VIROLOGY

182, 430-441 (1991)

Infectious

RNA Transcripts from Ross River Virus cDNA Clones and the Construction and Characterization of Defined Chimeras with Sindbis Virus RICHARD J. KUHN, HUBERT G. M. NIESTERS,’ ZHANG HONG, AND JAMES H. STRAUSS’ Division of Biology, California Received December

Institute of Technology, 20, 1990; accepted

Pasadena, February

California

9 1 125

12, 199 1

We have constructed a full-length cDNA clone of the virulent T48 strain of Ross River virus, a member of the alphavirus genus. Infectious RNA can be transcribed from this clone using SP6 or T7 RNA polymerase. The rescued virus has properties indistinguishable from those of the T48 strain of Ross River virus. We have used this clone, together with a full-length cDNA clone of Sindbis virus, to construct chimeric plasmids in which the 5’ and the 3’ nontranslated regions of the Sindbis and Ross River genomes were exchanged. The nontranslated regions of the two viral genomes differ in both size and sequence although they maintain specific conserved sequence elements. Virus was recovered from all four chimeras. Chimeras containing heterologous 3’ nontranslated regions had replicative efficiencies equal to those of the parents. In contrast, the chimeras containing heterologous 5’ nontranslated regions were defective in RNA synthesis and virus production, and the severity of the defect was dependent upon the host. Replication of a virus containing a heterologous 5’ nontranslated region may be inefficient due to the formation of defective protein-RNA complexes, whereas, the presumptive complexes formed between host or virus proteins and the 3’ nontranslated region to promote RNA synthesis appear to function normally in the chimeras. o 1991 Academic Press,

Inc.

The genome of the prototype RR strain T48 consists of a single-strand molecule of RNA of positive polarity, 11,851 nucleotides in length exclusive of the poly(A) tail. The complete genomic sequences of the T48 strain and the mouse avirulent NB 5092 strain have been determined by sequencing both genomic RNA and cDNA clones (Dalgarno eT al,, 1983; Faragher et al,, 1988; Strauss et a/., 1988). The genomic RNA is translated into four nonstructural proteins (called nsP1 to nsP4), which are required for virus replication (Strauss and Strauss, 1986). These nonstructural proteins, possibly together with host proteins, form the viral replication complex which is responsible for the synthesis of plus- and minus-sense genome-length RNA as well as a subgenomic RNA which codes for the structural proteins of the virus. Comparative studies of alphaviruses have identified potentially important conserved sequences and structures that may serve as promoters for various functions during the replication of the viral genome (Ou eT al., 1981, 1983; Strauss and Strauss, 1986). The development of a cDNA clone to SIN from which infectious RNA can be transcribed in vitro has permitted molecular genetic analyses of these conserved sequences (Rice et al., 1987; Grakoui et al., 1989; Levis et al., 1990; Kuhn ef a/., 1990; Niesters and Strauss, 1990a,b). From these studies, it is clear that the conserved sequence elements are essential for virus replication and that single nucleotide substitutions or dele-

INTRODUCTION Ross River virus (RR), a member of the alphavirus genus, is the causative agent of epidemic polyarthritis in humans in Australia (Doherty et al., 1964; for a review see Kay and Aaskov, 1989). The virus is primarily restricted to Australia, but recently it spread to Fiji and the American Samoa and Cook Islands, infecting tens of thousands of people (Marshall and Miles, 1984). RR was first isolated in 1963 from a pool of Aedes vigilax mosquitoes (Doherty et al., 1963) and only later was it recovered from an epidemic polyarthritis patient (Aaskov et a/., 198 1). RR has an extremely broad host range, but in nature produces clinical infections only in humans and possibly horses. Infection of humans often leads to polyarthritis accompanied by fever and rash. An age-dependent infection in mice can also be induced in the laboratory which leads to hind limb paralysis and death (Taylor and Marshall, 1975). Unlike many other alphaviruses, including Sindbis (SIN) whose primary virus reservoir is birds, RR is normally maintained in small mammals (Kay and Aaskov, 1989). The reason for this difference in virus ecology is not known but may be related to the replicative capacity of the virus in different hosts. ’ Present address: Diagnostic Centre SSDZ, P.O. Box 5010, 2600 GA Delft, The Netherlands. ’ To whom reprint requests should be addressed. 0042-6822191

$3.00

Copyright 0 1991 by Academic Press. Inc. All rights of reproduction I” any form reserved.

430

ROSS RIVER AND SINDBIS CHIMERIC

tions can have deleterious effects on virus growth. These studies also suggest that host proteins may interact directly with these conserved sequence elements. In this paper we describe a new approach toward understanding how these sequences may function and the relationships between conserved sequences among different alphaviruses. We describe the construction of a full-length cDNA clone of the virulent RR strain T48 from which infectious RNA transcripts can be produced in vitro. We have used this clone together with a full-length SIN clone to construct a series of chimeric viruses which have exchanged either the 5’ or the 3’ nontranslated regions (NTRs) of RR and SIN. These chimeras exchange conserved sequence elements found within the 5’and 3’ NTRs along with other sequences which do not appear to be conserved. The properties of these viruses are described and suggest that the 5’ and 3’ NTRs function as distinct units in the viral genome.

MATERIALS AND METHODS Ceils and viruses Ross River virus (T48 strain) (Berge, 1975) was isolated and grown as previously described (Dalgarno et a/., 1983). The heat-resistant, small plaque parental SIN used in these experiments was derived from cDNA clone pToto52 (Kuhn et al., 1990). The parental RR refers to the T48 strain of RR rescued from cells transfected with pRR64 RNA. Viruses recovered following transfection were plaque-purified twice in BHK cells and were then grown in Vero cells to obtain high titer stocks. Mosquito C6/36 cells were maintained as previously described (Kuhn et al., 1990).

