ViralPathogenesis and MolecularBiologytstep during multiplication ofVSV,Sindbis, po-lio, andsimianvirus 40 affected byDIparticles is that ofgenomereplication (see reference 16). Since

BACTEhIOLOGICAL REVIEWS, Dec. 1977, p. 811-821Copyright © 1977 American Society for Microbiology

Vol. 41, No. 4Printed in U.S.A.

Viral Pathogenesis and Molecular BiologytALICE S. HUANG

Department ofMicrobiology and Molecular Genetics, Harvard Medical School, Boston,Massachusetts 02115

INTRODUCTION .................................... 811STRUCTURE OF VSV DI PARTICLES ....................................... 812INTERFERENCE BY DI PARTICLES ......................................... 812

EFFECT OF AGE ON DI PARTICLE-MEDIATED INTERFERENCE .......... 812Encephalitis Caused by VSV in Young or Old Mice 813

Assay for DI Particles 814

Infection of Cells in Culture from Young or Old Mice .... 815Cell Cultures for Studies on Viral Pathogenesis .. .. ........ 815

ORIGIN OF VSV DI GENOMES 815

Subgenomic and Extragenomic Sequences .... ................. 815Complementary Sequences ................................................. 816Generation of DI Genomes During Replication of RNA 817

Significance of Complementary Termini in Viral Genomes 818

CONCLUSION .......................................................... 819LITERATURE CITED ................. 820

INTRODUCTIONMolecular biology applied toward understand-

ing viral diseases has come of age. An excitingfuture is in store when questions of viral patho-genesis can be answered by biochemical analysisof animal viruses and their infected cells. Thisis not, however, a one-way street. Our experiencehas been that in designing experiments whichattempt to answer questions related to diseaseprocesses new information is often gathered thatcontributes significantly to our thinking aboutgeneral problems in molecular biology. This two-way exchange between molecular biology andviral pathogenesis will be illustrated here byrecent examples in our work as well as in thatof others.

Studies on the structure of animal viruseshave been most rewarding. Identification andseparation of viral structural proteins have ledto an understanding of the components on thevirion surface that determine attachment andpenetration into host cells. An example of theimportance of these viral surface proteins is thatof paramyxoviruses. For this group of viruses,the envelope is covered by glycoproteins thatmust be cleaved before the virus particle canbecome fully infectious (14, 51). This suggeststhat attachment and penetration of this virusinto new tissues and new hosts may be deter-mined by the presence or availability of specificproteases. A recent study positively correlatescleavage of the hemagglutinating and fusing pro-teins of Newcastle disease virus to its virulencefor chickens (25).

t Eli Lilly Award Address, May 1977.

Another way to study virulence is in the ma-nipulation of conditional lethal mutants bothgenetically and biochemically. Among the manyexamples is influenza virus, where some of thegenes responsible for virulence have been iden-tified and physically separated (37). Knowingthe gene product(s) that accounts for virulencemay provide a new direction for viral prophy-laxis or for treatment of a virus disease while itis in progress.The utility of analyzing viral structural pro-

teins is elegantly demonstrated by Arnon andcolleagues (28) where a crucial oligopeptide ofonly 20 amino acids, derived from the coat pro-tein of the bacteriophage ms-2, has been foundto be essential for bacteriophage attachment toits host cell. When this oligopeptide is linked toa carrier, neutralizing antibodies against the bac-teriophage are elicited. Modification of the an-tigenicity of this oligopeptide with other carriersand the ability to synthesize vast amounts of itchemically in the laboratory indicate a wholenew era of antiviral vaccines.Another potent tool for probing virus-infected

cells is the ability to copy viral messenger ribo-nucleic acid (RNA) into highly labeled deoxyri-bonucleic acid (DNA) in vitro by using a virion-associated polymerase (2, 56). Labeled DNA isa very sensitive probe for detecting commonsequences both in cells and among related vi-ruses. Such an approach has opened up a wholeera of study of endogenous viruses-viruses thatare integrated into the genetic information oftheir hosts and are vertically transmitted fromparent to offspring (32). The finding of identicalor related sequences in different species has re-

811

on February 21, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

812 HUANG

vealed a history of the persistence and spreadof endogenous viruses among different animalspecies during the evolution of these species.These endogenous viruses usually reside in acryptic form without any expression of theirinformation. What role thay play in viral dis-eases is only beginning to be understood.These are just a few examples to indicate the

areas where animal virology may bear fruit inthe next few years. The precision of biochemicaltechniques can now be applied to questions ofdirect relevance to disease processes. My intro-duction to molecular animal virology began inthe laboratory of Robert R. Wagner at JohnsHopkins University. We used rate-zonal centrif-ugation to separate infectious standard vesicularstomatitis virus (VSV) from its deletion mutants(18). This resulted in the first isolation and char-acterization of a defective interfering (DI) par-ticle. Since then DI particles have been shownto be an important determinant of the outcomeof viral infections (see reference 16).

