Island). BRV was isolated from IPTG and X-gal. DNA was isolated by alkaline– the brain tissue of a baboon with meningoencephalomy- SDS extraction (Sambrook et al., 1988), digested with elitis (Duncan et al., 1995). NBV was isolated from the NotI, and fractionated by agarose gel electrophoresis. heart blood of a flying fox (Gard and Compans, 1970) and Clones containing cDNA inserts the approximate size of was obtained from Terrence Dermody (Vanderbilt Univer- full length S-class genomic dsRNA were sequenced us- sity). The ARV isolates were plaque purified and ampli- ing a Licor automated sequencer at the NRC-Dalhousie fied to pass four using a multiplicity of infection of 0.01 in Core Sequencing Facility. All clones were sequenced a continuous quail cell line, QM5 (Antin and Ordahl, completely in both directions. Extensive sequence (600– Ј Ј 1991), as previously described (Duncan et al., 1996). NBV 700 nucleotides from the 3 - and 5 -termini) was ob- and BRV were similarly plaque purified and amplified in tained from clones generated from independent RT-PCRs monkey Vero cells. for the NBV S2 genome segment and the S3 genome segments of NBV, ARV-176, and ARV-138. No anomalies cDNA cloning and sequencing were detected from the redundant sequencing.

Viral dsRNA was isolated from concentrated virus Sequence analysis stocks as previously described (Duncan et al., 1995). The genomic RNA (20 ␮g) was denatured in 10 ␮lof90% Sequences were compiled and analyzed using the dimethyl sulfoxide (Sigma) in heat-sealed glass capillar- University of Wisconsin GCG software, version 8 (Deve- ies for 30 min at 45°C. The RNA was poly(A)ϩ-tailed by reaux et al., 1984). The cDNA sequences and the pre- expelling the denatured RNA into preheated reaction dicted amino acid sequences of the encoded gene prod- mixture (100 ␮l) containing3UofE. coli poly(A)ϩ poly- ucts were used in pairwise comparisons to determine merase ( Technologies Inc.) and incubating at 37°C the extent of amino acid and nucleotide identity of ho- for 10 min according to the protocol of Cashdollar et al. mologous genome segments, using the GAP algorithm of (1985). The reaction was stopped by the addition of EDTA the GCG software. Multiple alignments of the homolo- and SDS to 30 mM and 0.15%, respectively, and the tailed gous protein sequences of three prototype MRV strains, RNA was isolated by phenol–chloroform extraction and obtained from GenBank, and the four fusogenic orthoreo- ethanol precipitation. The tailed dsRNA was fractionated viruses sequenced in this report were obtained using the on 1% agarose gels, and the S-class genome segments sequential alignment algorithm of PILEUP in the GCG were excised from the gel and isolated using the RNaid software suite. In the case of the highly diverged ␴-class protocol (BIO 101). major outer capsid protein sequences, the gap weight Tailed, gel-purified dsRNA (2 ␮g) was mixed with 1 ␮g and gap extension penalties were increased, and the ␮ of an oligo(dT)18–NotI primer in a total volume of 12 lof preliminary end-weighted alignments were visually ad- water. The sample was heated to 95°C for 10 min, chilled justed. The aligned sequences were used as input for on ice, and added to preheated (42°C) reverse transcrip- phylogenetic analysis, performed using the maximum tase reaction mixture containing 4 ␮lof5ϫ Superscript parsimony software of PAUP, version 3.1 (Swofford, 1993). RT reaction buffer, 2 ␮lof0.1mMDTT,1␮lof10mM All gaps introduced to maintain the alignments were deoxynucleotides, and 1 ␮l (200 U) of Superscript RT treated as missing characters. Trees were constructed (Life Technologies Inc.). The reaction was incubated at for each of the three homologous proteins from the 42°C for 60 min and terminated by heating at 95°C for 10 seven taxa using exhaustive searches. In all cases, only ORTHOREOVIRUS PHYLOGENY 327

