structure and functionofthereovirusgenome · structure andfunction of thereovirus genome 487...

MICROBIOLOGICAL REVIEWS, Dec. 1981, p. 483-501 Vol. 45, No. 40146-0749/81/040483-19$02.00/0

Structure and Function of the Reovirus GenomeWOLFGANG K. JOKLIK

Department ofMicrobiology and Immunology, Duke University Medical Center,Durham, North Carolina 27710

INTRODUCTION ................ ................................ 483DISCOVERY OF REOVIRUSES .............................................. 483STRUCTURE OF THE REOVIRUS PARTICLE .................. ............. 484REOVIRUS GENOME................................................. 484PROTEINS OF REOVIRUS ................................................ 486INFORMATION CONTENT OF THE 10 REOVIRUS GENES ...... ............ 488ENZYMES IN REOVIRUS .................................................... 489TRANSCREPTION OF REOVIRUS RIBONUCLEIC ACID ....... .............. 489SEQUENCES OF REOVIRUS GENES AND MESSENGER RIBONUCLEIC

ACIDS ............................................ 490TRANSLATION OF REOVIRUS MESENGER RIBONUCLEIC ACIDS ........ 493ESSENTIAL FEATURES OF THE REOVIRUS MULTIPICATION CYCLE ..... 494

Reovirus Morphogenesis.................................................... 495Reovirus DeletionMutants.................................................. 496

TEMPERATURE-SENSITIVE MUTANTS OF REOVIRUS ...... ............... 496INTERACTION OF REOVIRUS WITH THE INTACT HOST ...... ............. 497LITERATURE CITED ......................................................... 498

INTRODUCTIONReovirses possess several unique properties:

their genome consists of double-stranded (ds)ribonucleic acid (RNA) and exists in the form of10 individual genes that are not linked cova-lently, and the reovirus particle, which consistsof two concentric icosahedral capsid shells, con-tains a ds -* single-stranded (ss) RNA transcrip-tase that transcribes the ds genome into messen-ger RNA (mRNA).There are three serotypes of mammalan reo-

virus, 1, 2 and 3, which are ubiquitous. There arealso several serotypes of avian reoviruses, anti-genically unrelated to the mammalian ones. Inaddition, there are numerous other viruses ofvertebrates, insects, and plants that share theunique properties listed above, but differ in thenumber of genes that they possess and in detailsof virus particle structure. These viruses aregrouped together into the family Reoviridae(Table 1). All appear to use the same strategy ofgenome expression.This review is concerned with the mammalian

orthoreoviruses. I will attempt to trace the high-lights of research on reovirus, demonstratinghow one discovery or concept led to the next,and focus on those unique aspects of reovirologythat have stimulated research on and providedinsight into other areas of virology and cell bi-ology.

DISCOVERY OF REOVIRUSESReoviruses were discovered in the early 1950s,

when tissue culture methods permissive of virusgrowth in vitro were being developed and when,under the influence of efforts to develop massvaccination against poliomyelitis, interest in en-teric viruses was intense. Beginning in 1950,many cytopathogenic agents that were neitherpoliovirus nor coxsackievirus were encounteredin the human alimentary tract (77, 79). Theseagents, many of which were isolated from ap-parently healthy persons, turned out to possessrather similar properties, and they were lumpedtogether under the acronym ECHO virus (en-teric cytopathogenic human orphan). Some-times referred to as viruses in search of diseases,many of them are in fact associated with clinicaldisease. They are now known to be members ofthe Enterovirus genus of the Picornaviridaefamily. Some isolates, however, although meet-ing the original and appropriately loose defini-tion of ECHO viruses, were soon found to bedifferent; the best known of these are the reovi-ruses. ECHO virus 10, in particular, was recog-nized as being much bigger and producing dis-tinctive cytopathic effects (77), and it becamereovirus serotype 1 (82). Soon two viruses pre-viously isolated from macaca monkeys and des-ignated SV12 and SV59 were shown to be iden-tical to reovirus serotypes 1 and 2, respectively(80), and a virus originally isolated from humansin suckling mice (107) and found to producehepato-encephalomyelitis in mice turned out tobe the reovirus that had been designated sero-type 3 (106).

483

on January 13, 2021 by guesthttp://m

mbr.asm

.org/D

ownloaded from

http://mmbr.asm.org/

TABLE 1. The Reoviridae

Genus Virus Host Symptoms in humans No. ofgenesOrthoreovirus Mammalian reoviruses (3 Humans and other None 10

serotypes) mammalsAvian reoviruses Birds

Orbivirus Numerous isolates; Mammals and insects Encephalitis (rare) 10transmitted by insects (culicoides, mosquitoes,Bluetongue virus ticks)African horsesickness virusEubenangee virus

Colorado tick fever virus Humans, ticks Encephalitis 12

Rotavirus Human rotavirus (2 Humans Infantile gastro- 11serotypes) enteritis (diarrhea)

Calf rotavirus (Nebraska calf Calvesdiarrhea virus)

Murine rotavirus (epizootic Micediarrhea of infant mice[EDIM])

Simian rotavirus (SAl1) MonkeysBovine or ovine rotavirus Cattle or sheep

("O" agent)Numerous other rotaviruses Other mammalian species

Cypovirus Cytoplasmic polyhedrosis Bombyx mori (silkworm) 10viruses (numerous strains) and other Lepidoptera,

Diptera, andHymenoptera

Phytoreovirus Wound tumor virus Plants, leaf hoppers 12Rice dwarf virus Plants, leaf hoppers

Fijivirus Maize rough dwarf virus Plants, leaf hoppers 10Fiji disease virus Plants, leaf hoppers

STRUCTURE OF THE REOVIRUSPARTICLE

The broad outlines of reovirus structure werelaid in the early 1960s. Starting with Rhim et al.(78), a long series of electron microscopic studiesled to the following conclusions. The reovirusparticle consists of two capsid shells with diam-eters of about 50 and 75 nm, respectively (Fig.1). Both consist of subunits or capsomers ar-ranged with icosahedral symmetry, but theirprecise arrangement has been very difficult todetermine and is, in fact, still not resolved.Whereas originally an outer capsid shell com-

posed of 92 columnar capsomers or 180 trun-cated pyramidal capsomers arranged around 92holes was suggested (111), more recently Luftiget al.(57) have proposed that there are 122 cap-somers, based on the observation that 20 cap-somers are clearly visible at the periphery ofnegatively stained virus particles. More recently,Palmer and Martin (72) have suggested thatthere are 32 morphological units arranged withT = 3 symmetry, each of which has the sym-metry of a T = 9 icosadeltahedron, being madeup of six separate wedge-shaped subunits that

are in turn made up of three smaller subunits. Aprominent feature of this structure is the sharingof subunits, an apparently unique feature of allmembers of the Reoviridae that makes preciseassignment of the symmetry system very diffi-cult.The outer capsid shell of reovirus is readily

digested by proteases like trypsin and chymo-trypsin, thereby yielding single-shelled cores.The arrangement of capsomers in the core iseven less defined than that in the outer capsidshell. Cores possess a unique structural feature:they exhibit 12 prominent icosahedrally locatedprojections or spikes that are hollow, thus pro-viding means of access to and exit from theparticle interior (57).

REOVIRUS GENOMEThe studies of Gomatos and collaborators in

the early 1960s that indicated that the RNA ofreovirus particles was double stranded createda great deal of interest, since they provided thefirst demonstration of the occurrence of ds RNAin nature. The original observation was thatcytoplasmic inclusions in L cells infected with

484 JOKLIK MICROBIOL. REV.


mbr.asm

.org/D

ownloaded from


STRUCTURE AND FUNCTION OF THE REOVIRUS GENOME 485

FIG. 1. Reovirus particles and reovirus cores. (A) Reovirus particles negatively stained with phospho-tungstic acid. Arrowspoint toparticularly clear capsomer-like structures. (B)A reoviruspartilephotographedafter 10-fold (C) and 12-fold (D) rotational enhancement. Note how structural details are brought out by theformer; in particular, the number ofperipheral capsomers is clearly seen to be 20, rather than a multiple of6. See Luftig et al. (57) for a discussion of the significance of this observation for constructing models forreovirus structure. (E and F) Reovirus cores stained with phosphotungstic acid. Note the icosahedralarrangement of the projections/spikes (57).

reovirus fluoresced pale green when treated withacridine orange (27). This is indicative of dsnucleic acids which bind only a small amount of

the dye and therefore stain orthochromatically;ss nucleic acids, which bind more dye, stainmetachromatically and fluoresce bright red. The

VOL. 45, 1981


mbr.asm

.org/D

ownloaded from


486 JOKLIK

inclusions proved to be resistant to deoxyribo-nuclease but susceptible to ribonuclease(RNase) at low salt concentrations, and it wassoon shown that the nucleic acid isolated frompurified virus particles was indeed ds RNA. Theevidence rests on the following facts, amongothers: (i) the RNA exhibits very sharp meltingprofiles, with the Tm depending on the ionicstrength (9, 88); (ii) it is resistant to RNase, theresistance depending on the concentration ofboth monovalent and divalent cations, as well ason the concentration of RNase (9, 88); (iii) it issusceptible to RNase Ill, which is specific for dsRNA; (iv) formaldehyde fails to induce hyper-chromicity (26); (v) its density in Cs2SO4 is 1.61g/ml, rather than 1.65 g/ml, which is character-istic of ss RNA (33, 88); (vi) its base compositionindicates equality of adenine and uridine, as wellas of guanine and cytosine (9, 26); and (vii) X-ray diffraction patterns are consistent with dou-ble strandedness (3, 49).

Attempts to extract this RNA from reovirusparticles in the form of one long molecule failed,and evidence soon accumulated that the reovirusgenome exists in the form of a collection ofdiscrete and unique segments that proved to begenes. The first indication of this was the elec-tron microscopic demonstration that reovirusRNA consists of a population of molecules thatexhibit a trimodal length distribution with max-ima at 1.1, 0.6, and 0.35 ,um, corresponding mo-lecular weights of about 2.5, 1.4, and 0.8 million(16, 110). Clearly, even the largest of these mol-ecules corresponded to only a portion of thegenome, since reovirus particles had alreadybeen shown to contain about 14.6% RNA, cor-responding to an aggregate molecular weight ofat least 10 x 106 (26). It was then shown that,regardless of the means used to liberate it, reo-virus RNA displays three size classes of mole-cules when analyzed on sucrose density gra-dients: these are the L, M, and S species ofmolecules, which sediment with 14, 12, and10.5S, corresponding to molecular weights ofabout 2.7, 1.4, and 0.7 million, respectively, orabout 4,500, 2,300, and 1,200 nucleotide basepairs (9, 88, 112). The molecules in these threesize classes were shown to be discrete segmentsrather than random fragments of larger mole-cules by the fact that they did not hybridize witheach other and that they were transcribed intospecific species of mRNA molecules within in-fected cells (7, 9, 114). These three size classeswere then further separated by polyacrylamidegel electrophoresis into 10 discrete and uniquemolecular species (Fig. 2) which are present inequimolar amounts and possess an aggregatemolecular weight of about 15 x 106 (93). Finalproof of the segmented nature of the reovirus

MICROBIOL. REV.