STRUCTURE OF VSV DI PARTICLESDI particles have all the same structural pro-

teins of the standard virus from which they arederived and are, therefore, antigenically indistin-guishable (18, 22, 58). Figure 1 illustrates theVSV structural proteins. There are only fiveproteins, one of which is glycosylated and twoof which are phosphorylated. All five proteinsare specified by the genome of standard VSV.The virions of VSV resemble bullets. Deletionmutants that form shorter bullets are readilyseparated from the standard virus. Analysis ofDI particles suggests that the length of the bulletcorresponds well with the length of the nucleo-capsid and the length of the genome RNA (20).Recently, particles equal in length to standardVSV have been detected, which contain RNAwith only one-third the complexity of the stan-dard genome, suggesting that several equivalentsof a DI RNA may become encapsulated intoone long, bullet-shaped particle (D. Rao and A.Huang, unpublished observations).The organizations of the nucleotide sequences

of these DI RNAs are peculiar (see below). Inthis paper, a molecular model to account for thereorganization of the VSV genome into DI ge-nomes will be presented. The potential signifi-cance of those rearrangements will be discussed.

INTERFERENCE BY DI PARTICLESThe finding of a DI agent immediately raises

the possibility that such an agent might be usefulin the prophylaxis of viral disease. A step to-wards achieving this goal is understanding themechanism of interference. Over the years the

important features of interference mediated byDI particles have been uncovered by using VSVas a prototype system for study. These featuresare applicable to other virus systems as well(see reference 16) and are summarized below:(i) interference occurs intracellularly and not atthe cell surface; (ii) it is specific in that it ishomotypic, i.e., interference occurs most stronglyagainst the standard virus from which the DIparticle was derived; with closely related virusesinterference occurs to a lesser degree, and withunrelated viruses it does not occur at all; (iii)this phenomenon can exist independently of theinterference mediated by interferon. The criticalstep during multiplication of VSV, Sindbis, po-lio, and simian virus 40 affected by DI particlesis that of genome replication (see reference 16).Since these viruses have different strategies forreplicating their nucleic acids, the molecular de-tails of interference by specific DI particles arelikely to vary for each individual virus group.This will also be reflected in other effects thatmay be secondary to interference with genomereplication. For instance, those steps that occurafter genome replication, or are dependent onprior genome replication, are likely to be in-hibited also. Thus, during morphogenesis, theremay be competition for limiting structural pro-teins.

Several parameters known to affect the degreeof interference exerted by DI particles have beensummarized (16). Most likely, others will beuncovered, and a knowledge of them may helpin manipulating DI particles in prophylaxis ofviral diseases. An important factor is the virusstrain. Different isolates of a virus, which maybe antigenically indistinguishable, show quanti-tative and qualitative differences with respectto the generation of DI particles. They may alsoshow a variable degree of susceptibility to inter-ference caused by a given amount ofDI particles.Of particular importance is the species of hostin which the virus is grown. Thus, the generationof DI particles and their ability to interfere arespecies specific and can even be seen in differenttissues from one animal. In certain cases it hasbeen shown that the degree of interference isaffected by a single genetic locus within aninbred strain of mice (7). This host functionaffecting DI interference is thought to act onviral nucleic acid replication.

EFFECT OF AGE ON DI PARTICLE-MEDIATED INTERFERENCE

In the 1930's Sabin and Olitsky (47-50) pub-lished a series of papers on VSV in which theyshowed that mice 2 months or older are rela-tively resistant to acute encephalitis caused by

BACTERIOL. REV.

on February 21, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

VIRAL PATHOGENESIS AND MOLECULAR BIOLOGY

a b c- L

-G

-N

-NS

-M

FIG. 1. Structural proteins of VSV. Purified standard VSV grown in suspended Chinese hamster ovarycells (54) in the presence of three different isotopes was disrupted, and the polypeptides were separated on a10% acrylamide gel containing sodium dodecyl sulfate (27). The VSV-specific protein-containing bands arelabeled L, G, N, NS, and M. (a) 32p; (b) I'Slmethionine; (c) [4C]glucosamine.

this virus. In contrast, VSV causes an encepha-litis resulting in paralysis and rapid death inmice that are only a few weeks old. Sabin andco-workers were unable to detect in these ani-mals any general systemic host response againstVSV that occurs rapidly enough to neutralizethe virus or to inhibit virus replication (36).They proposed the existence of "localized bar-riers" that develop immediately after virus in-fection and surround the sites of virus introduc-tion so that the virus does not spread to thecentral nervous system. One mechanism for the

development of such barriers may be productionof DI particles.

Encephalitis Caused by VSV in Young orOld Mice

P. Sinarachatanant has confirmed these find-ings in a recent study of VSV-caused encepha-litis in mice. Figure 2 shows the lethal dose(LD5o)ofVSV necessary to cause a 50% mortalityin young or old mice plotted against the daysafter intranasal inoculation. Young mice were 3weeks old and old mice were 2 months old. The

813VOL. 41, 1977

on February 21, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

814 HUANG

\\~- 0hId Mice, Survival 9/48

\Mr

Mice, Survival 1/50

detected in the infected animal by making brainhomogenates, especially if care was taken toinclude the rhinencephalon of the mouse alongwith the rest of the brain. Unfortunately, thisassay was unable to detect DI particles from asingle brain. While this work was in progress,Holland and Villarreal (13) reported that theywere able to detect in vivo synthesized DI par-ticles of VSV only if they pooled the brains ofsix mice. It became clear in our measurementsthat as the titer of standard virus rose in theindividual brains, death of the animal was pre-saged. Sinarachatanant was unable to detect DIparticles in individual brains throughout thecourse of these VSV infections, whether themice had expired or not. Neutralizing antibodyin the brain was detected after day 12 in theolder mice.