Construction clones

of full-length

Ross River virus cDNA

The method used for constructing cDNA clones from RR genomic RNA has been previously described (Strauss et al., 1988). The resulting clones were mapped on the viral genome by restriction enzyme mapping. The termini of the clones were confirmed by direct sequencing (Sanger et al., 1977; Maxam and Gilbert, 1980). These clones represented the entire RR genome with the exception of the first 373 nucleotides. A specific primer (dGGCAACAGCACGTCAGGG), complementary to nucleotides 13 12-l 330, was synthesized and used to prime first strand cDNA synthesis in order to obtain the 5’terminal sequences. After synthesizing the second strand, the cDNA was digested with the restriction enzymes BarnHI and HindIll to generate a 1202-base pair fragment (nucleotides 36 to 1238). This fragment was purified from low melting agarose

VIRUSES

431

and cloned into pGEM3Z (Promega). The clone was designated pRRV4. Using pRRV4 and the other subclones depicted in Fig. 1, stepwise ligations were performed which resulted in a RR cDNA clone representing all but the first 36 nucleotides of RR. The sequence of the entire construct was determined by the dideoxy method @anger et a/., 1977) after subcloning into M13mp18 or Ml 3mpl9, or directly on plasmid DNA (Chen and Seeburg, 1985) using Sequenase (United States Biochemicals). The 3’terminal poly(A) tract consists of 30 A residues followed by an EcoRl site. To ensure a unique restriction site immediately downstream of the 3’ poly(A) for run-off transcription, several intermediate plasmids were constructed which resulted in the placement of a unique Sacl restriction site immediately 3’ of the EcoRl site. To determine the sequence of the 5’ end of the T48 genome RNA, a primer annealing to nucleotides 1312 to 1330 was used to produce cDNA. The singlestranded cDNA was purified by electrophoresis in low melting agarose and was treated with terminal deoxynucleotide transferase (Pharmacia) in the presence of 1 mNI dGTP. The dG-tailed cDNA was extracted with phenol/chloroform and precipitated with ethanol. The DNA was amplified by the polymerase chain reaction using an oligo(dC) primer together with the original cDNA primer. The product was then purified byelectrophoresis in low melting agarose and sequenced by conventional dideoxy techniques (Sanger et a/., 1977). The 5’ terminus of the RR genome was chemically synthesized as a double-stranded adaptor of 56 nucleotides which contained (from 5’ to 3’): a Sacl overhang, an SP6 RNA polymerase promoter, a single G residue to promote transcription initiation, and the first 36 nucleotides of RR up to the BarnHI site. This synthetic adapter with the 5’ end of RR and the SP6 transcription promoter was ligated to the remainder of the RR cDNA clone to generate a full-length cDNA clone, designated pRR64.

Site-directed mutagenesis and the construction recombinant plasmids

of

Oligonucleotide-directed mutagenesis was carried out essentially as described elsewhere (Kunkel, 1985; Zoller and Smith, 1984; Kuhn et al., 1990). Restriction sites were engineered in pRR64 and pToto52 for the purpose of reciprocal exchange involving the NTRs of the two genomes. Table 1 lists the restriction sites created and the nomenclature of the resulting clones. The chimeric plasmids were constructed using these and other restriction sites and are shown in Table 2. The construction of pRFU3’SIN was slightly more involved.

432

KUHN ET AL TABLE 1

Transfections and plaque assays

SUMMARYOFcDNA CLONESCONTAINING NEWRESTRICTION SITES Designation

Virus

Restriction site inserteda

pRR64 pRR35 pToto52 pToto55 pToto57

Ross River Ross River Sindbis Sindbis Sindbis

none Xbal (74) BstEll (11380) BamHl(48) Xbal (54)

a Nucleotide position of the 5’end of the recognition site is given in parenthesis. Numbering begins at the 5’end of the viral cDNA insert.

The 363-bp BstEII-EcoRI fragment from pToto52 was purified by electrophoresis in low melting agarose. Two oligonucleotide primers were made for use in a polymerase chain reaction. Primer 1 contained a 5’ Hindlll restriction site followed by a BstEll restriction site and 9 nucleotides representing the 5’ end of the purified SIN fragment. Primer 2 contained a 5’Sacl restriction site followed by an EcoRl restriction site and then 12 T residues. These primers were used to amplify the BstEIl-EcoRI fragment and add sticky ends. The 5’ HindIll end facilitated cloning into the RR clone while the 3’Sacl end maintained the unique restriction site used to linearize templates. The PCR product was digested with HindIll-Sacl, and the fragment was purified and ligated to the appropriate fragments from pRR64 to generate the chimeric clone pRR/3’SIN. In vitro transcriptions The full-length cDNA clone of RR was prepared for transcription by linearization at the unique Sac1 restriction site which follows the poly(A) tract. The recombinant clones were linearized with either Sac1 or Xhol. Transcription was carried out using SP6 RNA polymerase as described previously (Rice et al., 1987).

Confluent monolayers of BHK-2 1 (clone 15; obtained from Joel Dalrymple; these cells have been cured of the previously described mycoplasma contamination) in 35-mm tissue culture plates were used for transfection using DEAE-dextran as previously described (Rice et al., 1987). Following transfection, cells were overlayed with 1% agarose (Seakem ME, FMC Corp., Marine Colloids Division, Rockland, ME) in MEM containing 5% fetal bovine serum and plaques were allowed to develop at 37” for 2 to 3 days. Plaques were picked prior to neutral red staining to improve the viability of the virus. Alternatively, transfection using lipofectin (Life Technologies, Inc., Gaithersburg, MD) was employed according to the manufacturer’s suggestions. The RNA-lipofectin mixture was incubated with the cells for 2 to 4 hr at 37” after which it was removed and an agar overlay was added. One-step growth analysis and RNA synthesis Differential growth curves were pet-formed by infecting confluent Vero cell monolayers or C6/36 mosquito cell monolayers (35-mm plate) at a multiplicity of infection of 5 (as determined in Vero cells) in PBS containing 1% fetal calf serum. Following a I-hr incubation at room temperature, the inoculum was removed, Eagle’s minimal essential medium containing 5% fetal calf serum was added, and the cells were incubated at either 37” for Vero cells or 30” for C6/36 cells. The culture fluid was removed and fresh medium added every 30 min for the first 2 hr, and every hour after that. Culture fluid samples were assayed for plaque-forming virus by titration on Vero cell monolayers at 37”. The results shown are averages of two independent growth experiments. RNA synthesis was analyzed in cells infected at a multiplicity of 5 in PBS containing 1% fetal calf serum. Following 1 hr of incubation at room temperature, the

TABLE 2

Ross RIVER-SINDBIS CHIMERICVIRUSES Designation

Sindbis sequences8

Ross River sequencesb

Recovered virus

5’ SIN/RR 5’ RR/SIN 5’ RR 36/SIN RW3’SlN SIN0 RR

nt 1 to nt 54 (XbaF) nt 55 (XbaF) to Poly(A) nt 49 (f3amHIC)to Poly(A) nt 11,381 (5stEIIC)to Poly(A) nt 1 to nt 11,385 (B.stEIIC)

nt 75 (Xbal’) to Poly(A) nt 1 to nt 74 (Xbalc) nt 1 to nt 36 (BarnHI”) nt 1 to nt 1 1,334 (HindIll) nt 11,330 (HindIll) to Poly(A)

Yes Yes No Yes Yes

a 5’ NTR:59 nt; 3’ NTR:322 nt (nt 11,382-l 1,703). b 5’ NTR:78 nt; 3’ NTR:524 nt (nt 11,328-l 1,851). ’ Engineered restriction sites.