II I1 1 I I IILI L2 L3 MI M2 M3 Si S2 S3 S4

FIG. 2. Autoradiogram of a gel in which the 10genes of the genome of reovirus serotype 3 (strainDearing) had been electrophoresed. The direction ofelectrophoresis was from left to right. (Courtesy ofA.R. Schuerch.)

genome came with the demonstration that evenbefore extraction there are 20 3' termini in theRNA within each reovirus particle (62).Although there is no doubt that the 10 reovi-

rus genes are discrete molecules, the nature oftheir arrangement within virus particles is stillnot clear. Granboulan and Niveleau (28) wereable to liberate, with very low frequency anddemonstrable only by electron microscopic anal-ysis, very long linear arrays of RNA from reovi-rus particles, and more recently Kavenoff et al.(40) were able to release the RNA in the forn ofspider-like arrangements, with all genes con-nected at a common focus. These authors hy-pothesize that within the virus particles the 10genes are linked noncovalently by protein mol-ecules. It should be noted in this regard thatsome highly specific mechanism, possibly involv-ing association with protein molecules, must beresponsible for assembling 10 discrete species ofRNA into each virus particle (see below). Thismechanism must be very efficient, since underoptimal conditions the ratio of reovirus particlesto infectious units has been reported to approachunity (105). The pattern of transcription of in-dividual genes within reovirus cores, however, ismore in accord with the view that the individualgenes are unlinked, for they are transcribed notwith equal frequencies, but with frequencies thatare inversely proportional to their size (99). Thisis impossible if they are linked in a linear arrayif the catalytic sites of the transcriptase are fixedin space (which is most probably the case; seebelow), unless they are arranged in decreasingorder of transcription frequency, as are the genesof vesicular stomatitis virus. This is unlikely,since the relative transcription frequencies canbe made to change significantly by varying theconcentration of Mg2e and nucleoside triphos-phates. Thus, the genes are probably not linkedwithin the virus particle. Analysis by X-ray crys-tallography suggests well-ordered packaging ofthe genes with adjacent helices being packedparallel to one another (30).

PROTEINS OF REOVIRUSThe protein composition of reovirus particles

was first examined by Loh and Shatkin (56) andSmith et al. (102). The nomenclature of reovirus


mbr.asm

.org/D

ownloaded from



proteins was developed by the latter group ofworkers.

Reovirus particles are composed of nine spe-cies of proteins (Table 2). The outer capsid shellis composed of three species: the major species,ulC (a cleavage product of ,l) and a3, whichtogether make up over 60% of the mass of thereovirus particle, and the minor species al,which is present in reovirus particles to theextent of only about 24 molecules. The core,which comprises about one-third of the reovirusparticle's protein mass, contains six protein spe-cies: the major components Xl, X2, and a2, andthe minor components X3, ,il, and ,u2. The struc-tural, antigenic and enzymatic functions of theseproteins, as far as presently known, are as fol-lows.

(i) No function can yet be assigned to theminor protein species X3, ,Al, and ,2 (12 mole-cules or less per virus particle). All are presentin cores, but their location is not known: theymay be located on the inner surface of the coreshell, or they may be associated with the viralgenome, or they may occur free within the core.Conceivably, they may be responsible for someof the enzymatic activities exhibited by reoviruscores (see below), or they may serve to link thegenes (see above).

(ii) Protein X2 is the major, if not the sole,component of the 12 core projections/spikes(120) through which, as Gillies et al. showed in1971 (24), the ss transcripts of reovirus genesthat are synthesized in cores are liberated. Asdemonstrated via cross-linking studies (73), eachspike is a pentamer of A2, so that there are 60X2 molecules per virus particle. Recently Lee etal. (51) isolated and cloned a series of hybrid-omas that secrete immunoglobulin Gs (IgGs)directed against a variety of reovirus-coded pro-teins and found two that secrete IgG's directed

TABLE 2. Reovirus proteins

% in Approx no.Protein spe- MOl w vir- of mole- Location

cies Molwt iorn cules/viruscies ~~~~~~particleStructuralAl 155,000 15 105 CoreA2 150,000 11 90 CoreA3 135,000 52 '12 Core

p1 80,000 2 20 CorepUlC 72,000 35 550 Outer shellp2 70,000 s2 '12 Core

al 42,000 1 24 Outer shella2 38,000 7 200 Corea3 34,000 28 900 Outer shell

NonstructuralpNS 75,000aNS 36,000

against protein X2. Interestingly, these IgG's alsoneutralize reovirus infectivity, prevent hemag-glutination, and precipitate virus particles. Thisevidence indicates that, contrary to what hadbeen believed, the projections/spikes projectthrough the outer capsid shell to the surface ofthe virus particle (31). Models of reovirus archi-tecture will thus have to take into account thefact that there are 12 unique icosahedrally dis-tributed structural environments on the surfaceof reovirus particles. It is remarkable that theyhave not been seen.

(iii) Proteins X1 and a2 are the other twomajor core components. It is tempting to spec-ulate that the core shell is composed of capso-mers made up of ln Xl and 2n a2 molecules (justlike the outer reovirus capsid shell, which mayconsist of capsomers made up of ln AlC and 2na3 molecules; see below). Since protein Xl ismore readily iodinated than a2 (120), it may becloser to the core surface (assumng that bothproteins can be iodinated to similar degrees,other factors being equal, which has not yet beenshown).

(iv) Of proteins u1C and a3, protein ,llC is acleavage product of protein ,ul. It is the principalcomponent of the reovirus outer capsid shell, aswell as of reovirus particles. Antibodies againstit do not neutralize infectivity or prevent he-magglutination, although they may precipitatevirus particles (31).

Protein a3 is the other principal constituent ofthe outer capsid shell. It possesses strong affinityfor u1C, as is shown by the fact that well overhalf of the unassembled form of both of thesetwo proteins exists in the cytoplasm of infectedcells complexed with each other (32, 51). Pre-sumably, therefore, these proteins are also inti-mately associated with each other in virus par-ticles, in which they may exist in the form ofcapsomers with the constitution ln ,ulC:2n a3. Itshould be noted, however, that (i) chymotrypsinremoves a3 from reovirus particles before ,ulC isdegraded (37), (ii) after infection, reovirus par-ticles are converted to subviral particles fromwhich a3 is removed completely, whereas 1Closes only a polypeptide with a molecular weightof about 12,000, being converted to protein 8 (13,96, 98), and (iii) antibody to a3 neutralizes infec-tivity and possesses hemagglutination inhibitionactivity, whereas antibodies to ulC do not (51).Thus, although, closely associated, protein a3and,ulC can also react independently.

Protein a3 possesses the remarkable propertyof having strong affinity for ds RNA; also, thatportion of it that occurs in free form (that is, notcomplexed with ulC) in the cytoplasm of in-fected cells can be isolated by adsorption to andelution from polyinosinic-polycytidylic acid (32).

VOL. 45, 1981


mbr.asm

.org/D

ownloaded from


488 JOKLIK

The significance of this property, a remarkableone for a protein that is a component of theouter capsid shell, is not known. Perhaps a3 hassome function during reovirus morphogenesis.

(v) Protein al is only a minor component ofreovirus particles (only 24 molecules), but it isvery important, since it specifies how reovirusparticles interact with host cells and with thehost. Antibody to X2, but not antibodies to ,Al/y1C or a3, prevents antibody against al frombinding to al; this indicates that al is locatedclose to where the projections/spikes penetratethrough the outer capsid shell to the outer par-ticle surface (52). Presumably two al moleculesare associated with each of these structures.Protein al, which is able to adsorb to cells byitself, is the reovirus cell attachment protein(52). It is also the reovirus hemagglutinin (119)and elicits the formation ofneutralizing antibody(116). It is responsible for the development ofdelayed hypersensitivity (117) and for the gen-eration of suppressor T cells (19) and of cytolyticT lymphocytes (18). It also determines the ex-tent to which reovirus particles interact withmicrotubules (5) and specifies reovirus tissuetropism and virulence (115) (see below). It is byfar the most type specific of all reovirus proteins:the ability of antibodies against it to precipitatehomologous and heterologous al molecules (thatis, those of serotypes 3 and of 2 and 1, respec-tively), to neutralize virus, and to inhibit hemag-glutination is almost completely type specific(23, 31, 51). By contrast, antibodies against X2and a3 display hemagglutination inhibition ac-tivity that is mostly type specific, protein-precip-itating ability that is partially type specific, andneutralizing ability that is group specific (31, 51).It seems that the antibodies against X2 aremainly responsible for the group-specific neu-tralizing elements in antisera against reovirus,and that the antibodies against al are responsi-ble for the type specificity of the neutralizingactivity of such antisera.

INFORMATION CONTENT OF THE 10REOVIRUS GENES

By the late 1960s and early 1970s, it had beenestablished that the three size classes of reovirusgenes, namely, the L, M, and S genes (6), codefor the three size classes of reovirus proteins,namely, the A, ,, and a size class proteins (102).It was realized, however, that the largest L genedid not necessarily code for the largest X protein,and so on. Techniques for identifying the infor-mation content of individual genes became avail-able in the late 1970s. They came from two quitedifferent directions, but yielded identical an-swers. The first involved translation in an invitro protein-synthesizing system. Attempts to

translate reovirus mRNA's in vitro had beenmade since the early 1970s, when McDowell andJoklik (60) demonstrated that polyribosomesisolated from infected cells were capable of in-corporating labeled amino acids into proteinswith the electrophoretic mobilities of authenticreovirus proteins. Shortly thereafter, McDowellet al. (61) devised cell-free systems from severalmammalian cells, including HeLa cells, L cells,and Ehrlich ascites tumor cells, that were capa-ble of translating reovirus mRNA species tran-scribed in vitro by reovirus cores into a, ,A, andeven A size class proteins, and Both et al. (12)devised a similar system from wheat germ thatwas capable of translating all 10 species of reo-virus mRNA. In fact, it was in this study thatthe existence of the two minor protein speciesX3 and ,u2 was first demonstrated (not only arethese two proteins present in virus particles andin extracts of infected cells in very small amountsonly, but their electrophoretic migration ratesare also very close to those of the major capsidprotein species A2 and ,uC, so that they areusually obscured by them). The straightforwardway of determining the information content ofreovirus genes by this approach would be toisolate the 10 species of mRNA transcribed byreovirus cores in vitro, to hybridize them to thevarious genes to determine from which each wastranscribed, and to translate them individuallyin an in vitro protein-synthesizing system. It isdifficult technically, however, to isolate suffi-cient quantities of the larger mRNA species.The approach therefore adopted by McCrae andJoklik (59) was to isolate the 10 genes them-selves, to denature them at 50°C in 90% dimethylsulfoxide, and then to dilute them into a protein-synthesizing system prepared from wheat germ.At the RNA concentration and ionic strengthused in these incubated mixtures, reannealing ofthe plus and minus strands was sufficiently slowto permit the translation of even I size classmRNA molecules into complete protein mole-cules. Identification of proteins was achieved notonly by comparison of the electrophoretic mo-bilities of the translation products with those ofauthentic proteins, but also by V8 peptide map-ping. It was found that there was no absolutecorrespondence between relative gene and pro-tein size as judged by electrophoretic migrationrates. The gene-protein assignments that werefound are summarized in Fig. 3.Mustoe et al. (67) studied the same problem

by a completely different method. As describedabove, there are three serotypes of mammalianreovirus, serotypes 1, 2, and 3. The sizes of thegenes of virus strains belonging to these threeserotypes, and of the proteins encoded by them,differ slightly but detectably. As would be ex-