10.0

DAYS AFTER INTRANASAL INOCULATION

FIG. 2. In vivo assay of VSV in young and oldmice. CD-I outbred female mice, 50 at 3 weeks of ageand 48 at 8 weeks of age, were each divided intogroups containing 8 to 10 mice. Each group was

inoculated intranasally with purified standard VSV(54) at the indicated concentrations. The day that50o of the mice in a group died (LD),J is plottedversus the concentration of virus inoculated. ( 0 ) 3-week-old mice; ( 0 ) 8-week-old mice.

old mice not only expired later than young mice,but a higher dose of purified standard virus wasnecessary to cause mortality. In addition, theoverall survival rate was higher for old mice.

Assay for DI ParticlesTo measure DI particles from mouse brains,

an assay system was developed that depends on

the ability of DI particles to inhibit uridineincorporation into virus-specific RNA. It was

demonstrated earlier that DI particles dramati-cally inhibit RNA synthesis by standard virus(19). Figure 3 shows the results of a reconstruc-tion experiment. Different concentrations of DIparticles were added to cells infected with a

given concentration of standard VSV. The sen-

sitivity of this assay, at 4.5 h after infection,indicates that DI particles at a ratio of 1:100 tostandard VSV produced a significant degree ofinhibition (30%). This ratio represents a lowerlimit of detection for DI particles of 2 x 105 per

ml. Standard plaque-forming VSV was readily

(/)

cZ~

,3d)2&

1.0

0.1

0.01

0 20 40 60 80 100

% INHIBIT/ON OF RNA SYNTHESISFIG. 3. In vitro radioactive assay of VSV DI par-

ticles. Chinese hamster ovary cells (8 x 106) were

infected with standard VSV at a multiplicity of 20plaque-forming units and different concentrations ofDI particles (19). The infected cells were incubatedat 34WC in medium containing 10 jig of actinomycinD and 0.6 IiCi of '4C]uridine. Virus-specific RNAsynthesis was measured as previously described (19).A control sample infected with only standard VSVgave the maximal amount of virus-specific RNA syn-thesis at each time period after infection (pi.). Thepercentage of inhibition of RNA synthesis by DIparticles is based on a comparison between VSVRNA synthesis in the presence of increasing concen-

trations of DI particles and VSV RNA synthesis inthe absence of DIparticles.

108

Id

/zZ~~~~~/ / */.5

,00

// 3/

/ / .

0*/ 4.5

* 4.50 ~~~~~A2.5

10 A15

BACTERIOL. REV.

109

on February 21, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

VIRAL PATHOGENESIS AND MOLECULAR BIOLOGY

A recent attempt to detect DI particles oflymphocytic choriomeningitis virus made in vivowas more successful. Popescu and Lehmann-Grube (45) were able by a very sensitive nega-tive-plaque assay, based on the protection ofcells by DI particles, to detect significantamounts of DI particles in various infectedmouse tissues. The distribution and amounts ofDI particles increased with age and led to theconclusion that DI particles indeed play a rolein chronic, persistent infections.

Infection of Cells in Culture from Youngor Old Mice

To adequately test whether or not DI particlesformed the localized barriers in mice to VSV,cell cultures from mice were prepared. Johnsonet al. found that fibroblast cells obtained fromyoung and old mice grown in culture still main-tain their relative susceptibility or resistance toinfection by a togavirus (21). Infection of cellcultures made from the hind limb muscles ofyoung and old mice with purified standard VSVat different multiplicities showed that cells fromold mice, in contrast to cells from young mice,were more resistant, produced less infectiousVSV per culture, and developed a cytopathicresponse more slowly, if at all (P. Sinarachatan-ant, unpublished observations). When these cellswere infected with different ratios of DI to stan-dard particles, it was shown that interference ofVSV growth by DI particles as measured byRNA synthesis is more evident in cells from oldanimals than in cells from young animals (Fig.4). As a control, it was determined that VSVattaches and penetrates equally well in bothtypes of cell cultures.The difference in the degree of interference

exerted by a given concentration of DI particlesbetween these two cell types is consistently 30%as measured by inhibition of RNA synthesis.This inhibition is significant because it is onlyan indirect measure of the inhibition of VSVreplication; a 30% inhibition of RNA could re-flect a reduction of 99% in the production ofinfectious standard virus during one cycle ofgrowth (38). Therefore, these preliminary exper-iments indicate that DI particles may play arole, particularly during the initial stages of aninfection, in limiting the spread of VSV in oldanimals. Considerable work along these lines isneeded to define the precise relation betweenage and the synthesis of DI particles.

Cell Cultures for Studies on ViralPathogenesis

Studies at the interface of molecular virologyand viral pathogenesis are difficult to perform

10.0

Ci)IL.

01Q)

0.1

/o/

0 20 40 60 80 100

% INHIBIT/ON OF RNA SYNTHESISFIG. 4. Interference by VSV DI particles in fibro-

blast cells from young and old mice. Fibroblast cellcultures were made from the hind limb muscles oftwo-week-old or one-year-old CD-1 outbred mice, ac-cording to Johnson et al. (21). Infection of these cellsand RNA synthesis were measured as shown for Fig.3. Only the results at 4 h postinfection were plotted.

in the whole animal. There are too many param-eters that the experimenter cannot control.Therefore, development of such systems, wherecells from the animal can be put into cultureand their environment can be carefully con-trolled, offers a great potential for these typesof studies. A similar approach is to establishlong-term cultures by transformation of fibro-blast cells with simian virus 40 (7).