a single most parsimonius tree was generated. These syncytium-inducing and pathogenic capabilities. Virology 250, 263– unrooted trees were displayed as phylograms that were 272. rooted at the midpoint of the longest branch. Duncan, R., Horne, D., Cashdollar, L. W., Joklik, W. L., and Lee, P. W. K. (1990). Identification of conserved domains in the cell attachment proteins of the three serotypes of reovirus. Virology 174, 399–409. Accession numbers Duncan, R., Murphy, F. A., and Mirkovic, R. (1995). Characterization of a novel syncytium-inducing baboon reovirus. Virology 212, 752–756. The accession numbers for the various genome seg- Gard, G., and Compans, R. W. (1970). Structure and cytopathic effects of ments of the fusogenic orthoreoviruses sequenced in Nelson Bay virus. J. Virol. 6, 100–106. this report are as follows: ARV-176: S2-AF059716, S3- Gillian, A. L., and Nibert, M. L. (1998). Amino terminus of reovirus AF059720, S4-AF059724; ARV-138: S2-AF059717, S3- nonstructural protein sigma NS is important for ssRNA binding and AF059721, S4-AF059725; NBV: S2-AF059718, S3- nucleoprotein complex formation. Virology 240, 1–11. Huismans, H., and Joklik, W. K. (1976). Reovirus-coded polypeptides in AF059726, S4-AF059722; and BRV: S1-AF059719, S2- infected cells: Isolation of two native monomeric polypeptides with AF059723, S3-AF059727. The homologous genome affinity for single-stranded and double-stranded RNA, respectively. segments of the nonfusogenic mammalian reoviruses Virology 70, 411–424. were obtained from GenBank: MRV-La: S2-L19774, S3- Imani, F., and Jacobs, B. L. (1988). Inhibitory activity for the interferon- ␴ M18389, S4-M13139; MRV-Jo: S2-L19775, S3-M18390, S4- induced protein kinase is associated with the reovirus serotype 1 3 protein. Proc. Natl. Acad. Sci. USA 83, 7887–7891. X60066; and MRV-De: S2-L19776, S3-U35349, S4-K02739. Jacobs, B. L., and Langland, J. O. (1998). Reovirus ␴3 protein: dsRNA binding and inhibition of RNA-activated protein kinase. Curr. Topics ACKNOWLEDGMENTS Microbiol. Immunol. 233, 185–196. Joklik, W. K., and Roner, M. R. (1995). What reassorts when reovirus The expert technical assistance of Zhaoxia Chen and Jingyun Shou genome segments reassort? J. Biol. Chem. 270, 4181–4184. is gratefully acknowledged. The author thanks John Logsdon for useful Kedl, R., Schmechel, S., and Schiff, L. (1995). Comparative sequence discussions on parsimony analysis. This research was supported by a analysis of the reovirus S4 from 13 serotype 1 and serotype 3 grant from the Medical Research Council of Canada. field isolates. J. Virol. 69, 552–559. Kowalik, T. F., and Li, J. K.-K. (1991). Bluetongue virus evolution: Se- REFERENCES quence analysis of the genomic S1 segments and major core protein VP7. Virology 181, 749–755. Ahne, W., Thomsen, I., and Winton, J. (1987). Isolation of a reovirus from Kuno, G., Chang, G.-J. J., Tsuchiya, K. R., Karabatsos, N., and Cropp, the snake, Python regius. Arch. Virol. 94, 135–139. C. B. (1998). Phylogeny of the genus . J. Virol. 72, 73–83. Antczak, J. B., and Joklik, W. K. (1992). Reovirus genome segment Li, W.-H. (1997). “Molecular Evolution.” Sinauer Associates, Massachu- assortment into progeny genomes studied by the use of monoclonal setts. antibodies directed against reovirus proteins. Virology 187, 760–776. Lloyd, R. M., and Shatkin, A. J. (1992). Translational stimulation by Antin, P. B., and Ordahl, C. P. (1991). Isolation and characterization of an reovirus polypeptide ␴3: Substitution for VAI RNA and inhibition of avian myogenic cell line. Dev. Biol. 143, 111–121. phosphorylation of the ␣ subunit of eukaryotic initiation factor 2. Bigot, Y., Drezen, J.