MICROBIOL. REV.


mbr.asm

.org/D

ownloaded from



MW. LI -xi 14MW2.5-2.7 .)1'1-- X 14-15.5

X106 L13 --)3 X

M.W. rMl1.2-1.4 M2x106 M3

CSIM.W. S2_s,s0.6-S

X106 S

x2 W- Lic*-

_ ~~~~~~~~~2x104_ * r ~~~- 0'3

FIG. 3. Gene-protein assignments for reovirus se-rotype 3 (strain Dearing) (59).

pected, the genes of these three serotypes un-

dergo extensive reassortment in mixedly infectedcells. Thus, by examining polyacrylamide gelelectrophoresis profiles of the RNA and proteinspecies of cloned recombinants after pairwisemixed infection, it is possible to determine whichprotein changes accompany specific genechanges. The gene-protein assignments made bythis method are exactly the same as thoseyielded by the direct in vitro translation methoddescribed above.

ENZYMES IN REOVIRUSIn 1967, Kates and McAuslan (39) discovered

the first synthetic virus-coded virus-associatedenzyme, deoxyribonucleic acid (DNA)-depend-ent RNA polymerase of vaccinia virus. A yearlater, Borsa and Graham (10) and Shatkin andSipe (91, 92) independently discovered an en-

zyme with a similar function, namely, to tran-scribe the viral genome into mRNA, in reovirusparticles or, more acurately, in reovirus cores:

this is a ds -. ss RNA polymerase, or "transcrip-tase." Shortly thereafter, Kapuler et al. (38) andBorsa et al. (11) demonstrated the presence ofnucleoside triphosphate phosphohydrolase ac-

tivity in reovirus cores, and in 1975 Furuichi andShatkin and their collaborators discovered thatterminal guanylyltransferase and the enzymesthat methylate the cap G and the ribose of theoriginal 5'-terminal residue are also present (21,89). Together these enzymes constitute a mech-anism that enables reovirus cores to transcribethe 10 viral genes into 10 capped species ofmRNA which are extruded through the 12 pro-

jections/spikes located on cores (24).The five mRNA-synthesizing enzyme activi-

ties are not expressed by intact reovirus parti-cles: they are expressed only when their outercapsid shell is partially or completely removed.Two types of particles in which these enzymesare expressed are the subviral particles (SVP),which are formed in infected cells and have lost

all outer shell proteins except a 60,000-daltonportion of protein ,ulC (see below), and the reo-virus cores, which have lost all outer capsidshell. Presumably, loss of integrity of the outercapsid shell causes structural changes in the coreas a result of which the transcriptase and thecapping enzymes are activated. Interestingly,the process is reversible, for addition of a3 toSVP causes the enzymes to become latent (4).Nothing is known concerning the nature of thechanges that are induced in cores as a result ofwhich they become able to transcribe. One pos-sibility is that the A2 clusters that make up theprojections/spikes are closed in virus particlesand open in cores; alternatively, the crucial fac-tor may not be the configuration of the spikes,but rather changes in the configuration of thetranscriptase and associated enzymes which maybe locked into an inactive state if the outercapsid shell is intact.Nothing is known concerning the nature of

the transcriptase or of the four enzymes con-cerned with capping transcripts, since all enzymeactivity is lost as soon as cores are disrupted. Itis therefore thought that the enzymes are com-ponents of the core shell and that the genesmove past transcriptase catalytic sites fixed onthe interior side of the core shell, perhaps at thebase of the spikes, so that the ds templatesremain within the cores, whereas the transcriptsare fed into and through the spikes. If this iscorrect, the transcriptase catalytic site may com-prise sequences on both X2 and on neighboringprotein molecules (Xl or a2), whereas the cap-ping functions would be on X2 in the projection/spike channels. Indeed, when cores transcribe dsRNA in the presence of labeled pyridoxal phos-phate, which is hypothesized to react with theactive site of the enzyme, both Xl and X2 becomelabeled (63). It should be pointed out, however,that this is the only evidence that the transcrip-tase is located on the inner capsid shell. Asdiscussed above, no functions have yet beenassigned to the three major proteins, X3, ,ul, and,u2, that are also associated with cores, and theseproteins, either free in the interior of the virusparticle or attached to the inner core shell, maythemselves possess transcriptase or capping ac-tivity or both.

TRANSCRIPTION OF REOVIRUSRIBONUCLEIC ACID

The transcriptase activity of reovirus cores isvery stable; the enzyme is active for over 48 h at480C, which is its optimal temperature. Thetranscription that it catalyzes is totally asym-metric-strands with only one polarity (plus,that is, translatable) being synthesized-andconservative (that is, the first plus strand to beformed is already a transcript, not the plus

VOL. 45, 1981


mbr.asm

.org/D

ownloaded from


490 JOKLIK

strand of the ds molecule being transcribed). Inthe presence of optimal Mg2e concentrations, allgenes are transcribed at the same rate, i.e., fourand two times as many s and m class transcripts,respectively, are synthesized as I class tran-scripts (99). In other words, the frequency oftranscription initiation is not the same for eachgene, but is inversely proportional to gene size,which argues against the 10 genes being tran-scribed as a linked complex (see argumentsabove).The transcripts that are formed during the

early part of the reovirus multiplication cycle orin vitro in the presence ofS-adenosylmethionineare capped and methylated; that is, the structureof their 5' termini is m7G(5')ppp(5')GmpCp ...(21, 89). However, neither capping nor methyl-ation is tightly coupled to transcription. S-ad-enosylmethionine does not stimulate transcrip-tion, and under appropriate conditions not onlyunmethylated but also uncapped transcripts arereadily formed (22). Interestingly, S-adenosyl-methionine stimulates the transcriptase of orbi-viruses 2-fold (109) and that of cytoplasmic pol-yhedrosis viruses by up to 70-fold (20). Themechanism of this stimulation is thought to belowering of the Km of the initiating nucleotide atthe promoter site.Although not influenced by S-adenosylme-

thionine, the initiation of transcription of reovi-rus genes appears to be a complex process, forYamakawa et al. (M. Yamakawa, Y. Furuichi,K. Nakashima, A. J. LaFiandra, and A. J. Shat-kin, J. Biol. Chem., in press) have recently foundthat transcription aborts in more than 50% ofinitiation events before transcripts are morethan five residues long. This finding relates to aseries of observations made about 10 years agoto the effect that reovirus particles contain notonly ds RNA, but also ss RNA. In fact, about25% of the total RNA in reovirus particles is ssRNA (6, 92). This RNA is in the form of shortmolecules which are not breakdown productssince they contain PPP (as well as smalleramounts of PP and P) at their 5' ends. Thesemolecules fall into two classes (8, 69, 108). Aboutone-third contain only adenosine; these are theoligoadenylates, and there are about 900 ofthemin each reovirus particle. The remainder fall intothe following series: GCOH; GCUOH; GCUAOH;GCUA(U)14UOH; and GCUA(A)lj,AoH: theseare the 5'-G-terminated oligonucleotides. Thesequence of these 5'-G-terminated oligonucleo-tides is very interesting, since all reovirus plusstrands share the sequence GCUA at their 5'termini (41, 54; J. K.-K. Li, J. D. Keene, P. P.Scheible, R. Chmelo, J. Antzak, and W. K. Jok-lik, unpublished data). They are therefore

clearly products of abortive transcription. Thefact that they are sealed into reovirus particleslends support to the view (see above) that theprojections/spikes through which transcripts arenormally released are sealed in reovirus particlesand open in reovirus cores. (Incidentally, thishypothesis also implies that nucleoside triphos-phates are unable to enter intact reovirus parti-cles, thereby rendering transcription impossible,irrespective ofwhether the transcriptase in virusparticles exists in a configurationally active orinactive state). Presumably these oligonucleo-tides are sealed into reovirus particles during thefinal stage of morphogenesis, when transcriptionbecomes inhibited through the association of u3to the penultimate form of immature virus par-ticles (see above). Either the 5'-G-terminatedoligonucleotides may then represent the nor-mally occurring abortive transcripts demon-strated by Yamakawa et al. (in press) whichbecome trapped, or transcription elongation mayat that time be inhibited before transcriptioninitiation, the abortion rate then rising briefly to100% (8). Furthermore, the synthesis of oligoad-enylates may then be an expression of a partialactivity of the transcriptase in the process ofbeing inactivated: its Km for adenosine triphos-phate may then greatly exceed that for othernucleoside triphosphates, which would result inits catalyzing the template-independent polym-erization of adenosine triphosphate to short oli-goadenylates. Indeed, Silverstein et al. (95)found that oligoadenylates are formed withinnascent virus particles during the final stages ofmorphogenesis and suggested that the oligoad-enylate polymerase activity may be an alterna-tive activity of the viral transcriptase regulatedby outer capsid proteins. Furthermore, Johnsonet al. (36) found that late temperature-sensitivemutants of reovirus that synthesize noninfec-tious and therefore presumably defective parti-cles at nonpermissive temperatures do not syn-thesize oligoadenylates at nonpermissive tem-peratures, which again supports the conclusionthat they are synthesized during the final stagesof virus maturation.

SEQUENCES OF REOVIRUS GENESAND MESSENGER RIBONUCLEIC

ACIDSAttempts to sequence reovirus genes started

soon after it was realized that they were individ-ual RNA molecules, the purpose being to testthe possibility that the association of sets ofgenes during morphogenesis involved recogni-tion of complementary sequences at their ends.Although at that time, in the early 1970s, meth-ods were available for examining only the ter-

MICROBIOL. REV.


mbr.asm

.org/D

ownloaded from



minal two or three residues, it was soon shownthat all reovirus genes had identical terminaldinucleotide pairs, which ruled out the possibil-ity of terminal complementarity. The first seri-ous attempt to sequence a reovirus RNA mole-cule was when Nichols et al. (70) found thatunder certain limiting conditions reovirus coressynthesize only a single species of mRNA, thes4 species (15), and sequenced its first 25 5'-terminal residues. This work was followed bythe studies of Kozak and Shatkin, who, in aseries of elegant studies, sequenced the regionsin 6 of the 10 species of mRNA that bound toand were protected from digestion with RNaseby wheat germ 40S ribosomal subunits and 80Sribosomes (41, 44-46). In all cases the 40S sub-units protected sequences 50 to 60 residues longthat included the 5' terminus as well as theinitiation codon. The sequences protected byintact 80S ribosomes were subsets of these se-quences that were only about one-half as longand were centered around the initiation codons(41). Figure 4 shows the six sequences. Interest-ing features are: (i) all six 5' termini start withGCUA; (ii) the distance between the 5' terminusand the first initiation codon is short, only 15 to33 residues; (iii) the sequences between the 5'termini and the initiation codons are completelydifferent; (iv) these sequences lack secondarystructure features such as hairpins; and (v) eachmRNA possesses near its 5' terminus a Shineand Dalgarno (94) sequence, that is, a sequencethat is complementary to some sequence nearthe 3' end of the 18S ribosomal RNA.