ROSS RIVER AND SINDBIS CHIMERIC

433

VIRUSES

pFlRv4 Bl 5130 s37

S66 s12

A 30

FIG. 1. Library of cDNA clones used to construct the full-length RR clone. The top line is a schematic representation of the RR genome. Positions of proteins are denoted by arrows, and tick marks represent a kilobase of nucleotide sequence. cDNA clones are positioned on the genome as determined by sequencing of the termini. The vector used for all clones was pGEM3.

inoculum was removed, Eagle’s minimal essential medium containing 5% fetal calf serum and 5 hg/ml Actinomycin D was added, and the cells were incubated at the appropriate temperature. After 2 hr at the appropriate temperature 20 &i/ml [3H]uridine was added to the cultures. At the times indicated, the cells were washed twice in TSM buffer (150 mNI NaCI, 10 mM Tris-HCI, pH 7.5, 5 mM MgCI,) followed by the addition of TSM containing 0.2% NP-40. The cells were scraped from the plates and nuclei were removed by low speed centrifugation. SDS was added to the supernatant to a final concentration of 1% and nucleic acids were extracted by phenol/chloroform. Labeled RNA was detected by precipitation with Na,HPO, on DE-81 filters followed by liquid scintillation counting.

and dideoxy sequencing. Previous attempts to sequence the 5’terminus by direct dideoxy sequencing of the RNA produced strong stops 2 to 3 nucleotides from the end, preventing the unambiguous identification of the 5’ end (Faragher et a/., 1988; our own unpublished data). Resolution of the terminal sequences indicates that the RR 5’sequence for nucleotides 1 through 8 is identical to that found in Semliki Forest virus. In contrast, all other sequenced alphaviruses, including Middleburg virus which belongs to the Semliki Forest subgroup, contain an additional nucleotide at position 3. Using this sequence information, we synthesized a short stretch of double-stranded DNAfrom paired complementary oligonucleotides, which contained the authentic Ej’terminus of the RRgenome downstream from

RESULTS Sequence determination of RRV-T48 cDNA and construction of a full-length clone A library of RR cDNA clones in pGEM3 was obtained and characterized by restriction mapping and by sequencing the ends of the inserts. The subclones shown in Fig. 1 were sequentially assembled into a single plasmid representing all but the first 36 nucleotides of RR. To ensure that no mistakes were introduced during cDNA synthesis and/or construction of this clone, it was sequenced by the dideoxynucleotide method, either directly or after subcloning into either M 13mpl8 or M 13mpl9. Several nucleotide differences were found from the sequence reported previously for RR strain T48 (Dalgarno et a/., 1983; Faragher et al., 1988; Strauss et al., 1988) and several ambiguous nucleotides identified, and these are listed in Table 3 (numbering according to our sequence of 1 1,851 nucleotides for RR). In particular, we were able to determine the 5’terminal sequence of RRas AUGGC by tailing the cDNA representing the 5’ terminal sequence followed by polymerase chain amplification

TABLE 3 NUCLEOTIDEDIFFERENCESBETWEENTHE RR cDNA CLONE AND THE PUBLISHEDRR T48 SEQUENCE

Nucleotidea

Nucleotideb

Change

Amino acid substitution

Cap 1 55 537 538 788 805 1146 2593 8882

1 2 56 539 540 790 807 1148 2595 8884

N to Cap N to A deletion of a C N to T N to C N to T N toT A to C T to C YtoT

NC” NC x to Y Xto R x to v x to s No change Y to Hd No change

-

’ Nucleotide numbering according to the sequence of the RR cDNA clone (pRR64). b Nucleotide numbering according to the sequence of Faragher et al. ’ Noncoding region. dThis amino acid substitution is found in the attenuated RR Nelson Bay strain 5092.

434

KUHN ET AL. TABLE 4 RNAs

SPECIFICINFECTIVITYOFTRANSFECTED

Conditions PRR~~~ plus: Complete m’GWbwW)A No cap analog DNase I before transcription DNase I after transcription RNase A after transcription RR T48 viral RNA pToto52b

Infectivity

(PFU/rg RNA)

1.7 x lo3 1.3 x 103 0 0 0.8 X lo3 0 2.0 x lo5 2.0 x lo4

a Determined by titration of RNA using the DEAE-dextran transfection procedure. b Complete SP6 transcription reaction including m7G(5’)ppp(53G unless otherwise indicated.

a SP6 transcription promoter and restriction sites at both ends appropriate for cloning. This synthetic adaptor was ligated to the remainder of the genome to generate the full-length RR cDNA clone, pRR64. This plasmid contains a unique Sacl restriction site downstream of the poly(A) tract to ensure that run-off transcripts would terminate close to the authentic 3’ end of the genome. The 5’ end of transcripts synthesized in vitro from pRR64 contains an additional 5’ G, which corresponds to the major transcriptional start site for SP6 RNA polymerase (Kang and Wu, 1987). Infectious transcripts of Ross River cDNA clones Following linearization of pRR64 with Sacl, RNA was transcribed in vitro using SP6 polymerase in the presence of a cap analog (either m7G(5’)ppp(5’)G or m7G(5’)ppp(5’)A). The proportion of full-length RNAs was similar to that obtained by transcription of SIN pToto52 transcribed under identical conditions. The RNA transcripts were then added to BHK cells which were previously treated with DEAE-dextran or were added in the presence of cationic liposomes (lipofectin). Following a 30-to 60-min incubation at room temperature for DEAE-dextran or a 2- to 4-hr incubation at 37” for lipofectin, the RNA inoculum was removed and an agarose overlay was added. After 2 to 3 days plaques whose morphology was identical to plaques produced by transfection of T48 virion RNA were visible without staining. As shown in Table 4, however, the specific infectivity of pRR64 RNA is about 12-fold less than the specific infectivity of the pToto52 RNA and about 120-fold less than T48 RR virion RNA. The reason for the decreased efficiency is not clear. Virus stocks generated from transfection-derived virus produced plaques which were identical in mor-