ORIGIN OF VSV DI GENOMESSubgenomic and Extragenomic SequencesThe central issue, as well as the most intrigu-

ing aspects of interference by DI particles, liesin the particles' effect on viral nucleic acid syn-thesis (19, 39, 41). Most of the workers in thefield suspect that inhibition during nucleic acidreplication depends on competition between DIRNA and RNA for limiting enzymes. The suc-cess of this competition depends very much onthe sequence organization along the two ge-nomes. VSV DI particles contain less RNA thandoes standard virus (20). The three nucleotidesat the 5' and 3' termini of the standard VSVgenome are inverted complementary sequences,pppApCpGp ... CpGpUOH (4, 5, 11). If defectiveRNA in DI particles were generated by internaldeletions, one would expect that the two endsof the RNA would be identical to those of thestandard virus RNA. On the other hand, there

815VOL. 41, 1977

1.1

on February 21, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

816 HUANG

could be end deletions in which one or the otherof the 5' or 3' ends of the RNA molecule woulddiffer from the parental standard virus RNA.

Hybridization studies by Reichmann and co-workers (29, 52) have shown that deletions inthe DI RNA tend to be largely from the 3' endof the genome. However, Lazzarini and co-work-ers (23) found that the three nucleotides at the3' ends of RNA from several strains of DI par-ticles are identical to those of standard RNA.To examine more than just the nucleotide

sequences at the termini, D. Rao and G. Free-man in my laboratory have compared the T1ribonuclease-generated oligonucleotides of a DIRNA to its standard RNA (Fig. 5). There arefewer spots in the fingerprint of the DI RNAcompared with that of the standard virus RNA.This finding supports the idea of a lower com-plexity and confirms the finding that the molec-ular weight of the DI particle RNA is aboutone-third the size of a standard virus RNA (20).This analysis would, also, indicate whether ornot the DI RNA contained only subgenomicfragments of the standard virus RNA. Most ofthe oligonucleotide spots, at or below the brom-ophenol blue dye marker, generated from DIRNA appeared to be included among the spotsfor standard virus RNA. However, the appear-ance of two extra spots in the DI RNA not seenin the B particle was quite striking. One spot is

at the far right of the fingerprint for DI RNA,and the other is near the center, one-fourth ofthe way down from the X to the arrow markingthe dye, bromophenol blue. Similar observationshave been made by Coffin and Kang with otherpreparations ofVSV DI particles (personal com-munications).These fingerprints indicate clearly that DI

RNA not only contains certain portions of thegenome of the standard virus RNA, but thatadditional sequences have been included. Wheredo these additional sequences come from? Dothey come from other virus-specified RNAs? Orare the extra nucleotide sequences obtainedfrom nucleic acid molecules of the host?

Complementary SequencesSeveral recent results are pertinent here. To

repeat, Banerjee's group (4) and Lazzarini'sgroup (23) have shown by an analysis of onlythree nucleotides that the 3' ends contain theidentical three nucleotides in both standard andDI genomes. Since the sequence of these threenucleotides is complementary to that found atthe 5' end of standard RNA, it is expected thatsome complementary sequences may be foundin all RNAs of VSV DI particles. Jacques Per-rault, in John Holland's laboratory, has beenable to observe, in the electron microscope, cir-cularized DI RNA and, also, dimer linear struc-

FIG. 5. Ribonuclease Tl-generated oligonucleotides ofRNA from standard and DI particles of VSV. 32plabeled, purified standard VSV and DI particles were extracted and their RNAs purified on sucrosegradients. The RNAs were digested by Ti ribonuclease and fingerprinted on two-dimensional gels (8). (X)Xylene cyanol dye marker; (B) bromophenol blue dye marker.

BACTERIOL. REV.

on February 21, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

VIRAL PATHOGENESIS AND MOLECULAR BIOLOGY 817

tures (42, 43). Similar circularization of Sindbisvirus RNA (15) and Sendai virus DI RNAs (26)as well as the segmented bunyavirus RNA (12)has been detected. Such circularization suggestscomplementarity at the termini of these RNAs.In some cases, the RNA circles included "pan-handle" structures, suggesting extensive in-verted complementary sequences at the RNAtermini (26, 42, 43).A large variety of DI particles of VSV have

now been generated (44, 46). Several groupshave noted the ability of RNA molecules fromthese DI particles to rapidly self-anneal from afew to as much as 80% (30, 40, 42, 43). Theseresults indicate that the RNA may have longregions of nucleic acid homology that are cova-lently linked within one molecule. This propertyof rapid intramolecular self-annealing has re-sulted in the term "snap-back" RNA and sug-gests very strongly that extra oligonucleotidesequences found in DI RNA may come fromcomplementary RNA sequences.

Generation of DI Genomes DuringReplication ofRNA

During replication of VSV, there is an orderedsynthesis of nucleic acids. The strategy of RNAsynthesis of VSV begins with the incoming vir-ion that contains a 40S RNA genome definedas a minus strand (17, 35). The virion becomesuncoated to the nucleocapsid stage, and thenucleocapsid serves as template for the virion-associated polymerase during transcription ofmessenger RNA (3). The messenger RNAstrands are of opposite polarity, complementaryto virion RNA, and smaller than genome RNA(17, 35). These messenger RNAs are designatedplus strands. When virus-specific protein synthe-sis is accomplished, replication of the genomecan proceed (19). This procedure is thought tooccur by synthesis of a complete complementof the genome into a plus strand of 40S RNA,which serves as template for the synthesis ofminus-strand 40S RNA genome (34, 53). Howin this scheme of synthesis can short RNA piecesbe generated that contain sequences from bothminus and plus strands?Our hypothesis for generation of these DI

genomes is presented in Fig. 6. The standardVSV genome RNA is represented by the alpha-bet with A and A' at the 3' and 5' ends, respec-tively. A and A' designate complementary se-quences at the termini. Under the usual circum-stances of replication, the minus strand is copiedinto its complement. Upon completion, anotherfull-sized molecule, the plus strand, is synthe-sized with the termini A and A' at the 3' and 5'ends, respectively (Fig. 6, no. 1).