Recently Li et al. (54) sequenced the 3' endsof both strands of the Si genes of serotypes 1, 2,and 3. Since the plus and minus strands ofreovirus genes are exactly the same length andthe plus strands of reovirus genes are identicalwith reovirus mRNA molecules (55, 68), this

ml

analysis provided the sequences at both ends ofthe sl mRNA molecules of the three reovirusserotypes. As pointed out above, the Si genecodes for protein al, the most type specific of allreovirus proteins. Examination of the sequencesof the three Si genes should therefore revealmost clearly those features that are conservedand those that may be varied. In fact, one wouldexpect few similarities among their coding se-quences, but significant similarities in their ter-minal sequences, which should be able to rec-ognize RNA polymerases, ribosomes, encapsi-dation signals, and, perhaps, sequences on otherreovirus genes. Li et al. (54) did, in fact, find thatall three Si genes possess at least six and insome cases nine identical base pairs at the endsthat contain the 5' ends of the plus strains (Fig.5); that the initiation codons are 14 or 15 basepairs downstream and that the sequences be-tween the common terminal ones and the initi-ation codons, for even these three homologousgenes, are completely different; that the threecoding sequences are quite different (althoughthe ratios of neutral to charged to hydrophobicamino acids are similar); and that at the otherend of the genes there is substantial homology,including a terminal block of five base pairs andthree other blocks of homology from five toseven base pairs long separated by regions ofvarying length. Similar studies on the three L3,M3, and S2 genes have shown that all possessthe same four and five respectively terminal basepairs that are also present on Si genes; that thesequences between the 5' termini of plus strandsand the initiation codons of homologous genesare identical (rather than completely different,as in Si genes); that the coding sequences ofhomologous genes are identical for the first 10to 20 codons; and that at the other ends the L3,M3, and S2 genes of serotypes 1, 2, and 3 are

15C7 (pG CUA UUC, GCU GUC -AUG GCU UAC AUC GCA C-

31M2 u7G Gs C UM UCU GCU GAC CGU UAC UCU GCA MG AUG GGG MC GCU CWU CCU AUC G-ppp

13

S2

20u7G GM CU MA GUG ACC GUG GUC AUG GCU UCA UUC MG GGA WUC tJCC G-

ppp

207GpppGM CU AUU CGC UGG UCA GUU AUG GCU CCC UGC GCG UUC CUA UUC MG-

29S3 7Gppp CU MA GUC ACG CCU GUC GUC GUC ACU AUG GCU UCC UCA CUC AG-

33S4 *7GpppG CUA UUU UGC CUC UUC CCA GAC GUU GUC GCA AUG GAG GUG UGC UUG CCC MC G-

FIG. 4. 5'-Terminal sequences ofsix species ofmRNA ofreovirus serotype 3 (strain Dearing). The sequencesshown are those protected by the 40S ribosomal subunits of wheat germ (41). The gene assignments are fromDarzynkiewicz and Shatkin (15).

VOL. 45, 1981


mbr.asm

.org/D

ownloaded from


492 JOKLIK

A

slYPE

2

3

3

GCUAUU

GCUAUU

5,C U A U L

5,

C UAUU

MICROBIOL. REV.

GE10 20 30 40 S0 60 70

CGC:GCCUAU6GAU6CAUCUCUCAUUACAGA6AUACGGAAAAUA6UACUCCAACUAUCUGUAUCA AGCAAUSAV l.A OW U.U ILE 118 61.1 ILE AN LYS IL E.WL U 6.1KM U OM VAL Wr SW AN

GC ACUCAU6UCGGAUCUA6UGCA6CUCAUAAGAA6666AAUCUUACU6UUAACUGGGAAU666AGAAUCA SAP L1 WL " 11 n s s . ILE UU UU1.8 vwAl aL o

GGUCXGAU66AUCCUC6CCUAC6U6AA6AA6UA6UAC66CU6AUAAUC6CAUUAAC6A6UGAUAAUG6AG 5

P M IM LUUMB6 oU au WL v. An LAU ILE ILE MA LO 11 W AV 6LY

10 20 30 40 50 60 70

C6C'UGGUCAGUUAUGGCU'CGCGCU6CSUUCCUAUUCAAGACUGUU666UUU6'U6GGUCUGCAAAAUGUGCMA MW MA M IE LO PE LYi 11 L 61LY MP LY LY LO 6. AU VM

S2

B

GENE SEYP70 10 S0 40 30 20 10 3,

Sl GGGCCUCGUGGACGAUCAUGUACUCAU6CAAU6UGA UiAAUCUAGCGiAA C6GCAC A6GUCAA CAUC-G 1

S.X....... CAUUAUGUAGCCAUA6UCU6G AUCGGUGCUCACUCGGCAC GUGGCGACIUCAUCI- 2

S5 GGUGCUGCUAGACCGUUUCGUACCCGCGUAGUUUCiC UXAGACtCACC66C AC CAUIUC -IN 3

S2 CGAUUCGA6CCCUUAUCU6AUCW6UCUU6W6M UA66GUCCCCCCAC.ACCCCUCACG.ACUG CCACACA,UCAuc'- 3-CA'G' ACACACAUAU3!............

FIG. 5. Sequences ofthe 5' termini (A) and 3' termini (B) oftheplus strands, that is, the mRNA 's, ofthe Sigenes ofreoviru serotypes 1, 2, and 3 and of the S2 gene of reovirus serotype 3 (54).

also much more similar than the three Si genes(Joklik et al., unpublished data).The possible secondary structure of reovirus

mRNA's is interesting; Fig. 6 shows a typicalexample, the sl mRNA of serotype 2 (54). Thereis little tendency to form hairpin loops withinthe terminal 70 to 100 residues at either end; thesole exception is a hairpin loop near the 3' endthat contains three in-phase termination codons.Some mRNA species such as the s2 mRNA ofserotype 3, do not even possess this loop. Thereare complementary sequences 6 to 10 residueslong near the 5' and 3' ends of all reovirusmRNA's that have been examined, the free en-ergy of association of which is quite high (-10to -20 kcal) (see also Li et al. [55]). There is alsoa six-residue-long sequence near the 5' end ofserotype 2 sl mRNA that is complementary toa sequence near the 3' end of 18S ribosomalRNA (a Shine and Dalgarno sequence) and thatoverlaps almost entirely the sequence that iscomplementary to the 3'-terminal sequences.Each reovirus mRNA species possesses a differ-

ent Shine and Dalgarno sequence. For someRNA species, such as serotype 3 sl or s3 mRNA,it is no more than four residues long and prob-ably very unstable; for some RNA species, suchas serotypes 3 s2 niRNA (55), the Shine andDalgarno sequence does not overlap at all withthe sequence that is complementary to the 3'-terminal sequence. Thus it seems that sequencesimmediately upstream from the initiation codonof reovirus mRNA's can associate either withsequences near their own 3' terminus or with 3'-terminal 18S ribosomal RNA sequences. Con-ceivably, such associations could regulate thefrequency of translation of the individual speciesof reovirus mRNA. As will be discussed below,some reovirus mRNA species are translatedmuch more (up to 50-fold) frequently than oth-ers. The system of 10 species of reovirus mRNAmolecules, which are all translated in the sameenvironment at the same time and the relativeconcentrations ofwhich within infected cells canbe measured accurately by hybridization to dsRNA, is uniquely suited for elucidating the na-

7


mbr.asm

.org/D

ownloaded from



G C, OH'6' 'U''.U. CA: U'U 'C'

3 A-U-U-A-C-U-A-G-6-A-A-U C

18s RIBOSOMAL RA 4 ..C-G

6 * *C- - 15.0 CAL

5# . . . C-C" C-UU-A

C

A

INITIATION COD01 U

U

C

6

6

C

A

6

6

C

U

C

A A

U C

C U

U C

A

GU

6C

A6C

U

-C

AU

AA

GAAGG6AGAUCU ......

6

U

GGC

U

A 13.0 KCAL

AGUX6CX C

66C CAUA

U 3 TER.MNATIOmA CODWIS IN PHASEU

AC

FIG. 6. Possible partial secondary structure of thesl mRNA ofreovirus serot)pe 2 (54).

ture of the controls that regulate frequency oftranslation; indeed, elucidation of these controlsis one of the principal reasons for sequencing thereovirus genes.

TRANSLATION OF REOVIRUSMESSENGER RIBONUCLEIC ACIDSThe translation of reovirus mRNA's is con-

trolled at several levels. First, there is controlmediated by the 5'-terminal cap. The reovirusmRNA's that are transcribed during the first 4h or so are capped and methylated: they aretranslated efficiently in extracts of uninfectedcells, but not in extracts of cells infected withreovirus (101), whereas uncapped reovirusmRNA molecules are translated poorly in ex-tracts of uninfected cells, but efficiently in ex-

tracts of infected cells. Not surprisingly, theimmature reovirus particles, which synthesizeabout 95% of the total viral mRNA that isformed in infected cells (see below), synthesizeuncapped mRNA's: their guanylyltransferaseand methylase activities are latent (100). Itseems, therefore, that some of the factors of the

protein-synthesizing mechanism of the cell un-dergo profound changes during the course of thereovirus multiplication cycle, but we do not yetknow what these factors are or how they aremodified.