phology to those produced by the parental T48 strain in various cell lines tested. In BHK and Vero cells, the pRR64-derived virus and the wild-type T48 virus produced indistinguishable growth curves (data not shown, but see also below). In addition, biological testing of the rescued virus in 1-week-old mice produced a disease which was temporally and pathologically identical to the disease produced by the T48 parental strain (D. E. Griffin, R.J.K.,J.H.S., unpublished data). Thus although the specific infectivity of pRR64 RNA is low, the resulting virus appears to be identical to or at least very similar in its properties to T48. Modifications to the transcription 5’ end of the RR genome

promoter and the

Prior to the construction of pRR64, we had synthesized a full-length RR clone containing the 5’ terminal sequence AUUGGCGG deduced from the sequence of Faragher et al. (1988) which is identical to the SIN 5’ end for the first 8 nucleotides. This clone was positioned downstream of a T7 transcription promoter and two G residues to promote transcription initiation (see Fig. 2). This full-length clone (pRR34) yielded infectious virus which, by growth analysis, was identical to the wild-type T48 virus. The specific infectivity of the RNA, however, was quite low (3-5 X 1O2PFU/pg RNA) compared to the infectivity of the pToto clones of SIN (for comparison see Table 4). In an attempt to increase the specific infectivity of the RNA, one of the two G residues used to promote transcription initiation was deleted (clone pRR38, Fig. 2) but the specific infectivity of this transcript was only twofold higher. 5’primer exten-

+I I Genome

m’Gppp A

RNA

Direct Sequencing of Genome RNA

Pm

5 10 I I U G G C G G A C G

NN-------.

T7 pmmoter

r’ mTGpppGGAU-

-

-

-

-

-

-

-

-...

-

-

-

-

-

-

-

- .

Tl promoter

Pm

SP6 promoter

I+ m?Gppp

G A -

FIG. 2. Modifications of the 5’ end of the RR full-length cDNA. The sequence of the first 10 nucleotides of the genome RNA of RR is shown on the top line. The first nucleotide, A, is linked through a 5’-5’triphosphate to the m7G cap. Direct sequencing of the genome RNA by Faragher et a/. (1988) is shown underneath with the first two ambiguous positions denoted by N. pRR34, pRR38, and pRR64 are full-length clones of RR which differ in their transcription promoter, the number of G residues used to promote transcription initiation, and the sequence corresponding to the first nucleotide of the RR genome. All three clones produced infectious RNA.

ROSS RIVER AND SINDBIS CHIMERIC

sion products derived from RNA isolated from cells infected by pRR38 RNA or T48 RNA comigrated on a denaturing polyacrylamide gel, suggesting that extra nucleotides were processed from the 5’terminus of the pRR38 transcript RNA (data not shown). We assume that the first 2 nucleotides of pRR38 (GA) are removed and that this results in the substitution of U for A at nucleotide 1 of the genome. Since the specific infectivity of the transcript RNA is low, however, and this virus behaves like T48 in growth experiments, it is possible that the virus arising from the transfection mix also has the 5’ terminal A corrected. pRR64 differs from pRR38 by the replacement of the T7 transcription promoter with the SP6 transcription promoter and by deletion of the extra U (Fig. 2). This change in promoter did not seem to affect the quantity or quality of full-length RNA transcripts as determined by electrophoresis in agarose gels. Although the specific infectivity increased by a factor of 3, it is not possible to correlate this increase with the change in promoter since the 5’ end of the viral RNA was also affected. Construction of chimeric River and Sindbis viruses

cDNAs

between

Ross

Previous studies have identified four conserved nucleotide sequence elements which are essential for virus growth (Ou et a/., 1982a,b, 1983; Grakoui et a/., 1989; Levis et a/., 1990; Kuhn et al., 1990; Niesters and Strauss, 1990a,b). Two of these elements are located in the 5’ and 3’ NTRs, respectively. The 3’ 1gnucleotide conserved sequence element is highly conserved among the alphaviruses. SIN and RR, however, appear to have the least similarity, with RR having 2 nucleotide substitutions and 4 nucleotide insertions relative to the SIN sequence. Upstream of this conserved element their sequences are completely unrelated and the lengths of the NTRs are quite different: 322 nucleotides for SIN and 524 nucleotides for RR. The sequences in the 5’ NTRs are not nearly as conserved as the 19-nucleotide conserved element. As with the 3’ NTRs, the 5’ NTRs of the two viruses differ in their lengths: SIN has a 59nucleotide NTR and RR has a 78-nucleotide region. Interestingly, the 5’ terminal secondary structures which can be drawn for the RNAs appear to be significantly different. The proposed hairpin structure at the 5’ end of SIN is 44 nucleotides in length compared with a 29-nucleotide hairpin structure proposed for RR. Experiments performed by Niesters and Strauss (1990a) suggest that the formation of the 5’ hairpin structure in SIN is required for normal growth since mutations which destabilized the stem structure produced viruses that were defective in growth.

VIRUSES

435

We wished to study the functional significance of the 5’ and 3’ NTRs of these two distantly related alphaviruses. By exchanging the NTRs we could determine whether these sequences could function as independent units of the viral genome. To facilitate reciprocal exchanges of the NTRs, several new restriction sites were inserted within the RR cDNA or within the SIN cDNA by site-directed mutagenesis. These cDNA clones were summarized in Table 1 and the chimeric viral cDNAs made with these clones were described in Table 2. Chimeras in the 5’ NTRs were designed such that the junction would occur 5 nucleotides upstream of the AUG codon for the nonstructural polyprotein so as to not interfere with the context for translation initiation. Chimeras in the 3’ NTRs were constructed such that the entire 3’ NTR of the heterologous virus was placed immediately downstream of the termination codon for the structural polyprotein. Four possible combinations of chimeric cDNAs which exchange the 5’ and 3’ NTRs were generated, RNA was transcribed and transfected into BHK cells, and virus was recovered from all four (Table 2). Stocks were generated from plaque-purified chimeric viruses and used for experiments described below. The identity of the chimeric viruses was confirmed by dideoxy sequencing of cDNA across the junction of the two viral sequences. These data indicate that nucleotides 1 to 74 of the RR genome are sufficient to confer replication competency to a SIN genome lacking its homologous 5’ NTR. The reciprocal exchange also produces infectious virus. At the 3’ end of the genome, exchange of the 524-nucleotide 3’ NTR of RR for the smaller 322-nucleotide 3’ NTR of SIN produces a replication competent virus, as does the reciprocal exchange. The results suggest that a heterologous NTR (either at the 5’ or 3’end) which replaces the original NTR of a virus can provide sufficient primary and/or secondary sequences to support replication of the chimeric virus even though the NTRs differ in both size and sequence. Another chimeric plasmid was constructed which contained about half of the RR 5’ NTR (nucleotides 1 to 36) and from nucleotide 49 through to the poly(A) tail of SIN. This clone, p5’RR36/SIN, failed to produce virus, even after numerous transfection attempts using lipofectin, which increases the specific infectivity of other RR transcripts by lo- to 20-fold (data not shown). ln vitro translation of p5’RR36/SIN RNA suggested that the lethal lesion had no effect on protein translation (data not shown). Niesters and Strauss have shown previously that deletions in SIN between nucleotide 41 and the translation start site at nucleotide 60 had little effect on the replication competency of the resulting

436

KUHN ET AL.