If replication of this RNA is abortive and the3' end of the nascent strand forms a hairpin(Figure 6, no. 2), the polymerase can then con-tinue synthesis by copying the nascent strand,displacing the template, and synthesizing in thedirection of the 5' end of the nascent molecule.This is demonstrated by the two "panhandle"structures in Fig. 6, no. 3. What is generated isa plus strand that has, not only original terminiA and A' from the standard RNA molecule, butalso other complementary sequences, such as B'and B and C' and C (Fig. 6, no. 4). When aminus strand is copied from this defective plusstrand, a defective minus-strand genome is gen-erated with the same A and A' terminal se-quences and the additional complementary se-quences of B and B' as well as C and C' (Fig. 6,no. 5).Other complementary sequences can readily

be generated by introducing a second hairpinloop during synthesis of this minus-strand-defec-tive genome. This would result in sequencesfrom the standard genome, the minus strand,interspersed by complementary sequences fromthe plus strand (Fig. 6, no. 6). Such extensivecomplementarity will allow considerable self-an-nealing within the RNA molecule itself.

In general, however, most DI particles ofVSVcontain RNA with a majority of their sequencesfrom the 5' end of the standard genome (51).The model shown in Fig. 6 generates defectiveRNA with sequences from the 3' end of thestandard genome. Synthesis of RNA duringVSV replication is asymmetrical in that muchmore minus-strand 40S RNA is made than plus-strand 40S RNA. A model for generation ofVSV DI genomes from the 5' end is formallysimilar to that shown in Fig. 6 and is moreaccurately depicted in Fig. 7. Formation of thehairpin loop in the nascent strand occurs duringsynthesis of the minus strand while full-sized,plus-strand 40S RNA is utilized as the template(Fig. 7, no. 2). The defective genome generatedin this way would then contain the sequencesT, U, V, W, X, Y, Z, and A', all from the 5' endof the standard genome, as well as sequencesY', Z', and A, which would be complementaryto the sequences from the 5' end of standardRNA. Gene order of the various cistrons ofVSV,as determined by Ball and White (1), is shownon the bottom of Fig. 7, no. 5. Comparing theVSV gene order to the defective RNA generatedin Fig. 7, complementary sequences will existbetween the defective genome and the plus-strand messenger RNA for the L protein. Suchresults have been found by Reichmann and co-workers (29, 52).

If such a model (Fig. 7) were true for VSV, it

VOL. 41, 1977

on February 21, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

818 HUANG

(6-1) - 3' ABCDEFG H I J+ 5' ABC'D'E'F'G'H'-)

(6-2) -3' A B C D E F G H I J+5'C 'ID

+5' AIBC *. H

3'

(6-3) E FGIEI GIDI H'

C'IB'IA'

5'

3 '

K L M N O P Q R S T U V W X Y Z A'

K L M N 0 P Q R S T U V W X Y Z A'

E'FI

EG,

C'CB BA A

51 3'

(6-4) +5'

(6-5) -3'

A'B'CIDI EIFIG'HC B A

A B C D E F G H C B A'

CD E F G H CIBIA(6-6) -3'

CD E F G H C B A 3

OR

-3' A B C H' G F'IE 'D'C D E F G H C B'AFIG. 6. Model for the generation of VSV DI RNA from the 3' end.

would be expected (i) that some of the extragen-omic, or complementary, sequences would residenear the 3' terminus of the RNA extracted fromthe DI particle and (ii) that self-annealed RNAsequences would be enriched for sequences atboth the 3' and the 5' ends. We have obtainedpreliminary evidence indicating that one of theextragenomic oligonucleotides of DI RNA,shown in Fig. 5, is indeed near the 3' end andanneals intramolecularly to form ribonuclease-resistant fragments (G. Freeman, D. Rao, andA. Huang, unpublished observations).The model for generation of VSV DI genomes

presented in Fig. 6 and 7 may be applicable toall linear viral genomes, whether they are doublestranded or single stranded, and whether theycontain DNA or RNA. A similar method hasbeen discussed for the observed generation ofdefective adenovirus DNA by Daniell (6). It istempting to speculate that viruses from subacutesclerosing panencephalitis which contain infor-mation in excess from the standard measles ge-nome (10) and that Sindbis virus deletion mu-tants (9, 24) are generated in a similar manner.In addition, this model would predict that not

only smaller defective genomes are generated,but that defective genomes larger than the stan-dard virus genome could be generated if thehairpin occurred more than halfway from theterminus initiating replication ofthe nucleic acidand if most of the nascent strand is copied intocomplementary sequences covalently linked tothe defective genome.