Second, there is profound control over thefrequency with which the 10 species of reovirusmRNA are translated, some species being trans-lated much more frequently than others (Table3). Interestingly, the relative efficiencies oftranslation seem to be the same both early dur-ing the multiplication cycle, when the mRNA'sare capped, and late, when they are not capped.It is likely that the major, if not the sole, mech-anism for regulating the translation frequenciesof these 10 species of mRNA, which exist in thesame environment at the same time, resides intheir sequence content. Experiments to deter-mine the order of the 10 sequences ofmRNA arecurrently underway via a two-pronged approachinvolving sequencing the 3' termini of the minusstrands of reovirus genes, on the one hand, andcloning reovirus genes and sequencing theirDNA transcripts, on the other (Joklik et al.,unpublished data). Elucidation of the features ofmRNA sequences that specify efficiency oftranslation could clearly have far-reaching sig-nificance. It should be pointed out in this regardthat differences in translation frequencies thatexist in vivo also exist in vitro: thus, the four sspecies of mRNA are translated in wheat germprotein-synthesizing systems in vitro with effi-ciencies that differ by as much as sevenfold (53),the same factor that is observed in vivo (Table3).The initial events in the translation ofreovirus

mRNA's have been examined in some detail byKozak and Shatkin. The first step is the bindingof 40S ribosomal subunits at the 5' terminus (41,44-46). Since there is no binding intemally, even

TABLE 3. Approximate relative frequencies oftranscription and translation of the 10 reovirus

genesTranslation fre-

Gene Transcription Translation quency/tran-frequency frequency scription fre-quency

Li 0.05 0.03 0.6L2 0.05 0.15 3L3 0.05 0.1 2

Ml 0.15 0.03 0.2M2 0.3 1.0 3.3M3 0.5 0.5 1

Si 0.5 0.05 0.1S2 0.5 0.2 0.4S3 1.0 0.3 0.3S4 1.0 0.7 0.7

VOL. 45, 1981


mbr.asm

.org/D

ownloaded from


494 JOKLIK

when the RNA is denatured to expose putativemasked initiation codons (43), it seems that thebinding is specific for 5' termini, the affinity forcapped RNAs being greater than that for un-capped ones which, however, do bind to a sig-nificant extent (47). The 40S ribosomal subunitscover a surprisingly large stretch of mRNA,namely, 50 to 60 residues. After they havebound, they move along the RNA molecule untilthey reach the first initiation codon, where theypause to combine with 60S ribosomal subunits,methionyl transfer mRNA, and whatever factorsare necessary to initiate protein synthesis (48).The stretches ofRNA covered by 80S ribosomesare shorter than those covered by 40S subunitsand are subsets of them; the sequences coveredby 80S ribosomes are 25 to 35 residues long andare centered about the initiation codon (41).Pausing at the first AUG is a critical event ofthe translation process; it depends on secondaryRNA structure, since 40S ribosomal subunits donot pause there-and do not combine with 60Sribosomal subunits-if the secondary structureof mRNA is weakened or destroyed (42, 43) orin the presence of inhibitors of protein synthesissuch as edeine. It will be fascinating to determinewhich of these various stages regulates transla-tion frequency (i.e., efficiency).

ESSENTIAL FEATURES OF THEREOVIRUS MULTIPLICATION CYCLEFigure 7 shows the essential features of the

reovirus multiplication cycle. Reovirus particlesadsorb to specific receptors that are the samefor all three serotypes (52). The reovirus attach-ment organ is protein ol, which, as pointed outabove, is located pairwise in 12 icosahedrallydistributed positions on the virus particle sur-

ReoveusPorticle

,ISVP

EorlyMessenger

RNA

- t

IPProteins

face; this protein can adsorb to cells by itself andcompete with virus particles for attachmentsites. Presumably the initial attachment of reo-virus particles to host cells occurs via two almolecules.The reovirus particle is then taken up into

cells in phagocytic vacuoles that fuse with lyso-somes (97) within which the reovirus outer cap-sid shell is partially digested: proteins al and a3are removed entirely, and a 12,000-dalton frag-ment is removed from protein ,ulC, leaving a60,000-dalton protein that is sometimes referredto as protein 8 (13,96,98). This partial disruptionof the outer capsid shell, which generates the so-called SVPs, activates the transcriptase that ispresent within them, and once the SVP havebeen released from the lysosomes into the cyto-plasm, the reovirus genes begin to be tran-scribed.When reovirus genes are transcribed by cores

in vitro, in the presence of 5 mM Mg2+, they aretranscribed with a frequency that is inverselyproportional to their size (99) (see above). Thefrequency of transcription of the 10 genes in vivodiffers somewhat from this pattem in that theamount of s transcripts usually exceeds theamount of m transcripts, which in turn exceedsthe amount of I transcripts that are formed; thisis most probably due to the fact that the intra-cellular Mg2e concentration is much lower than5 mM. In addition, it seems that there is someregulation over the rate of transcription of indi-vidual reovirus genes during the early stages ofthe multiplication cycle (113). At that time fourspecies of mRNA are synthesized predomi-nantly: 11, m3, s3, and s4. This pattern soonchanges to that described above, but persists inthe presence of cycloheximide (50) and also oc-

Immature RewusVirus Particle

Partilces

Lots+ essengr/ - RNA

Prote,ns

FIG. 7. Schematic representation of the reovirus multiplication cycle.

MICROBIOL. REV.


mbr.asm

.org/D

ownloaded from



curs under nonpermissive conditions, such as inmammalian cells infected with avian reovirus(104). Furthermore,-Shatkin and LaFiandra (90)reported that whereas infectious SVPs (whichare produced by digesting reovirus particles withchymotrypsin in vitro in the presence of 0.15 MNaCl and which appear to be identical to theSVPs produced in vivo) transcribe all 10 speciesof mRNA in vitro, they transcribe only species11, m3, s3, and s4 in cells treated with cyclohex-imide. These data suggest that the transcriptionof reovirus genes may be regulated early duringthe multiplication cycle, possibly by a host cellfactor the function of which is neutralized byone or more of the proteins coded by genes, Li,M3, S3, and S4 (namely, the very minor speciesX3, the outer capsid shell component a3 [whichhas strong affinity for ds RNA], and the twononstructural proteins yNS and aNS). It is dif-ficult to see, however, how transcription withinSVPs can be influenced from the outside whennot even pancreatic RNase, a very small protein,can enter them. Clearly, more work remains tobe done on the mechanism of this mysteriousphenomenon.The 10mRNA's are translated into 10 primary

translation products. Sizable amounts of theseproteins are present in the cytoplasm of infectedcells (32). They migrate in density gradientsaccording to their monomeric molecular weightsexcept proteins a3 and ,ulC; about 80% of eachof these latter proteins sediments in the form ofa complex the molecular composition of whichappears to be 1 l1C:l a3. The existence of thesecomplexes is also confirmed by treating infectedcell extracts with monoclonal antibodies, whenantibodies directed against,lC and a3 precipi-tate not only the free form of each protein butalso the complex (51). Interestingly, one anti-body precipitates the complex but neither of thefree proteins, which indicates that the antigenicsite against which this antibody is directed com-prises amino acid sequences of both ,ulC and a3.

Several of the primary reovirus translationproducts are processed to cleavage products.Interestingly enough, in no case is cleavage com-plete; that is, in all cases some of the primarytranslation product remains uncleaved, whichsuggests that both it as well as the cleavageproduct may have a function during reovirusmultiplication. One of these pairs of proteins hasalready been discussed above: it comprses pro-teins Al and AlC, the products of gene M2. Theother two pairs are ,NS and ANSC, and X2 andX2C (51). These latter two cleavage productswere discovered when monoclonal IgG's againstyNS and X2, respectively, were added to extractsof cells infected with reovirus. In the case of

,uNS, a second protein some 5,000 daltonssmaller, which is present in infected cells inroughly the same amount asu,NS, was also pre-cipitated, and in the case of X2, a small amount,corresponding to about 10% of the amount of X2,of a protein with a molecular weight of about120,000 was also precipitated. Thus most (pl),about one-half (uNS), or only a small amount(A2) of the primary gene product is processed toa cleavage product. The function of,llC is known(it is the principal component of the reovirusouter capsid shell); that of uNSC and X2C is not.

Reovirus MorphogenesisSome 4 h after infection, presumably when

sufficient amounts of viral transcripts and pro-teins have been formed, morphogenesis com-mences. The earliest immature virus particlescomprise virus-coded proteins among which Xl,X2, g2, and ,ulC are prominent (64) and containplus-stranded transcripts in RNase-sensitiveform (1). Within these particles, the plus strandsare transcribed into minus strands once and onceonly, and the minus strands remain associatedwith the plus strands, thereby forming the prog-eny ds RNA molecules. The nature of the mech-anism that specifies that each progeny virusparticle receives 1 of each 10 RNA molecules isnot known and is one of the principal unsolvedproblems of reovirology. The reason is that theimmature reovirus particles are both multiple innature and unstable, so that it is difficult topurify them. It has so far been impossible toisolate individual species ofsuch particles and todetermine their protein composition, the natureof the RNA that they contain, and the enzymeactivities that they possess. All that is known isthat (i) assortment ofRNA molecules occurs atthe stage of ss RNA (1, 84, 85, 122); (ii) theearliest class (or classes) of immature reovirusparticles possesses an ss -. ds RNA polymerase(a replicase), although it is not clear whether allRNA molecules are transcribed simultaneously(in fact, available evidence suggests that tran-scription of the s, m, and I class plus strands intominus strands may proceed sequentially [121]);(iii) morphogenesis proceeds through the se-quential exchange or addition of virus-codedproteins (124); (iv) immature reovirus particlesthat already contain ds RNA transcribe it intoss transcripts (mRNA) (65, 83) (in fact, this iswhere reovirus "multiplies," about 95% of viralmRNA formed in infected cells being synthe-sized by progeny immature reovirus particles);(v) morphogenesis probably proceeds through astage of corelike particles (84) (particles resem-bling cores also accumulate in cells infected withcertain classes of temperature-sensitive mutants

VOL. 45, 1981


mbr.asm

.org/D

ownloaded from


496 JOKLIK

at nonpermissive temperatures [64]); and (vi)the final stage of morphogenesis appears to beaddition of protein a3, since if infected cells are

pulse-labeled briefly with radioactively labeledamino acids, the virus that can be isolated fromthem is labeled only in protein a3 (P. W. K. Leeand W. K. Joklik, unpublished data). In fact, a3associates even with parental SVPs, which arethus liberated together with newly formed prog-eny as particles with an intact inner shell andwith an outer shell that consists of proteins al,8, and a3 (13). It should be noted that theaddition of a3 is the event that triggers theinhibition of the transcriptase and thus causes itto become latent (4) (and, incidentally, maycause it to synthesize oligoadenylates and shorttranscripts for a brief period oftime which wouldthen be sealed into the maturing virus particles;see above).Thus, the salient feature of the reovirus mul-

tiplication cycle is that the assortment of RNAsegments into progeny virus particles occurs atthe level of ss RNA, which provides the physicaland genetic link between parent and progenyvirus particles; ds RNA never exists in free,unencapsidated form within infected cells (25).The nature of the replication of ds RNA is thusquite different from that of ds DNA: for ds RNA,its plus and minus strands are synthesized atvery different times, and have different fates anddifferent functions. It is also clear that recom-binants among reovirus strains are formed at thelevel of ss RNA, not ds RNA, for it seems thatno mechanism exists for "mixing" ds RNA mol-ecules.Nothing is known concerning the mechanism

that specifies that each progeny virus particle isapportioned a set of 10 unique RNA molecules.One approach to investigating this problem is touse antisera to individual reovirus proteins toisolate increasingly complex particles on themorphogenetic pathway. Needless to say, mor-phogenesis would have to be synchronized forthis type of study, but this could be achieved byusing temperature-sensitive mutants blocked atthe first stage of morphogenesis at nonpermis-sive temperatures (such as ts447) (58) and thenshifting down. As for the antibodies, a set of 19monoclonal IgG's directed against 7 of the 10reovirus-coded proteins has recently been iso-lated and characterized (51) and could be usedfor this purpose.