A.Vero

B. Mosquito -8

RR RR I 3’ SIN

6

8

10

12

4

Hours pi

6

8

10

12

Hours pi

SIN SIN /3’ RR

Hours pi

Hours pi

FIG. 3. One-step differential growth curves for the 3’ NTR chimeras. Growth curve experiments were performed as described under Materials and Methods. The 3’ chimeras and their parents were tested in Vero cells at 37” (A and C) and in C6/36 mosquito cells at 30” (B and D). SIN refers to the wild-type SIN virus rescued from cells transfected with pToto52 RNA; RR refers to the wild-type T48 strain of RR rescued from cells transfected with pRR64 RNA. Symbols for the viruses are shown in the corresponding legends to the right of the graphs. Results are expressed as log (PFU/ml/hr) released as measured by plaque assays on Vero monolayers at 37” and are the results of two independent growth curve exoeriments.

viruses (Niesters and Strauss, 1990a). Therefore, the present result suggests that an important domain within the 5’ NTR of RR has been disrupted and is the cause of the lethal lesion. A stem and loop structure can be formed at the 5’ end of the RR genome, beginning at nucleotide 3 and ending at nucleotide 29, 7 nucleotides upstream of the RR/SIN junction in p5’RR36/SIN. Growth properties of the chimeric viruses To test the replication efficiency of the chimeric viruses relative to the parental viruses, one-step differential growth curves were performed in which the rate of virus release at different times after infection was determined. Such growth rates are very sensitive to differences in replication efficiency. Growth curves were carried out in two cell lines that support both SIN and RR: Vero cells (a mammalian cell line) and C6/36

mosquito cells. We have previously found that mutants in the 5’ and 3’ conserved regions of SIN often show more severe defects in growth when assayed in mosquito cells (Kuhn et al., 1990; Niesters and Strauss, 1990a,b). Growth curves for the 3’ NTR chimeras in Vero cells are shown in Figs. 3A and 3C, along with the T48 RR (derived from pRR64 transcripts) and HRSP SIN (derived from pToto52 transcripts) parental viruses (the parent virus of a chimera will refer to the virus from which the coding sequence is derived). RR and SIN grew at about the same rate throughout the infection cycle, with SIN reaching a slightly higher rate at 12 hr after infection. The chimeric viruses grew at rates that were essentially indistinguishable from one another and their parents. Figures 3B and 3D show the growth rates in C6/36 mosquito cells. Interestingly, RR and SIN grew at sub-

ROSS RIVER AND SINDBIS CHIMERIC

437

VIRUSES

A. Vero 7-

-Q-a--

2

4

6

8

10

12

4

Hours pi

6

8

10

SIN 5’RR/SIN RR 5’SIN/RR

12

Hours pi

FIG. 4. One-step differential growth curves for the 5’ NTR chimeras. Growth curve experiments were performed as described under Materials and Methods. The 5’ chimeras and their parents were tested in Vero cells at 37” (A) and in C6/36 mosquito cells at 30” (5). SIN refers to the wild-type SIN virus rescued from cells transfected with pToto52 RNA; RR refers to the wild-type T48 strain of RR rescued from cells transfected with pRR64 RNA. Symbols for the viruses are shown in the corresponding legends to the right of the graphs.

stantially different rates. RR had a shorter lag period and late in infection released virus at a rate about 50fold higher than SIN. The reason for this differential growth is not known. In these cells the 3’NTR chimeras grew at rates that were identical to their respective parental virus. These results show that the 3’ NTR is not involved in the differential growth rate observed for SIN and RR in mosquito cells. In Fig. 4A the chimeras resulting from exchange of the 5’ NTRs were examined in Vero cells. In contrast to the 3’ NTR chimeras, both 5’ NTR chimeras show a reduced growth rate when compared to either parent. 5’SIN/RR has a lag in virus production compared to RR. At later times, the rate of virus release begins to approach the parental RR level although never quite reaching it. 5’RR/SIN exhibits a more severe defect in growth, with a 3-hr lag$ in virus production relative to SIN and a final rate of virus release approximately 30fold lower than that of SIN. Figure 4B shows growth curves for the 5’ NTR chimeras in mosquito C6/36 cells. As noted above, RR grows at a higher rate than SIN in mosquito cells. In these cells, 5’SIN/RR was greatly delayed in growth, with a lag 3 to 4 hr longer than SIN or RR. At 12 hr after infection it produced virus at about 10% the rate for SIN and l/300 the rate for RR. The rate of virus release for 5’RR/SIN in mosquito cells was extremely low and did not change significantly during the infection cycle. At 12 hr after infection it produced virus at about l/3000 the rate for SIN and less than 10e5 the rate of RR. These results show that the replacement of the RR 5’ NTR with the corresponding region of SIN produced a viable virus which had a reduced growth capacity. This defect in growth is accentuated when examined in

mosquito cells. The reciprocal virus, 5’RR/SIN, was also viable and had a reduced growth capacity in Vero cells. In mosquito cells this virus was severely crippled, suggesting that the alteration in the 5’ end of the RNA disrupted the normal contribution of the host cell required for virus proliferation. In contrast, the chimeras which exchanged the 3’NTRs of SIN and RR were indistinguishable in growth from their parents in both Vero and mosquito cells and therefore the 3’ NTRs from these viruses represent independent functional units of the viral genomes.