Significance of Complementary Termini inViral Genomes

The panhandle structures depicted in Fig. 6and 7 are of potential significance. When the 3'and the 5' ends are juxtaposed, the sequencesA and A' are annealed to each other. The for-mation of similar double-stranded termini withA and A' annealed to each other can occur notonly within one molecule by the formation of apanhandle, but plus and minus 40S strands ofVSV annealed together contain ends which alsohave A and A'. In addition, nascent chain syn-thesis, once it is under way, immediately pre-sents an end with the A' sequences ofthe nascentstrand 5' terminus and the A sequences of thetemplate 3' terminus annealed together. Such a

5'

5'

3'

5'

5'

5'

BACTERIOL. REV.

on February 21, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

VIRAL PATHOGENESIS AND MOLECULAR BIOLOGY

(7-1) + 5' A'B'C'D'E'F'G'H'I'J'K'L'M'NO0PIQIR'S'T'U'V'W'X'Y'Z'AAU V W X Y Z A

(7-2) +5' A B'CIDIEIFIGIH'IIJ'K'L'MINIOPIQIR'SIT'U'V'W'X'Y'Z' AU V wX

ZA3'

(7-3) UVT X

zIzA A'

3' 5'

5'(7-4) - 3'

- 3 UPGPCP Ai IM

N NS M G-GPCPAPPP 5'

L

FIG. 7. Model for the generation of VSV DI RNA from the 5' end.

juxtaposition of the 3' and 5' termini by comple-mentarity between A and A' may serve as a

recognition site for polymerase attachment andgenome replication. Therefore any RNA mole-cule, or molecules together, that can presentdouble-stranded termini consisting of a 3' and a5' terminus annealed to each other would havethe potential of being recognized by the polym-erase and would be copied at the 3' end withconcomitant displacement of the 5' end. Such amodel for polymerase recognition of double-stranded or single-stranded templates has beensuggested by Lechner and Kelly (31) for initiat-ing replication of adenovirus DNA (55).

It is also highly likely that the extent of com-plementarity, as well as the base compositioninternal to the termini, determines in some waystability of the binding of the polymerase or

rate of replication of the molecule. By increasingthe extent of homology at the termini, VSV DIRNA may compete more or less effectively forthe polymerase than standard VSV RNA. Thesecomplementary termini may be important forother functions. For instance, Marcus and Sek-ellick (33) have found that DI particles with a

great deal of genome complementarity are themost effective inducers of interferon.

CONCLUSIONThe study of inverted terminal complemen-

tary sequences may have potential impact inseveral important areas. For viral diseases, thesestudies may reveal the underlying principles bywhich certain defective genomes can competeand successfully inhibit synthesis of standardvirus. For molecular biology, these studies mayreveal the link between linear viral genomesthat can circularize and the mechanism by whichtheir nucleic acids are replicated.For generating DI genomes, this model, based

on the presence of inverted complementary ter-mini on linear molecules, precludes random re-combination between RNA molecules as ameans of generating DI genomes. This modelpredicts that the DI genomes will arise onlyfrom one or the other end of the genome. Com-plementary or extragenomic sequences arise byhairpin formation during synthesis. How inter-nal hairpin loops are made during nucleic acidreplication is not clearly understood, but suchevents are thought to occur when the reversetranscriptase functions in vitro in the absenceof actinomycin D (57; I. Verma, personal com-

munication). Therefore, these studies on the ge-nomes of VSV DI particles, which were initiatedbecause of the intrinsic relevance of interferenceto viral diseases, are likely to lead to answersbasic for the replication of nucleic acids and forthe generation of deleted and inverted se-quences.

3'

5'

(7-5)

A ZIyIT U V W X Y Z A I

819VOL. 41, 1977

on February 21, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

820 HUANG

ACKNOWLEDGMENTSPreviously unpublished material has been contrib-

uted by members of my laboratory: Pantipa Sinara-chatanant, Gail M. Clinton, Sheila P. Little, GordonFreeman, and Donald Rao. I thank Tom Kelly andJohn Coffin for stimulating discussions and JacquesPerrault, Ellen Daniell, and C. Y. Kang for preprints.

This work was supported by Public Health Serviceresearch grants AI 10100 from the National Instituteof Allergy and Infectious Diseases and grant VC 63from the American Cancer Society. Alice S. Huang isa Public Health Service Research Career Develop-ment awardee.

LITERATURE CITED1. Ball, L. A., and C. N. White. 1976. Order of

transcription of genes of vesicular stomatitisvirus. Proc.,Natl. Acad. Sci. U.S.A. 73:442-446.

2. Baltimore, D. 1970. RNA-dependent DNA po-lymerase in virions of RNA tumour viruses.Nature (London) 226:1209-1210.

3. Baltimore, D., A. S. Huang, and M. Stampfer.1970. Ribonucleic acid synthesis of vesicularstomatitis virus. II. An RNA polymerase in thevirion. Proc. Natl. Acad. Sci. U.S.A. 66:572-576.

4. Banerjee, A., and D. P. Rhodes. 1976. 3' termi-nal sequence of vesicular stomatitis virus ge-nome RNA. Biochem. Biophys. Res. Commun.68:1387-1394.

5. Colonno, R. J., and A. K. Banerjee. 1976. Aunique RNA species involved in initiation ofvesicular stomatitis virus RNA transcription invitro. Cell 8:197-204.

6. Daniell, E. 1976. Genome structure of incompleteparticles of adenovirus. J. Virol. 19:685-708.

7. Darnell, M. B., and H. Koprowski. 1974. Ge-netically determined resistance to infection withgroup B arboviruses. II. Increased productionof interfering particles in cell cultures from re-sistant mice. J. Infect. Dis. 129:248-256.

8. DeWachter, R., and W. Fiers. 1972. Preparativetwo-dimensional polyacrylamide gel electro-phoresis of 32P-labeled RNA. Anal. Biochem.49:184-197.