Reovirus Deletion MutantsAlthough the mechanism that apportions

RNA molecules to nascent reovirus progencyparticles is efficient (since most reovirus parti-cles contain 10 genes), it is not infallible, andseveral examples of reovirus particles with less

than 10 genes are known. For example, repeatedcultivation of the Dearing strain of reovirus se-

rotype 3 under conditions of high multiplicityresults, within seven to eight passages, in theproduction of noninfectious virus particles thatlack gene Li (71, 87). Several temperature-sen-sitive mutants of reovirus lose genes even morerapidly, the most extreme example being mutantts447, a mutant totally (that is, more than 99%)defective in the ability to form ds RNA-synthe-sizing particles (i.e., the earliest immature virusparticles) at nonpermissive temperatures, yieldsof which comprise even at permissive tempera-tures equal numbers of virus particles that con-tain and lack gene L3 (87). When this mutant iscultivated repeatedly at high multiplicity, ityields, within four passages, particles that lackgenes Li, L3, and Mi (2, 87). Virus particleslacking genes are also associated with thechronic (persistent) intracerebral infections thatresult in about 50% of 2-day-old rats inoculatedsubcutaneously with wild-type reovirus or someof its temperature-sensitive mutants (103).

TEMPERATURE-SENSITIVE MUTANTSOF REOVIRUS

Fields and Joklik (17) isolated a series of tem-perature-sensitive mutants that fell into six com-plementation and gene reassortment groups.During the next several years, the lesions inmany of these mutants were assigned to specificgenes by the use of both biochemical and genetictechniques; the four remaining mutant groupswere isolated by Ramig and Fields in 1979 (74).Their genetic and phenotypic properties aresummarized in Table 4.Reovirus exhibits the phenomenon of extra-

genic suppression to a remarkable extent. In1977, Ramig et al. (76) found that a spontaneousrevertant of a temperature-sensitive mutant of

TABLE 4. Reovirus temperature-sensitive mutantsa

Type Lesion in gene'mentgroup__A Late L2 (86) or M2 (66)B Late L2 (66)C Early S2 (58, 75)D Early M2 (35) or Li (75)E Very early S3 (34, 75)F Late ?G Late S4 (66)H ?I ?J? ?

a Mutant groups were isolated as foliows: A to E, byFields and Joklik (17), F and G by Cross and Fields(14); and H to J by Ramig and Fields (74).

b References providing evidence concerning the lo-cations of the lesions are given in parentheses.

MICROBIOL. REV.


mbr.asm

.org/D

ownloaded from



reovirus (the group A mutant ts2Wl) did notresult from back mutation to wild type, butrather from the presence of a second mutation(a "suppressor" mutation) elsewhere in its ge-nome. This was shown by backcrossing the re-vertant to wild-type virus and demonstratingthat about one-half of the progency still har-bored the original temperature-sensitive lesion;the backcross served to separate the tempera-ture-sensitive from the suppressor mutation.This phenomenon of extragenic suppression isvery common in reovirus. Thus Ramig andFields (74) found that 25 of 28 spontaneousrevertants of a variety of temperature-sensitivemutants belonging to all mutant groups stillcontained temperature-sensitive lesions, andthey were even able to isolate new suppressedtemperature-sensitive lesions belonging to newmutant groups from the pseudorevertants,which may have acted as temperature-sensitivesuppressors. The significance of this work stemsfrom the fact that the suppressor mutationsprobably act by producing compensating alter-ations in some other protein that interacts phys-ically with the protein coded by the gene withthe temperature-sensitive lesion, and one wouldtherefore expect the existence ofprecise pairwisepatterns of temperature-sensitive suppressor le-sions. Their elucidation should reveal the exist-ence of interactions among reovirus-coded pro-teins that may provide important clues for reo-virus morphogenesis.

INTERACTION OF REOVIRUS WITHTHE INTACT HOST

As pointed out above, reovirus was originallyisolated from two distinct sources: the humangastrointestinal tract and the brains of mice inwhich it had produced runting disease. Fieldsand his collaborators have recently attemptedto define which genes of reovirus are responsiblefor virulence and tissue tropism. They have usedfor this purpose strains of the three serotypes ofmanmalian reovirus, as well as some of thetemperature-sensitive mutants in the collectionof Fields and Joklik. They have found that atleast two genes govem the interaction ofreoviruswith its host. First protein al, coded by gene Si,is crucial for specifying the interaction betweenreovirus and its host. It is the reovirus cell at-tachment protein (52) and to that extent speci-fies whether reovirus particles can or cannotinteract with particular types of cells; further-more, the fact that it interferes, in its free nativeform, with the adsorption of virus particles of allthree serotypes of reovirus demonstrates that itinteracts with specific receptors on cell surfaces.It is also the reovirus hemagglutinin (119) and isone of three reovirus proteins that gives rise to

neutralizing antibody (the others being X2 anda3) (31, 116). It also controls the intracellularassociation of reovirus particles with microtu-bules (5). Furthermore, it is responsible for thegeneration of cytolytic T lymphocytes after in-fection, the development of delayed-type hyper-sensitivity (117) and the generation of suppres-sor T cells (19, 29). It also controls tissue tropismto the extent that it determines neurovirulencepatterns (115, 118): intracerebral inoculation ofreovirus serotype 3 into newborn mice causes anecrotizing encephalitis that is uniformly fatal,whereas reovirus serotype 1 causes ependymalcell damage and hydrocephalus, but the animalsgenerally survive. By using recombinants of reo-virus serotypes 1 and 3, it is readily shown thatthe neurovirulence pattern is associated solelywith the nature of the Si gene that they contain,and is presumably predicated by the interactionof al with receptors on neuronal/ependymalcells.

Protein y1C also controls virulence and tissuetropism, but by a totally different mechanism.That it can be shown to do so is a strikingillustration of the often suspected but difficult todemonstrate principle that virulence is an ex-tremely complex phenomenon mediated by nu-merous processes and pathways. Protein ,tdC, itwill be recalled, is the major constituent of thereovirus outer capsid shell which is removedwhen treatment with proteases such as chymo-trypsin converts reovirus particles to cores,which are far less infectious. It appears that the,ulC of serotype 1 is much more resistant toproteases than that of serotype 3 (81). As aresult, when it reaches the intestinal tract, reo-virus serotype 1 is resistant to the proteases thatit then encounters and is able to cause infectionin intestinal tissue, whereas serotype 3 strainsare inactivated and fail to initiate significantinfections. Interestingly, when a recombinant isinoculated into the upper intestinal region thatpossesses an M2 gene of serotype 1 and an Sigene of serotype 3, fatal intracerebral infectionresults from initial multiplication in intestinaltissue and subsequent spread to the brain (seeabove). Gene M2 thus represents-a second reo-virus virulence gene, for diseases or dietary fac-tors that reduce the secretion or activity of gas-tric or pancreatic enzymes could clearly increasesusceptibility to reovirus through the mediationof the protein that it encodes.This type of study, the use of recombinants in

a system in which gene reassortment both occursreadily and is readily detected, is an excellentexample of the application of basic research toclinical problems, which in the final analysis isthe major goal of virologists, be they basic orclinical.

VOL. 45, 1981


mbr.asm

.org/D

ownloaded from


498 JOKLIK

LITERATURE CITED1. Aces, G., H. Klett, M. Schonberg, J. Christ-

man, D. H. Levin, and S. C. Silverstein.1971. Mechanism of reovirus double-strandedribonucleic acid synthesis in vivo and in vitro.J. Virol. 8:684-689.

2. Ahmed, R., and A. F. Graham. 1967. Persistentinfections in L cells with temperature-sensitivemutants of reovirus. J. Virol. 23:250-262.

3. Arnott, S., F. Hutchinson, M. Spender, M. H.F. Wilkins, W. Fuller, and R. Langridge.1960. X-ray diffraction studies of double helicalribonucleic acid. Nature (London) 211:227-230.

4. Astell, C., S. C. Silverstein, D. H. Levin, andG. Acs. 1972. Regulation of the reovirus RNAtranscriptase by a viral capsomere protein. Vi-rology 48:648-654.

5. Babiss, L. E., R. B. Luftig, J. A. Weatherbee,R. R. Weihing, U. R. Ray, and B. N. Fields.1979. Reovirus serotypes 1 and 3 differ in theirin vitro association with microtubules. J. Virol.30:863-874.

6. Bellamy, A. R., and W. K. Joklik. 1967. Stud-ies on the A-rich RNA of reovirus. Proc. Natl.Acad. Sci. U.S.A. 58:1389-1395.

7. Bellamy, A. R., and W. K. Joklik. 1967. Stud-ies on reovirus RNA. II. Characterization ofreovirus messenger RNA and of the genomeRNA segments from which it is transcribed. J.Mol. Biol. 29:19-26.

8. Bellamy, A. R., J. L. Nichols, and W. K.Joklik. 1972. Nucleotide sequences of reovirusoligonucleotides: evidence for abortive RNAsynthesis during virus maturation. Nature(London) New Biol. 238:49-51.

9. Bellamy, A. R., L. Shapiro, J. T. August, andW. K. Joklik. 1967. Studies on reovirus RNA.I. Characterization of reovirus genome RNA.J. Mol. Biol. 29:1-18.

10. Borsa, J., and A. F. Graham, 1968. Reovirus:RNA polymerase activity in purified virions.Biochem. Biophys. Res. Commun. 33:896-901.

11. Borsa, J., J. Grover, and J. D. Chapman.1970. Presence of nucleoside triphosphatephosphohydrolase activity in purified virionsof reovirus. J. Virol. 6:295-300.

12. Both, G. W., S. Lavi, and A. J. Shatkin. 1975.Synthesis of all the gene products of the reo-virus genome in vivo and in vitro. Cell 4:173-180.

13. Chang, C., and H. J. Zweerink. 1971. Fate ofparental reovirus in infected cell. Virology 46:544-555.

14. Cross, R. K., and B. N. Fields. 1972. Temper-ature-sensitive mutants of reovirus type 3:studies on the synthesis of viral RNA. Virology50:799-809.

15. Darzynkiewicz, E., and A. J. Shatkin. 1980.Assignment of reovirus mRNA ribosome bind-ing sites to virion genome segments by nucleo-tide sequence analysis. Nucleic Acids Res. 8:337-350.

16. Dunnebacke, T. H., and A. K. Klein8chmidt.1967. Ribonucleic acid from reovirus as seen in

protein monolayers by electron microscopy. Z.Naturforsch. 22b:159-164.

17. Fields, B. N., and W. K. Joklik. 1969. Isolationand preliminary genetic and biochemical char-acterization of temperature-sensitive mutantsof reovirus. Virology 37:335-342.