RNA synthesis by the chimeric viruses In order to further examine the functions of the NTRs in virus replication, the synthesis of RNA by the chimerit viruses was examined. In this set of experiments we have examined the total incorporation of [3H]uridine. In contrast to the one-step differential growth experiments presented above, the RNA synthesis data represent cumulative incorporation. At 2 hr postinfection, cells were labeled with [3H]uridine and incubated until harvesting at which time the monolayers were lysed and incorporation assayed. Whereas the rates of virus release for SIN and RR in Vero cells were about equal throughout the 12-hr growth curve (Figs. 3A, 3C, and 4A), the quantity of RNA synthesized by the two parental viruses was quite different (Figs. 5A and 6A). SIN-infected Vero cells synthesized approximately 30fold more RNA than did RR-infected cells. The RNA synthesis by the chimeras containing exchanges in the 3’ NTRs deviated from the parental viruses in a hostdependent manner (Figs. 5A and 5B), even though both chimeras produced growth patterns which were

438

KUHN ET AL.

I

-

RR

--D-

RR13’SIN

--e

SIN

-Q-

SIN/3’RR

u-

I’ 4 4

I 6

, a

10

12

Hours pi

4

6

a

10

12

Hours pi

FIG. 5. RNA synthesis by the 3’ NTR chimeras. RNA synthesis experiments were carried out as described under Materials and Methods. The 3 chimeras and their parents were tested in Vero cells at 37” (A) and in C6/36 mosquito cells at 30” (6). Symbols for the viruses are shown in the corresponding legends to the right of the graphs. Cells were labeled beginning at 2 hr postinfection, samples were harvested at the appropriate times, and incorporation was measured by liquid scintillation counting. Results are expressed as log (cpm of [3H]uridine incorporated).

identical to those of the parental viruses (Figs. 3A, 3C, and 4A). In Vero cells the RNA synthesized by SIN/ 3’RR, which contains the coding sequences of SIN, approximates the amount synthesized by SIN and therefore the 3’ NTR of RR functions as well in SIN as the SIN 3’ NTR does. The reciprocal virus, RR/3’SlN, had a lower incorporation of RNA at early times but caught up to RR at later times. This defect did not appear when virus growth alone was assayed (Fig. 3). Analysis of RNA synthesis in mosquito cells showed a different result (Fig. 5B). Once again there was a significant difference in the levels of RNA synthesized by the parental viruses. From Fig. 3B, the growth rate for RR was approximately 50-fold higher than the SIN rate in mosquito cells and this difference was reflected, although not quantitatively, in the levels of [3H]uridine incorporation. Likewise, the chimera containing the coding sequence of RR and the 3’ NTR of SIN (RR/ 3’SIN) produced an equivalent amount of RNA as did RR and significantly more than SIN. The SIN/3’RR chimera, however, showed a reduction in the amount of RNA relative to SIN synthesized at late times in infection. The reason for this differential effect dependent upon the host cell is obscure, but may reflect the affinity of host proteins for the viral replication complex. Figure 6A presents the RNA synthesis data for the 5’ NTR chimeras and their parents in Vero cells. For the chimera 5’SIN/RR, there was a lower amount of RNA synthesized at 4 hr compared with RR. As the infection cycle proceeded the amount of RNA synthesized remained less than the amount synthesized by RRand by 12 hr the chimera had incorporated only one-third the amount of [3H]uridine as had the parent RR. A similar pattern of incorporation was seen with 5’RR/SIN. The

initial incorporation level, which was higher than wildtype RR,was reduced compared to the parental SIN. At 12 hr after infection it had incorporated approximately one-third the amount of [3H]uridine as SIN. The RNA synthesis patterns for the 5’ NTR chimeras and their parents in mosquito cells are shown in Fig. 6B. The chimera 5’SIN/RR synthesized significantly smaller amounts of RNA, and at 12 hr incorporation was approximately sixfold below that of the wild-type parent RR.As expected 5’RR/SIN had a very low level of RNA synthesis in mosquito cells and the labeled RNA did not accumulate within cells. DISCUSSION We have constructed a full-length cDNA clone of Ross River virus which can be transcribed in vitro to produce infectious RNA. The rescued virus is indistinguishable from the T48 parental strain of RR by the criteria of plaque morphology, host range, mortality in mice, and growth rate in BHK and Vero cells. This cDNA clone can be used for the construction of defined mutations and permits RRpathogenesis and replication to be studied on a molecular genetic level. In addition to constructing intratypic chimeras with the attenuated Nelson Bay strain of RRto analyze the determinants of pathogenesis, it is possible to construct intertypic chimeras with other alphaviruses to study functional domains of the genome. The specific infectivity of the RNA transcripts is quite reduced compared with the infectivity of pToto52, which utilizes the identical SP6 transcription promoter. Transcripts from pToto52 are approximately 1O-fold reduced in specific infectivity compared with wild-type

ROSS RIVER AND SINDBIS CHIMERIC

439

VIRUSES

SIN 5’ RR / SIN RR B’SINIRR

d. 4

6

6

10

12

4

Hours pi

6

6

10

12

Hours pi

FIG. 6. RNA synthesis by the 5’ NTR chimeras. RNA synthesis experiments were carried out as described under Materials and Methods. The 5’ chimeras and their oarents were tested in Vero cells at 37” (A) and in C6/36 mosquito cells at 30” (B). Symbols for the viruses are shown in the corresponding legends to the right of the graphs.

SIN RNA, whereas pRR64 transcripts are 120-fold reduced compared with the wild-type T48 RNA. Although this value is low, it is within the range of previously reported infectivities of in vitro transcribed RNAs such as polio virus (20-fold reduced, T7 polymerase, (van der Wet-f et a/., 1986)) and yellow fever virus (20- to 200-fold reduced, SP6 polymerase, (Rice et al., 1989)). The reason for the difference in specific infectivities may be a reflection of the different fidelities with which the SP6 (or T7) polymerase transcribes different RNAs. Previously, we have reported that sequence-dependent errors were introduced into certain pToto transcripts at high frequency using both SP6 and T7 polymerases (Kuhn et a/., 1990). It is also possible that the viral RNA replicases differ in the efficiency in which they can utilize and correct RNA templates that possess extra nucleotides at the 5’ and 3’ ends. Both the 5’ and the 3’ NTRs of alphavirus genomes contain within them short sequences of nucleotides which are highly conserved and have been proposed to function as promoter elements in RNA replication (Ou et a/., 1982a,b, 1983). The conserved sequence element present in the 5’ NTR is not nearly as highly conserved as the one present in the 3’ NTR. Accordingly, the predicted secondary structures for the RNA within this 5’ NTR are quite different for SIN and RR and suggest that the interaction between the RNA and the host and/or virus proteins may be significantly different for the two viruses. Niesters and Strauss (1990a) have found that mutations within the putative hairpin structure encompassing the first 44 nucleotides of the SIN 5’ NTR had deleterious effects on virus replication and in many cases the effects appeared to be host-dependent. For the 3’ NTR, a highly conserved sequence of