9. Guild, G., and V. Stollar. 1977. Defective inter-fering particles of Sindbis virus. V. Sequencerelationships between SV.STD 42S RNA and in-tracellular defective viral RNAs. Virology77:175-188.

10. Hall, W. W., and V. terMeulen. 1976. RNAhomology between subacute sclerosing panen-cephalitis and measles viruses. Nature (London)264:474-477.

11. Hefti, E., and D. H. L. Bishop. 1975. The 5'nucleotide sequence of vesicular stomatitis viralRNA. J. Virol. 15:90-96.

12. Hewlett, M. J., R. F. Pettersson, and D. Bal-timore. 1977. Circular forms of Uukuniemi vir-ion RNA: an electron microscopic study. J. Vi-rol. 21:1085-1093.

13. Holland, J. J., and L. P. Villarreal. 1975. Puri-fication of defective interfering T particles ofvesicular stomatitis and rabies viruses gener-ated in vivo in brains of newborn mice. Virology67:438-449.

14. Homma, M. 1972. Trypsin action on the growthof Sendai virus in tissue culture cells. I. Resto-ration of the infectivity for L cells by directaction of trypsin on L cell-borne Sendai virus.J. Virol. 8:619-629.

15. Hsu, M. T., H. J. King, and N. Davidson. 1971.An electron microscope study of Sindbis virusRNA. Cold Spring Harbor Symp. Quant. Biol.38:943-950.

16. Huang, A. S., and D. Baltimore. 1977. Defectiveinterfering animal viruses, p. 73-116. In H.Fraenkel-Conrat and R. R. Wagner (ed.), Com-prehensive virology, vol. 10. Plenum PublishingCorp., New York.

17. Huang, A. S., D. Baltimore, and M. Stampfer.1970. Ribonucleic acid synthesis of vesicularstomatitis virus. III. Multiple complementarymessengerRNAmolecules. Virology42:946-957.

18. Huang, A. S., J. W. Greenawalt, and R. R.Wagner. 1966. Defective T particles of vesicu-lar stomatitis virus. I. Preparation, morphology,and some biologic properties. Virology30:161-172.

19. Huang, A. S., and E. K. Manders. 1972. Ribo-nucleic acid synthesis of vesicular stomatitisvirus. IV. Transcription by standard virus inthe presence of defective interfering particles.J. Virol. 9:909-916.

20. Huang, A. S., and R. R. Wagner. 1966. Com-parative sedimentation coefficients of RNA ex-tracted from plaque-forming and defective par-ticles of vesicular stomatitis virus. J. Mol. Biol.22:381-384.

21. Johnson, R. T., H. F. McFarland, and S. E.Levy. 1972. Age-dependent resistance to viralencephalitis: studies of infections due to Sindbisvirus in mice. J. Infect. Dis. 125:257-262.

22. Kang, C. Y., and L. Prevec. 1969. Proteins ofvesicular stomatitis virus. I. Polyacrylamide gelanalysis of viral antigens. J. Virol. 3:404413.

23. Keene, J. D., M. Rosenberg, and R. A. Lazza-rini. 1977. Characterization of the 3' terminusof RNA isolated from vesicular stomatitis virusand from its defective interfering particles. Proc.Natl. Acad. Sci. U.S.A. 74:1353-1357.

24. Kennedy, S. I. T., C. J. Bruton, and S. Schles-inger. 1976. Defective interfering passages ofSindbis virus: nature of the defective virion. J.Virol. 19:1034-1043.

25. Klenk, H.-D., and R. Rott. 1977. Virus infectionand the cell surface. In G. Poste and L. G.Nicholson (ed.), Cell surface reviews, vol. 2.Elsevier-North Holland Publishing Co., NewYork.

26. Kolakofsky, D. 1976. Isolation and characteriza-tion of Sendai virus DI-RNAs. Cell 8:547-555.

27. Laemmli, W. K. 1970. Cleavage of structural pro-teins during the assembly of the head of bacte-riophage T4. Nature (London) 227:680-682.

28. Langbeheim, H., R. Arnon, and M. Sela. 1976.Antiviral effect on ms-2 coliphage obtained withsynthetic antigen. Proc. Natl. Acad. Sci. U.S.A.73:4636-4640.

29. Leamnson, R. N., and M. E. Reichmann. 1974.The RNA of defective vesicular stomatitis virusparticles in relation to viral cistrons. J. Mol.

BACTERIOL. REV.

on February 21, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

VIRAL PATHOGENESIS AND MOLECULAR BIOLOGY

Biol. 85:551-568.30. La rini, R. A., G. H. Weber, L. D. Johnson,

and G. M. Stamminger. 1975. Covalentlylinked message and anti-message (genomic)RNA from a defective vesicular stomatitis virusparticle. J. Mol. Biol. 97:289-308.

31. Lechner, R. L, and T. J. Kelly. 1977. The struc-ture of replicating adenovirus-2 DNA mole-cules. Cell, in press.

32. Lowy, D. R., W. P. Rowe, N. Teich, and J. W.Hartley. 1971. Murine leukemia virus: high-frequency activation in vitro by 5-iododeoxy-uridine and 5-bromodeoxyuridine. Science174:155-156.

33. Marcus, P. I., and M. J. Sekellick. 1977. Defec-tive interfering particles with covalently linked[±]RNA induced interferon. Nature (London),in press.