18. Finberg, R., H. L. Weiner, B. N. Fields, B.Benacerraf, and S. J. Burakoff. 1979. Gen-eration of cytolytic T lymphocytes after reovi-rus infection: role of S1 gene. Proc. Natl. Acad.Sci. U.S.A. 76:442-446.

19. Fontana, A., and H. L. Weiner. 1980. Interac-tion of reovirus with cell surface receptors. II.Generation of suppressor T cells by the he-magglutinin of reovirus type 3. J. Immunol.125:2660-2664.

20. Fnuruichi, Y. 1981. Allosteric stimulatory effectof S-adenosylmethionine on the RNA polym-erase in cytoplasmic polyhedrosis virus: amodel for the positive control of eukaryotictranscription. J. Biol. Chem. 256:483-493.

21. Furuichi, Y., M. Morgan, S. Muthukrishnan,and A. J. Shatkin. 1975. Reovirus messengerRNA contains a methylated, blocked 5'-termi-nal structure: m7G (5')ppp(5')GmpCp-. Proc.Natl. Acad. Sci. U.S.A. 72:362-366.

22. Furuichi, Y., and A. J. Shatkin. 1977. 5'-Ter-mini of reovirus mRNA: ability of viral coresto form caps posttranscriptionally. Virology77:566-578.

23. Gaillard, R. K., and W. K. Joklik. 1980. Theantigenic determinants of most of the proteinscoded by the three serotypes of reovirus arehighly conserved during evolution. Virology107:533-536.

24. Gillies, S., S. Bullivant, and A. R. Bellamy.1971. Viral RNA polymerases: electron micros-copy of reovirus reaction cores. Science 174:694-696.

25. Gomatos, P. J. 1967. RNA synthesis in reovirusinfected L929 mouse fibroblasts. Proc. Natl.Acad. Sci. U.S.A. 58:1798-1804.

26. Gomatos, P. J., and I. Tamm. 1963. The sec-ondary structure of reovirus RNA. Proc. Natl.Acad. Sci. U.S.A. 49:707-712.

27. Gomatos, P. J., L. Tamm, S. Dales, and R. M.Franklin. 1962. Reovirus type 3: physicalcharacteristics and interactions with L cells.Virologv 17:441-449.

28. Granboulan, N., and A. Niveleau. 1967. Etudeau microscope electronique du RNA de reovi-rus. J. Microsc. 6:23-27.

29. Greene, M. I., and H. L Weiner. 1980. Delayedhypersensitivity in mice infected with reovirus.II. Induction of tolerance and suppressor Tcells to viral specific gene products. J. Immu-nol. 125:283-287.

30. Harvey, J. D., A. R. Bellamy, W. C. Earn-shaw, and C. Schutt. 1981. Biophysical stud-ies of reovirus type 3. IV. Low-angle X-raydiffraction studies. Virology 112:240-249.

31. Hayes, E. C., P. W. K. Lee, S. E. Miller, andW. K. Joklik. 1981. The interaction of a seriesof hybridoma IgGs with reovirus particles:demonstration that the core protein X2 is ex-

MICROBIOL. REV.


mbr.asm

.org/D

ownloaded from



posed on the particle surface. Virology 108:147-155.

32. Huismans, H., and W. K. JokIlk. 1976. Reo-virus-coded polypeptides in infected celLs: iso-lation of two native monomeric polypeptideswith affinity for single-stranded and double-stranded RNA, respectively. Virology 70:411-424.

33. Iglewski, W. J., and R. M. Franklin. 1967.Purification and properties of reovirus ribonu-cleic acid. J. ViroL 1:302-307.

34. Ito, Y., and W. K. JokikL 1972. Temperature-sensitive mutants of reovirus. II. Anomalouselectrophoretic migration of certain hybridDNA molecules composed of mutant plusstrands and wild-type minus strands. Virology50:202-208.

35. Ito, Y., and W. K. JoklHk 1972. Temperature-sensitive mutants of reovirus. III. Evidencethat mutants ofgroup D ("RNA-negative") arestructural polypeptide mutants. Virology 50:282-286.

36. Johnson, R. B., IL Soeiro, and B. N. Fields.1976. The synthesis of A-rich RNA by temper-ature-sensitive mutants of reovirus. Virology73:173-180.

37. Joklik, W. K. 1972. Studies on the effect ofchymotypsin on reovirions. Virology 49:700.

38. Kapuler, A. M., N. Mendelsohn, H. Klett, andG. Ams. 1970. Four base-specific 5'-triphospha-tases in the subviral core of reovirus. Nature(London) 225:1209-1211.

39. Kates, J. R., and B. R. McAuslan. 1967. Mes-senger RNA synthesis by a "coated" viral ge-nome. Proc. Natl. Acad. Sci. U.S.A. 57:314-319.

40. Kavenoff, R., D. Talcove, and J. A. Mudd.1975. Genome-sized RNA from reovirus parti-cles. Proc. Natl. Acad. Sci. U.S.A. 72:4317-4321.

41. Kozak, M. 1977. Nucleotide sequences of 5'-ter-minal ribosome-protected initiation regionsfrom two reovirus messages. Nature (London)269:390-394.

42. Kozak, M. 1980. Binding of wheat germ ribo-somes to bisulfite-modified reovirus messengerRNA: evidence for a scanning mechanism. J.Mol. Biol. 144:291-304.

43. Kozak, M. 1980. Influence of mRNA secondarystructure on binding and migration of 40S ri-bosomal subunits. Cell 19:79-90.

44. Kozak, M., and A. J. Shatkin. 1976. Charac-terization of ribosome-protected fragmentsfrom reovirus messenger mRNA. J. Biol.Chem. 251:4259-4266.

45. Kozak, M., and A. J. Shatkin. 1977. Sequencesof two 5'-terminal ribosome-protected frag-ments from reovirus messenger RNAs. J. Mol.Biol. 112:75-96.

46. Kozak, M., and A. J. Shatkin. 1977. Sequencesand properties of two ribosome binding sitesfrom the small size class of reovirus messengerRNA. J. Biol. Chem. 252:6895-6908.

47. Kozak, M., and A. J. Shatkin. 1978. Identifi-cation offeatures in 5'-terminal fragments fromreovirus mRNA which are important for ribo-

some binding. Cell 13:201-212.48. Kozak, M., and A. J. Shatkin 1978. Migration

of 40S ribosomal subunits on messenger RNAin the presence of edeine. J. Biol. Chem. 253:6568-6577.

49. Langridge, R., and P. J. Gomatos. 1963. Thestructure of RNA. Reovirus RNA and transferRNA have similar three-dimensional struc-tures, which differ from DNA. Science 141:694-697.

50. Lau, R. Y., D. Van Alstyne, R. Berckmans,and A. F. Graham. 1975. Synthesis of reovi-rus-specific polypeptides in cells pretreatedwith cyclohexiimide. J. Virol. 16:470-478.

51. Lee, P. W. K., E. C. Hayes, and W. K. Joklik.1981. Characterization of anti-reovirus immu-noglobulins secreted by cloned hybridoma celllines. Virology 108:134-146.

52. Lee, P. W. K., E. C. Hayes, and W. K. Joklik.1981. Protein al is the reovirus cell attachmentprotein. Virology 108:156-163.

53. Levin, K. H., and C. E. SamueL 1980. Biosyn-thesis of reovirus-specified polypeptides. Puri-fication and characterization of the small-sizedclass mRNAs of reovirus type 3: coding assign-ments and translational efficiencies. Virology106:1-13.

54. Li, J. K.-K., J. D. Keene, P. P. Scheible, andW. K. Joklik. 1980. Nature of the 3'-terminalsequences of the plus and minus strands of theS1 gene of reovirus serotypes 1, 2 and 3. Virol-ogy 105:41-51.

55. Li, J. K.-K., P. P. Scheible, J. D. Keene, andW. K. Joklik 1980. The plus strands of reo-virus gene S2 are identical with its in vitrotranscripts. Virology 105:282-286.

56. Loh, P. C., and A. J. Shatkin- 1968. Structuralproteins of reovirus. J. Virol. 2:1353-1359.

57. Luftig, R. B., S. S. Kilham, A. J. Hay, H. J.Zweerink, and W. K. Joklik. 1972. An ultra-structure study of virions and cores of reovirustype 3. Virology 48:170-181.

58. Matsuhisa, T., and W. K. Joklik. 1974. Tem-perature-sensitive mutants of reovirus. V.Studies on the nature of the temperature-sen-sitive lesion of the group C mutant ts447. Vi-rology 60:380-389.

59. McCrae, M. A., and W. K. Joklik 1978. Thenature of the polypeptide encoded by each ofthe double-stranded RNA segments of reovirustype 3. Virology 89:578-593.

60. McDowell, M. J., and W. K. Joklik. 1971. Anin vitro protein synthesizing system frommouse L fibroblasts infected with reovirus. Vi-rology 45:724-733.

61. McDowell, M. J., W. K. Joklik, L. Villa-Ko-maroff, and H. F. Lodish. 1972. Translationof reovirus messenger RNAs synthesized invitro into reovirus polypeptides by severalmammalian cell-free extracts. Proc. Natl. Acad.Sci. U.S.A. 69:2649-2653.

62. Millward, S., and A. F. Graham. 1970. Struc-tural studies on reovirus: discontinuities in thegenome. Proc. Natl. Acad. Sci. U.S.A. 65:422-426.

63. Morgan, E. M., and D. W. Kingsbury. 1980.

VOL. 45, 1981


mbr.asm

.org/D

ownloaded from


500 JOKLIK

Pyridoxal phosphate as a probe of reovirustranscriptase. Biochemistry 19:484-489.

64. Morgan, E. M., and H. J. Zweerink. 1974.Reovirus morphogenesis. Core-like particles incells infected at 39' with wild-type reovirusand temperature-sensitive mutants in groupsB and G. Virology 59:556-565.

65. Morgan, E. M., and H. J. Zweerink. 1975.Characterization of transcriptase and replicaseparticles isolated from reovirus-infected cells.Virology 68:455466.

66. Mustoe, T. A., R. F. Ramig, A. H. Sharpe,and B. N. Fields. 1978. A genetic map ofreovirus. III. Assignment of the double-stranded RNA-positive mutant groups A, B,and G to genome segments. Virology 85:545-556.

67. Mustoe, T. A., R. F. Ramig, A. H. Sharpe,and B. N. Fields. 1978. Genetics of reovirus:identification of the ds RNA segments encod-ing the polypeptides of the ,u and a size classes.Virology 89:594-604.

68. Muthukrishnan, S., and A. J. Shatkin. 1975.Reovirus genome RNA segments: resistance to51 nuclease. Virology 64:96-105.

69. Nichols, J. L, A. R. Bellamy, and W. K.Joklik. 1972. Identification of the nucleotidesequences of the oligonucleotides present inreovirions. Virology 49:562-572.