19 nucleotides (23 nucleotides for RR, 19 nucleotides for all other sequenced alphaviruses) is found directly upstream of the poly(A) tract. Defined mutations within this sequence element perturb the growth of the virus in a host-dependent manner (Kuhn et al., 1990). For the remainder of the 3’ NTR there appears to be little sequence in common among the alphaviruses. Although there are repeated sequences within the 3’ NTR of individual alphaviruses (Ou et al., 1982b), the sequences themselves are not conserved among all alphaviruses. However, the repeated sequence blocks do show a higher degree of conservation among closely related viruses than sequences immediately surrounding them, suggesting that they may play a ro’le as promoters or enhancing type elements (Ou et a/., 198213; Faragher and Dalgarno, 1986; Y. Shirako, eta/., 1991). The size of the 3’ NTR found in alphaviruses varies considerably from a short 121 nucleotides found in Venezuelan equine encephalitis virus to a substantially larger 524 nucleotides in RR. Interestingly, even among geographic isolates of RR the size of the 3’ NTR varies (Faragher and Dalgarno, 1986). One common feature found in these RR isolates is the presence of a sequence block, which is absolutely conserved and located in the same position relative to the poly(A) tail. In some isolates this sequence block is repeated several times, whereas in others only one copy of the block is found. Deletion of 93% of the SIN 3’ NTR, including three repeated sequence blocks, produced a virus which was growth-deficient in a host-dependent manner (Kuhn et al., 1990). These results suggest that the entire 3’ NTR is important for virus replication and that some elements may be utilized only in specific hosts.

440

KUHN ET AL.

The construction of chimeric viruses to study cisand trans-acting functions has been previously carried out in a number of systems (Semler et al., 1986; Johnson and Semler, 1988; Traynor and Ahlquist, 1990). In the present study, we have found that exchanges involving the 5’ and 3’ NTRs of SIN and RR gave rise to chimeric viruses which grew surprisingly well. The most general interpretation of these results suggests that the alphavirus genome exhibits a high degree of plasticity, tolerating not only numerous nucleotide substitutions but also substantial differences in the sizes of the 5’ and 3’ NTRs. However, the results obtained clearly indicate that the 5’ and 3’ NTRs do not function in the same manner since the chimeras exchanging the 5’ and 3’ ends are clearly distinguishable. Exchange of the 3’ NTR of either RRor SIN with the 3’ NTR of the other produced chimeric viruses which had little or no growth defect in either Vero or mosquito cells as assayed by either the rate of virus production or the level of RNA synthesis. Although the 3’ conserved sequence element is still present, there are substantial changes between the RR sequence element and the SIN sequence. The fact that these exchanges change the sequence of the conserved elements and produce wild-type viruses contrasts to previous studies on SIN where it was found that single substitutions or deletions within the 3’ conserved sequence element usually had a rather significant effect on the ability of the virus to grow in either mosquito cells or chick embryo fibroblasts (Kuhn et al., 1990). Therefore we conclude that the 3’ NTRs can function as independent units, possibly by interacting directly with host proteins. In contrast to the results at the 3’end of the genome, exchange of the 5’ NTRs of SIN and RR produced chimericviruses which were defective in growth. RNA synthesis for both chimeras was significantly reduced in Vero cells, suggesting that this may be a primary defect of the viruses. In mosquito cells RNA synthesis was greatly reduced as was the production of virus, especially for 5’RR/SIN which could barely replicate. Therefore, in these chimeric viruses, host proteins and possibly viral proteins may still interact with the putative secondary structures but at a greatly reduced rate. The construction of chimeric genomes between RR and SIN has resulted in a series of viruses which display a broad range of phenotypes. The ability to rescue viruses, which in some cases replicate quite efficiently, from these chimeric constructs suggests that in nature the alphaviruses may evolve rapidly by the exchange of units which may function in a semi-independent manner. In addition to studying possible evolutionary pathways for alphaviruses, the chimeric viruses provide an opportunity to dissect specific regions of the genome

such as the 5’ and 3’ NTRs and to probe the mechanisms by which they operate. ACKNOWLEDGMENTS We thank Frank Preugschat and Ellen Strauss for stimulating discussions. We also thank Ellen Strauss for critical reading of the manuscript. This work was supported by Grant DMB-8617372 from the National Science Foundation and by Grants Al10793 and Al2071 2 from the National Institutes of Health. R.J.K.was a recipient of a U.S. Public Health Setvice Individual National Research Service Award (A107869) from the National Institutes of Health. H.G.M.N. was supported in part by a fellowship grant from the Niels Stensen Foundation, a Gosney Fellowship from CIT, and a travel grant from The Netherlands Organization for the Advancement of Pure Research.

REFERENCES AASKOV,J. G., MATAIKA,J. U., LAWRENCE, G. W., RABUKAWAGA, V., TACKER,M. M., MILES, 1. A. R., and DAGLISH,D. A. (1981). An epidemic of Ross River virus infection in Fiji, 1979. Am. 1. Trap. Med. Hyg. 30, 1053-l 059. BERGE,T. 0. (1975). “International Catalogue of Arboviruses including certain other viruses of vertebrates,” Second ed. U.S. Department of Health Education and Welfare, Bethesda, MD. CHEN, E., and SEEBURG.P. H. (1985). Supercoil sequencing: A fast and simple method for sequencing plasmid DNA. DNA 4, 16% 170. DALGARNO, L., RICE,C. M., and STRAUSS, J. H. (1983). Ross River virus 265 RNA: Complete nucleotide sequence and deduced sequence of the encoded structural proteins. virology 129, 170-l 87. DOHERTY,R. L., GORMAN,B. M., WHITEHEAD, R. H., and CARLEY,J. G. (1964). Studies of epidemic polyarthritis: The significance of three group-A arboviruses isolated from mosquitoes in Queensland. Aust. Annu. Med. 13, 322-327. DOHERTY,R. L., WHITEHEAD,R. H., GORMAN,B. M., and O’GOWER,A. K. (1963). The isolation of a third group A arbovirus in Australia, with preliminary observations on its relationships to epidemic polyarthritis. Aust. 1. Sci. 26, 183-l 84. FAW\GHER, S. G., and DALGARNO,L. (1986). Regions of conservation and divergence in the 3’untranslated sequences of genomic RNA from Ross River virus isolates. J. Mol. Biol. 190, 141-l 48. FAFIAGHER, S. G., MEEK, A. D. J., RICE, C. M., and DALGARNO,L. (1988). Genome sequences of a mouse-avirulent and a mouse-virulent strain of Ross River virus. Virology 163, 509-526. GRAKOUI, A., LEVIS,R., RAJU,R.. HUANG.H. V.. and RICE,C. M. (1989). A &-acting mutation in the Sindbis virus junction region which affects subgenomic RNA synthesis. J. Viral. 63, 5216-5227. JOHNSON,V. H., and SEMLER,B. L. (1988). Defined recombinants of poliovirus and coxsackievirus: Sequence-specific deletions and functional substitutions in the 5’-noncoding regions of viral RNAs. Virology 162, 47-57. KANG,C., and Wu, C.-W. (1987). Studies on the SP6 promoter using a new plasmid vector that allows gene insertion at the transcription initiation site. Nucleic Acids Res. 15, 2279-2294. KAY,B. H., and AASKOV,J. G. (1989). Ross River virus (Epidemic Polyarthritis). In “The Abroviruses: Epidemiology and Ecology” (T. P. Monath, Ed.), pp. 93-l 12. CRC Press, Boca Raton, FL. KUHN,R., HONG,Z., and STRAUSS, J. H. (1990). Mutagenesis of the 3 nontranslated region of Sindbis virus RNA. 1. Viral. 64, 14651476. KUNKEL,T. A. (1985). Rapid and efficient site-specific mutagenesis