-M., Sizaret, P.-Y., Rabouille, A., Hamelin, M.-H., and J. Virol. 66, 6878–6884. Periquet, G. (1995). The genome segments of DpRV, a commensal Mabrouk, T., Danis, C., and Lemay, G. (1995). Two basic motifs of reovirus of the wasp Diadromus pulchellus (Hymenoptera). Virology reovirus ␴3 protein are involved in double-stranded RNA binding. 210, 109–119. Biochem. Cell Biol. 73, 137–145. Calisher, C. H., Horzinek, M. C., Mayo, M. A., Ackerman, H.-W., and Mabrouk, T., and Lemay, G. (1994). Mutations in a CCHC zinc-binding Maniloff, J. (1995). Sequence analysis and a unifying system of virus motif of the reovirus ␴3 protein decrease its intracellular stability. taxonomy: Consensus via consent. Arch. Virol. 140, 2093–2099. J. Virol. 68, 5287–5290. Cashdollar, L. W., Chmelo, R. A., Wiener, J. R., and Joklik, W. K. (1985). Mayo, M. A. (1996). Recent revisions of the rules of virus classification Sequences of the S1 genes of the three serotypes of reovirus. Proc. and nomenclature. Arch. Virol. 141, 2479–2484. Natl. Acad. Sci. USA 82, 24–28. Mayo, M. A., and Pringle, C. R. (1997). Virus taxonomy—1997. J. Gen. Chiu, C. J., and Lee, L. H. (1997). Cloning and nucleotide sequencing of Virol. 79, 649–657. the S4 genome segment of avian reovirus S1133. Arch. Virol. 142, Miller, J. E., and Samuel, C. E. (1992). Proteolytic cleavage of the 2515–2520. reovirus sigma 3 protein results in enhanced double-stranded RNA- Dermody, T. S., Schiff, L. A., Nibert, M. L., Coombs, K. M., and Fields, binding activity: Identification of a repeated basic amino acid motif B. N. (1991). The S2 gene nucleotide sequences of prototype strains within the C-terminal binding region. J. Virol. 66, 5347–5356. of the three reovirus serotypes: Characterization of reovirus core Mora, M., Partin, K., Bhatia, J., and Carter, C. (1987). Association of protein ␴2. J. Virol. 65, 5721–5731. reovirus proteins with the structural matrix of infected cells. Virology Devereux, J., Haeberli, P., and Smithies, O. (1984). A comprehensive set 159, 265–277. of sequence analysis programs for the VAX. Nucleic Acids Res. 12, Murphy, F. A., Fauquet, C. M., Bishop, D. H. L., Ghabrial, S. A., Jarvis, 387–395. A. W., Martelli, G. P., Mayo, M. A., and Summers, M. D. (Eds.) (1995). Drastini, Y., Kibenge, F. S. B., McKenna, P. K., and Lopez, A. (1992). “Virus Taxonomy: Sixth report of the International Committee on Comparison of eight different procedures for harvesting avian reo- Taxonomy of Viruses.” Springer-Verlag, Wien New York (Arch. Virol. viruses grown in Vero cells. J. Virol. Methods 39, 269–278. [suppl]10). Duncan, R., Chen, Z., Walsh, S., and Wu, S. (1996). Avian reovirus- Ni, Y., and Kemp, M. C. (1990). Selection of genome segments following induced syncytium formation is independent of infectious progeny coinfection of chicken fibroblasts with avian reoviruses. Virology 177, virus production and enhances the rate, but is not essential, for 625–633. virus-induced cytopathology and virus egress. Virology 224, 453– Ni, Y., and Kemp, M. C. (1992). Strain-specific selection of genome 464. segments in avian reovirus coinfections. J. Gen. Virol. 73, 3107–3113. Duncan, R., and Sullivan, K. (1998). Characterization of two avian reo- Nibert, M. L. (1998). Structure of particles. viruses that exhibit strain-specific quantitative differences in their Curr. Topics Microbiol. Immunol. 233, 1–30. 328 ROY DUNCAN