70. Nichols, J. L, A. J. Hay, and W. K. Jokl1972. 5'-Terminal nucleotide sequence of reo-virusmRNA synthesized in vitro. Nature (Lon-don) New Biol. 235:105-107.

71. Nonoyama, M., and A. F. Graham. 1970. Ap-pearance of defective virions in clones of reo-virus. J. Virol. 6:693-694.

72. Palmer, E. L., and M. L. Martin. 1977. The finestructure of the capsid of reovirus type 3. Vi-rology 76:109-113.

73. Ralph, S. J., J. D. Harvey, and A. R. Bellamy.1980. Subunit structure of the reovirus spike.J. Virol. 36:894-896.

74. Ramig, R. F., and B. N. Fields. 1979. Revert-ants of temperature-sensitive mutants of reo-virus: evidence for frequent extragenic sup-pression. Virology 92:155-167.

75. Ramig, R. F., T. A. Mustoe, A. H. Sharpe,and B. N. Fields. 1978. A genetic map ofreovirus. II. Assignment ofthe double-strandedRNA-negative mutant groups C, D, and E togenome segments. Virology 85:531-544.

76. Ramig, R F., R. M. White, and B. N. Fields.1977. Suppression of the temperature-sensitivephenotype of a mutant of reovirus type 3.Science 195:406-407.

77. Ramos-Alvarez, M., and A. B. Sabin. 1954.Characteristics of poliovirus and other entericviruses recovered in tissue culture from healthyAmerican children. Proc. Soc. Exp. Biol. Med.87:655-661.

78. Rhim, J. S., K. 0. Smith, and J. L. Melnick.1961. Complete and coreless fonns of reovirus(ECHO 10): ratio of number of virus particlesto infective units in the one step growth cycle.Virology 15:428-435.

MICROBIOL. REV.

79. Robbins, F. C., J. F. Enders, T. H. Weller,and G. L. Florentino. 1951. Studies on thecultivation of poliomyelitis virus in tissue cul-ture. V. The direct isolation and serologic iden-tification of virus strains in tissue culture frompatients with paralytic and nonparalytic poli-omyelitis. Am. J. Hyg. 54:56-63.

80. Rosen, L 1960. Serologic grouping of reovirusesby hemagglutination-inhibition. Am. J. Hyg.71:252-249.

81. Rubin, D. H., and B. N. Fields. 1980. Molecularbasis of reovirus virulence. Role of the M2gene. J. Exp. Med. 152:853-868.

82. Sabin, A. B. 1959. Reoviruse8: a new group ofrespiratory and enteric viruses formerly clas-sified as ECHO type 10 is described. Science130:1387-1389.

83. Sakuma, S., and Y. Watanabe. 1971. Unilat-eral synthesis of reovirus double-stranded ri-bonucleic acid by a cell-free replicase system.J. Virol. 8:190-196.

84. Sakuma, S., and Y. Watanabe. 1972. Incorpo-ration of in vitro synthesized reovirus double-stranded ribonucleic acid into virus corelikeparticles. J. Virol. 10:943-950.

85. Sakuma, S., and Y. Watanabe. 1972. Reovirusreplicase-directed synthesis of double-strandedribonucleic acid. J. Virol. 10:628-638.

86. Schuerch, A. R., and W. K. Joklik 1973. Tem-perature-sensitive mutants of reovirus. IV. Ev-idence that anomalous electrophoretic migra-tion behavior of certain double-stranded RNAhybrid species is mutant group-specific. Virol-ogy 56:218-229.

87. Schuerch, A. R., T. Mat8uhisa, and W. K.Joklik. 1974. Temperature-sensitive mutantsof reovirus. VI. Mutant ts447 and ts556 parti-cdes that lack either one or two genome RNAsegments. Intervirology 3:36-46.

88. Shatkin, A. J. 1965. Inactivity of purified reo-virus RNA as a template for E. coli polymerasein vitro. Proc. Natl. Acad. Sci. U.S.A. 54:1721-1726.

89. Shatkin, A. J. 1974. Methylated messengerRNA synthesis in vitro by purified reovirus.Proc. Natl. Acad. Sci. U.S.A. 71:3204-3207.

90. Shatkin, A. J., and A. J. LaFiandra 1972.Transcription by infectious subviral particlesof reovirus. J. Virol. 10:698-706.

91. Shatkin, A. J., and J. D. Sipe. 1968. RNApolymerase activity in purified reoviruses.Proc. Natl. Acad. Sci. U.S.A. 61:1462-1469.

92. Shatkin, A. J., and J. D. Sipe. 1968. Single-stranded, adenine-rich RNA from purified reo-viruses. Proc. Natl. Acad. Sci. U.S.A. 59:246-253.

93. Shatkin, A. J., J. D. Sipe, and P. C. Loh. 1968.Separation of ten reovirus genome segmentsby polyacrylamide gel electrophoresis. J. Virol.2:986-991.

94. Shine, J., and L. Dalgarno. 1974. The 3'-ter-minal sequence of Escherichia coli 16S ribo-somal RNA: complementarity to nonsense tri-plets and ribosome binding sites. Proc. Natl.Acad. Sci. U.S.A. 71:1342-1346.


mbr.asm

.org/D

ownloaded from



95. Silverstein, S. C., C. Astell, J. Christman, H.Klett, and G. Aces. 1974. Synthesis of reovirusoligoadenylic acid in vivo and in vitro. J. Virol.13:740-752.

96. Silverstein, S. C., C. Astell, D. H. Levin, M.Schonberg, and G. Acs. 1972. The mecha-nism 6f reovirus uncoating and gene activationin vivo. Virology 47:797-806.

97. Silverstein, S. C., and S. Dales. 1968. Thepenetration of reovirus RNA and initiation ofits genetic function in L-strain fibroblasts. J.Cell Biol. 36:197-230.

98. Silverstein, S. C., D. H. Levin, M. Schonberg,and G. Acs. 1970. The reovirus replicativecycle: conservation of parental RNA and pro-tein. Proc. Natl. Acad. Sci. U.S.A. 67:275-281.

99. Skehel, J. J., and W. K. Joklik. 1969. Studieson the in vitro transcription of reovirus RNAcatalyzed by reovirus cores. Virology 39:822-831.

100. Skup, D., and S. Millard. 1980. mRNA cappingenzymes are masked in reovirus progeny sub-viral particles. J. Virol. 34:490-496.

101. Skup, D., and S. Millward. 1980. Reovirus-in-duced modification of cap-dependent transla-tion in infected cells. Proc. Natl. Acad. Sci.U.S.A. 77:152-156.

102. Smith, R. E., H. J. Zweerink, and W. K.Joklik. 1969. Polypeptide components of viri-ons, top component and cores of reovirus type3. Virology 39:791-810.

103. Spandidos, D. A., and A. F. Graham. 1976.Generation of defective virus after infection ofnewborn rats with reovirus. J. Virol. 20:234-247.

104. Spandidos, D. A., and A. F. Graham. 1976.Non-permissive infection ofL cells by an avianreovirus: restricted transcription of the viralgenome. J. Virol. 19:977-984.

105. Spendlove, R. S., M. E. McClain, and E. H.Lennette. 1970. Enhancement of reovirus in-fectivity by extracellular removal or alterationof the viral capsid by proteolytic enzymes. J.Gen. Virol. 8:83-93.

106. Stanley, N. F. 1961. Relationship of hepatoen-cephalomyelitis virus and reoviruses. Nature(London) 189:687-689.

107. Stanley, N. F., D. C. Dorman, and J. Pons-ford. 1953. Studies on the pathogenesis of ahitherto undescribed virus (hepato-encephalo-myelitis) producing unusual symptoms in suck-ling mice. Aust. J. Exp. Biol. Med. Sci. 31:147-159.

108. Stoltzfus, C. M., and A. K. Banerjee. 1972.Two oligonucleotide classes of single-strandedribopolymers in reovirus A-rich RNA. Arch.Biochem. Biophys. 152:733-743.

109. Van Dijk, A. A., and H. Huismans. 1980. Thein vitro activation and further characterization

of the blue-tongue virus-associated transcrip-tase. Virology 104:347-356.

110. Vasquez, C., and A. K. Kleinschmidt. 1968.Electron microscopy of RNA strands releasedfrom individual reovirus particles. J. Mol. Biol.34:137-147.

111. Vasquez, C., and P. Tournier. 1962. The mor-phology of reovirus. Virology 17:503-510.

112. Watanabe, Y., and A. F. Graham. 1967. Struc-tural units ofreovirus ribonucleic acid and theirpossible functional significance. J. Virol. 1:665-677.

113. Watanabe, Y., S. Millward, and A. F.Graham. 1968. Regulation of transcriptase ofthe reovirus genome. J. Mol. Biol. 36:107-123.

114. Watanabe, Y., L Prevec, and A. F. Graham.1967. Specificity in transcription ofthe reovirusgenome. Proc. Natl. Acad. Sci. U.S.A. 58:1040-1047.

115. Weiner, H. L., D. Drayna, D. R. Averill, Jr.,and B. N. Fields. 1977. Molecular basis ofreovirus virulence: role of the S1 gene. Proc.Natl. Acad. Sci. U.S.A. 74:5744-5748.

116. Weiner, H. L, and B. N. Fields. 1977. Neutral-ization of reovirus: the gene responsible for theneutralization antigen. J. Exp. Med. 146:1305-1310.

117. Weiner, H. L, M. I. Greene, and B. N. Fields.1980. Delayed hypersensitivity in mice infectedwith reovirus. I. Identification of host and viralgene products responsible for the immune re-sponse. J. Immunol. 125:278-282.

118. Weiner, H. L., M. L. Powers, and B. N. Fields.1980. Absolute linkage of virulence and centralnervous system cell tropism of reoviruses toviral hemagglutinin. J. Infect. Dis. 141:609-616.

119. Weiner, H. L, R F. Ramig, T. A. Mustoe, andB. N. Fields. 1978. Identification of the genecoding for the hemagglutinin of reovirus. Vi-rology 86:581-584.

120. White, C. K., and H. J. Zweerink. 1976. Studieson the structure of reovirus cores: selectiveremoval of polypeptide 2. Virology 70:171-180.

121. Zweerink, H. J. 1974. Multiple forms of SSDS RNA polymerase activity in reovirus-in-fected cells. Nature (London) 247:313-315.

122. Zweerink H. J., Y. Ito, and T. Matsuhisa.1972. Synthesis of reovirus double-strandedRNA within virion-like particles. Virology 50:349-358.

123. Zweerink, H. J., and W. K. Joklik. 1970. Stud-ies on the intracellular synthesis of reovirus-specified proteins. Virology 41:501-518.

124. Zweerink, H. J., E. M. Morgan, and J. S.Skyler. 1976. Reovirus morphogenesis: char-acterization of subviral particles in infectedcells. Virology 73:442-453.

VOL. 45, 1981


mbr.asm

.org/D

ownloaded from


structure and functionofthereovirusgenome · structure andfunction of thereovirus genome 487...

Documents