ROSS RIVER AND SINDBIS CHIMERIC without phenotypic selection. Proc. Nat/. Acad. Sci. USA 82,488492. LEVIS, R., SCHLESINGER,S., and HUANG, H. V. (1990). The promoter for Sindbis virus RNA-dependent subgenomic RNA transcription. /. Viol. 64, 1726-l 733. MARSHALL, I. D., and MILES, J. A. R. (1984). Ross River virus and epidemic polyarthritis. In “Current TOPICS in Vector Research” (K. F. Harris, Ed.), pp. 31-56. Praeger, New York. M+w+M, A. M., and GILBERT,W. (1980). Sequencing end-labeled DNA with base-specific chemical cleavage. In “Methods in Enzymology” (L. Grossman and K. Moldave, Eds.), Vol. 65, pp. 499-560. Academic Press, San Diego. NIESTERS,H. G. M.. and STRAUSS,1. H. (1990a). Defined mutations In the 5’ nontranslated sequence of Sindbis virus RNA. 1. viral. 64, 4162-4168. NIESTERS,H. G. M., and STRAUSS,1. H. (1990b). Mutagenesis of the conserved 5 1 nucleotide region of Sindbis virus. J. Viral. 64, 16391647. Ou, J.-H., RICE, C. M., DALGARNO, L., STRAUSS, E. G., and STRAUSS, J. H. (1982a). Sequence studies of several alphavirus genomic RNAs In the region containing the start of the subgenomic RNA. Proc. Nat/. Acad. Sci. USA 79, 5235-5239. Ou, J.-H., STRAUSS, E. G., and STRAUSS, J. H. (1981). Comparative studies of the 3’.terminal sequences of several alphavirus RNAs. Virology 109, 281-289. Ou, J.-H., STRAUSS, E. G., and STRAUSS.J. H. (1983). The 5’.terminal sequences of the genomic RNAs of several alphaviruses. J. Mol. Biol. 168, 1-15. Ou, J.-H., TRENT, D. W., and STRAUSS,J. H. (1982b). The 3’.non-codIng regions of alphavirus RNAs contain repeating sequences. /. Mol. Biol. 156, 7 19-730. RICE, C. M., GRAKOUI, A., GALLER. R., and CHAMBERS, T. J. (1989). Transcription of infectious yellow fever RNA from full-length cDNA templates produced by in vitro ligation. New Biol. 1, 285-296. RICE, C. M., LEVIS, R., STRAUSS,J. H., and HUANG, H. V. (1987). Pro-

VIRUSES

441

duction of infectious RNA transcripts from Sindbis virus cDNA clones: Mapping of lethal mutations, rescue of a temperature sensitive marker, and in vitro mutagenesis to generate defined mutants. J. viral. 61, 3809-3819. SANGER, F., NICKLEN, S., and COULSON, A. R. (1977). DNA sequencing with chain-terminating inhibitors. Proc. Nat/. Acad. Sci. USA 74, 5463-5467. SEMLER, B. L., JOHNSON, V. H., and TRACY, S. (1986). A chlmeric plasmid from cDNA clones of poliovirus and coxsackievirus produces a recombinant virus that is temperature-sensitive. Proc. Natl. Acad. Sci. USA 83, 1777-l 781. SHIRAKO, Y., NIKLASSON, B., DALRYMPLE, J., STRAUSS, E. G. and STRAUSS,J. H. (1991). Structure of the Ockelbo virus genome and its relationship to other Sindbis viruses. Virology 182, 753-764. STRAUSS, E. G., LEVINSON. R., RICE, C. M., DALRYMPLE, J., and STRAUSS,J. H. (1988). Nonstructural proteins nsP3 and nsP4 of Ross River and O’Nyong-nyong viruses: Sequence and comparison with those of other alphaviruses. Virology 164, 265-274. STRAUSS,E. G., and STRAUSS,J. H. (1986). Structure and replication of the alphavirus genome. In “The Togaviridae and Flaviviridae” (S. Schlesinger and M. 1. Schlesinger, Eds.). pp. 35-90. Plenum, New York. TAYLOR, W. P., and MARSHALL, I. D. (1975). Adaption studies with Ross River virus: Laboratory mice and cell cultures. J. Gen. Viral. 28, 59-72. TRAYNOR, P., and AHLQUIST, P. (1990). Use of bromovirus RNA2 hybrids to map cis- and transacting functions in a conserved RNA replication gene. /. Viral. 64, 69-77. VAN DERWERF, S., BRADLEY,1.. WIMMER, E., STUDIER,F. W., and DUNN, J. J. (1986). Synthesis of infectious poliovirus RNA by purified T7 RNA polymerase. Proc. Nat/. Acad. Sci. USA 83, 2330-2334. ZOLLER, M. J., and SMITH, M. (1984). Laboratory Methods: Oligonucleotide-directed mutagenesis: A simple method using two oligonucleotide primers and a single-stranded DNA template. DNA 3, 477-488.