Nibert, M. L., Schiff, L. A., and Fields, B. N. (1996). Reoviruses and their Schnitzer, T. J. (1985). Protein coding assignment of the S genes of the replication. In “Fundamental Virology,” 3rd ed. (B. N. Fields, D. M. avian reovirus S1133. Virology 141, 167–170. Knipe, and P. M. Howley, Eds.). Lippincott-Raven Press, New York. Swofford, D. L. (1993). “PAUP: Phylogenetic Analysis Using Parsimony, Pereira, H. G. (1991). Double-stranded RNA viruses. Semin. Virol. 2, ver. 3.1.” Illinois Natural History Survey, Champaign, IL. 39–53. Vakharia, V. N. (1997). Cloning and expression of the S1 and S3 genome Poyry, T., Kinnunen, L., Hyypia, T., Brown, B., Horsnell, C., Hovi, T., and segments of avian reovirus strain 1733. GenBank, direct submission. Stanway, G. (1996). Genetic and phylogenetic clustering of enterovi- Van Regenmortel, M. H. V. (1990). Virus species, a much overlooked ruses. J. Gen. Virol. 77, 1699–1717. but essential concept in virus classification. Intervirology 31, 241– Pringle, C. R. (1991). The 20th meeting of the Executive Committee of 254. ICTV: Virus species, higher taxa, a universal database and other Van Regenmortel, M. H. V., Bishop, D. H. L., Fauquet, C. M., Mayo, M. A., matters. Arch. Virol. 119, 303–304. Maniloff, J., and Calisher, C. H. (1997). Guidelines to the demarcation Ramig, R. F., and Fields, B. N. (1983). The genetics of reovirus. In “The of virus species. Arch. Virol. 142, 1505–1518. Reoviridae” (W. K. Joklik, Ed.). Plenum Press, New York. Varela, R., and Benavente, J. (1994). Protein coding assignment of avian Ramig, R. F., Mustoe, T. A., Sharpe, A. H., and Fields, B. N. (1978). A reovirus strain S1133. J. Virol. 68, 6775–6777. genetic map of reovirus. II. Assignment of the double-stranded RNA- Vieler, E., Baumgartner, W., Herbst, W., and Kohler, G. (1994). Charac- negative mutant groups C, D, and E genome segments. Virology 85, terization of a reovirus isolate from a rattle snake, Crotalus viridis, 531–534. with neurological dysfunction. Arch. Virol. 138, 341–344. Ramig, R. F., and Ward, R. L. (1991). Genomic segment reassortment in Wang, Q., Bergeron, J., Mabrouk, T., and Lemay, G. (1996). Site-directed rotaviruses and other reoviridae. Adv. Virus Res. 39, 163–207. mutagenesis of the double-stranded RNA binding domain of bacte- Robertson, M. D., and Wilcox, G. E. (1986). Avian reovirus. Vet. Bull. 56, ␴ 726–733. rially expressed 3 reovirus protein. Virus Res. 41, 141–151. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1988). “Molecular Cloning: Wiener, J. R., and Joklik, W. K. (1987). Comparison of the reovirus A Laboratory Manual,” 2nd ed. Cold Spring Harbor Laboratory Press, serotype 1, 2, and 3 S3 genome segments encoding the nonstructural ␴ Cold Spring Harbor, New York. protein NS. Virology 161, 332–339. Schiff, L. A. (1998). Reovirus capsid proteins ␴3 and ␮l: Interactions that Wiener, J. R., McLaughlin, T., and Joklik, W. K. (1989). The sequences of influence , assembly, and translational control. Curr. Topics the S2 genome segments of reovirus serotype 3 and of the dsRNA- Microbiol. Immunol. 233, 167–183. negative mutant ts447. Virology 170, 340–341. Schiff, L. A., Nibert, M. L., Co, M. S., Brown, E. G., and Fields, B. N. Yin, H. S., Shieh, H. K., and Lee, L. H. (1997). Characterization of the (1988). Distinct binding sites for zinc and double-stranded RNA in the double-stranded RNA genome segment S3 of avian reovirus. J. Virol. reovirus outer capsid protein sigma 3. Mol. Cell Biol. 8, 273–283. Methods 67, 93–101. Schmechel, S., Chute, M., Skinner, P., Anderson, R., and Schiff, L. (1997). Yin, H. S., Shieh, H. K., and Lee, L. H. (1998). Identification and char- Preferential translation of reovirus mRNA by a sigma3-dependent acterization of RNA-binding activities of avian reovirus non-structural mechanism. Virology 26, 62–73. protein sigmaNS. J. Gen. Virol. 79, 1411–1413.