34. Morrison, T. G., M. Stampfer, H. F. Lodish,and D. Baltimore. 1975. In vitro translationof vesicular stomatitis virus messenger RNAsand the existence of a 40S "plus" strand, p.293-306. In B. W. J. Many and R. D. Barry(ed.), Negative strand viruses. Academic PressInc., New York.

35. Mudd, J. A., and D. F. Summers. 1970. Poly-somal ribonucleic acid of vesicular stomatitisvirus-infected HeLa cells. Virology 42:958-968.

36. Olitsky, P. K., A. B. Sabin, and H. R. Cox.1936. An acquired resistance of growing animalsto certain neurotropic viruses in the absence ofhumoral antibodies or previous exposure to in-

/ fection. J. Exp. Med. 64:723-737.37. Palese, P., and J. L. Schulman. 1976. Differ-

encs in RNA patterns of influenza A viruses.J. Tirol. 17:876-884.

38. Palma, E. L., and A. S. Huang. 1974. Cyclicproduction of vesicular stomatitis virus causedby defective interfering particles. J. Infect. Dis.129:402-410.

39. Palma, E. L., S. M. Perlman, and A. S. Huang.1974. Ribonucleic acid synthesis of vesicularstomatitis virus. VI. Correlation of defectiveparticle RNA synthesis with standard RNAreplication. J. Mol. Biol. 85:127-136.

40. Perrault, J. 1976. Cross-linked double-strandedRNA from a defective vesicular stomatitis virusparticle. Virology 70:360-371.

41. Perrault, J., and J. J. Holland. 1972. Absenceof transcriptase activity and transcription-in-hibiting ability in defective interfering particlesof vesicular stomatitis virus. Virology50:159-170.

42. Perrault, J., and R. W. Leavitt. 1977. Charac-terization of snap-back RNAs in vesicular sto-matitis defective interfering virus particles. J.Gen. Virol., in press.

43. Perrault, J., and R. W. Leavitt. 1977. Comple-mentary terminal repeat sequences in single-stranded RNAs and snap-back RNAs from ve-sicular stomatitis virus defective interfering par-ticles. J. Gen. Virol., in press.

44. Petric, M., and L. Previc. 1970. Vesicular sto-matitis virus: a new interfering particle, intra-cellular structures, and virus-specific RNA. Vi-

rology 41:615-630.45. Popescu, M., and F. Lehmann-Grube. 1977.

Defective interfening particles in mice infectedwith lymphocytic choriomeningitis virus. J.Gen. Virol., in press.

46. Reichmann, M. E., C. R. Pringle, and E. A. C.Follett. 1971. Defective particles in BHK cellsinfected with temperature-sensitive mutants ofvesicular stomatitis virus. J. Virol. 8:154-160.

47. Sabin, A. B., and P. K. Olitaky. 1937. Influenceof host factors on the neuroinvasiveness of ve-sicular stomatitis virus. I. Effect of age on theinvasion of the brain by virus instilled in thenose. J. Exp. Med. 66:15-34.

48. Sabin, A. B., and P. K. Olitsky. 1937. Influenceof host factors in neuroinvasiveness of vesicularstomatitis virus. II. Effect of age on the invasionof the peripheral and central nervous systemsby virus injected into the leg muscles or theeye. J. Exp. Med. 66:35-56.

49. Sabin, A. B., and P. K. Olitsky. 1938. Influenceof host factors on neuroinvasiveness of vesicularstomatitis virus. III. Effect of age and pathwayof infection on the character and localizationof lesions in the central nervous system. J. Exp.Med. 67:201-227.

50. Sabin, A. B., and P. K. Olitsky. 1938. Influenceof host factors on neuroinvasiveness of vesicularstomatitis virus. IV. Variations on neuroinva-siveness in different species. J. Exp. Med.67:229-249.

51. Scheid, A., and P. W. Choppin. 1974. Identifi-cation of biological activities of paramyxovirusglycoproteins. Activation of cell fusion, hemol-ysis, and infectivity by proteolytic cleavage ofan inactive precursor protein of Sendai virus.Virology 57:475-490.

52. Schnitzlein, W. M., and M. E. Reichmann.1976. The size and the cistronic origin of defec-tive vesicular stomatitis virus particle RNAs inrelation to homotypic and heterotypic interfer-ences. J. Mol. Biol. 101:307-325.

53. Soria, M., S. P. Little, and A. S. Huang. 1974.Characterization of vesicular stomatitis virusnucleocapsids. I. Complementary 40S RNAmoleculesinnucleocapsids.Virology6l:270-280.

54. Stampfer, M., D. Baltimore, and A. S. Huang.1971. Absence of interference during high mul-tiplicity infection by clonally purified vesicularstomatitis virus. J. Virol. 7:409-411.

55. Sussenbach, J. S., and M. G. Kuijk. 1977. Stud-ies in the mechanism of replication of adenovi-rus DNA. V. The location of termini of replica-tion. Virology 77:149-157.

56. Temin, H., and S. Mitzutani. 1970. RNA-de-pendent DNA polymerase in virions of Roussarcoma virus. Nature (London) 226:1211-1213.

57. Varmus, H. E., W. E. Levinson, and J. M.Bishop. 1971. Extent of transcription by theRNA-dependent DNA polymerase of Rous sar-coma virus. Nature (London) New Biol.233:19-21.

58. Wagner, R. R., T. C. Schnaitman, R. M. Sny-der, and C. A. Schnaitman. 1969. Proteincomposition of the structural components ofvesicular stomatitis virus. J. Virol. 3:611-618.

821VoL. 41, 1977

on February 21, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from