Proc. Nati. Acad. Sci. USAVol. 89, pp. 2185-2189, March 1992Biochemistry

RNA-dependent RNA polymerase consensus sequence of the L-Adouble-stranded RNA virus: Definition of essential domainsJUAN CARLOS RIBAS AND REED B. WICKNERSection on the Genetics of Simple Eukaryotes, Laboratory of Biochemical Pharmacology, National Institute of Diabetes and Digestive and Kidney Diseases,Building 8, Room 207, National Institutes of Health, Bethesda, MD 20892

Communicated by Herbert Tabor, November 27, 1991 (received for review October 2, 1991)

ABSTRACT The L-A double-stranded RNA virus of Sac-charomyces cerevisiac makes a gag-pol fusion protein by a -1ribosomal frameshift. The pol amino acid sequence includesconsensus patterns typical of the RNA-dependent RNA poly-merases (EC 2.7.7.48) of (+) strand and double-stranded RNAviruses of animals and plants. We have carried out "alanine-scanning mutagenesis" of the region of L-A including the twomost conserved polymerase motifs, SG...T...NT..N (. = anyamino acid) and GDD. By constructing and analyzing 46different mutations in and around the RNA polymerase con-sensus regions, we have precisely dermed the extent of domainsand specific residues essential for viral replication. Assumingthat this highly conserved region has a common secondarystructure among different viruses, we predict a largely fl-sheetstructure.

The (+) strand RNA viruses of animals and plants shareamino acid sequence patterns in their RNA-dependent RNApolymerase (EC 2.7.7.48) regions (1-3) (Fig. 1). The samepatterns are present in most of the dsRNA viruses whosepolymerase segment has been sequenced, including the L-Avirus of Saccharomyces cerevisiae (4), reovirus (5), blue-tongue virus (6), and rotavirus (7). Two other dsRNA viruses[46 phage (8) and infectious bursal disease virus (9)] show aless clear fit to the pattern. The extent of these commonsequence patterns suggests that there is a common structureand function among the more than 50 viral enzymes for whichsequence data are available and that information obtainedabout one of them may be applicable to others. Whiledifferent enzymes must recognize different sites on viralRNA and interact with different viral and host proteins, thereare also common functions which they must carry out,including binding of Mg2' and rNTPs, holding onto thetemplate RNA chain, and the chain elongation reaction itself.It is likely that these conserved sequence patterns are part ofdomains responsible for such common functions.The L-A dsRNA virus of yeast replicates by a conservative

mechanism with (+) and (-) strands made inside the viralparticle at different points in the replication cycle (reviewedin ref. 10). In vitro systems for replication, transcription, andpackaging are available, and the signals for the replicationstep and for packaging have been defined (11-14). L-Aencodes its 70-kDa major coat protein (called gag) (4, 15) anda 170-kDa gag-pol fusion protein (4, 16) formed by a -1ribosomal frameshift indistinguishable in mechanism fromthat used by retroviruses for a similar purpose (17). The polregion of the gag-pol fusion protein (4) has all the character-istic sequence patterns identified by Kamer and Argos (1) astypical of (+) strand RNA viral RNA-dependent RNA poly-merases.The M1 satellite virus of L-A encodes a protein, the killer

toxin, that is secreted by cells carrying M1 and kills cells

lacking M1 (reviewed in refs. 10 and 18). M1 depends on L-Afor its coat and replication proteins (19). MAK10 is one ofthree chromosomal genes needed for L-A virus propagationwithin yeast cells (20). In a maklO host, L-A proteinsexpressed from a cDNA clone of L-A support the replicationof the M1 satellite virus but (for unknown reasons) do notsupport propagation of the L-A virus itself (21). Thus, whileL-A requires the MAK10 product itself, M1 requires MAK10only because it requires the L-A-encoded proteins. We haveused this phenomenon to assay the importance for viralpropagation of specific sequences encoded by L-A. Previouswork to examine the importance of the amino acid patternsconserved among RNA-dependent RNA polymerases hasdealt with the enzymes encoded by Qf phage, poliovirus, andbrome mosaic virus (22-24).We have defined the regions surrounding the two most

highly conserved RNA polymerase consensus patterns thatare necessary for viral propagation. We show that, althoughthese domains are highly conserved among a broad range ofviruses in both primary structure and predicted secondarystructure, they are not interchangeable.

MATERIALS AND METHODSStrains and Media. YPAD, YPG, 4.7MB, SD, and synthetic

complete medium (25) and LB medium (26) have beendescribed. S. cerevisiae strains 2955p0 (MATa trpl adel his3maklO-1 L-A-o M-o p0), 2629 (MATa leul karl-i L-A-HNBM1), and 5X47 (MATa/MATa hisl /+ trpl/+ ura3/+ [KIL-o]) were used. DNA sequencing was done by the dideoxymethod of Sanger et al. (27) with deoxyadenosine 5'-[a-[35S]thio]triphosphate, using a Sequenase kit [United StatesBiochemical (28)].

Site-Directed Mutagenesis. Mutagenesis of the L-A cDNAexpression plasmid pI2L2 (21) was carried out as describedby Kunkel (29), using the Muta-Gene kit from Bio-Rad. Allmutations were confirmed by sequencing. To ensure that lossof activity was due to the introduced mutation and not tochanges elsewhere in the L-A sequence or vector, two orthree mutants were checked for each change, and 16 wild-type clones (nonmutant at the site where mutagenesis wasattempted) isolated during attempts to make mutations weretested. All mutants of a given type were either all inactive orall active, and all of the wild-type clones isolated were active.This approach may be ofgeneral utility in extensive localizedmutagenesis projects when complete resequencing of eachmutant isolate or insertion of mutagenized segments intounmutagenized vectors would be impractical.

Substitution of Homologous Domains of Other Viruses byAsymmetric PCR. Fragments of 81, 105, and 225 nucleotidesfrom reovirus segment Li (5) and of 87, 159, and 180nucleotides from Sindbis virus (30) were amplified in a PCRprocess as described (31), in two steps using the Gene Amp

Abbreviations: ssRNA and dsRNA, single-stranded and double-stranded RNA; ORF, open reading frame; aa, amino acids.

2185

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Dec

embe

r 21

, 202

0

2186 Biochemistry: Ribas and Wickner

+ Strand RNA Viruses:Carnat 524 GCRNMDNNALGNCLLACLITKH..............LIKIRSRLINNflRCVLITobEch 2580 KGNNEQPSXVVDIMNVIIAMLY .... .... TCEKCGINKEEIVYYVNMDDLLIABromMV 521 FQR~fDAFYFG=LVTNAMIAY............ASDLSDCDCAIFSGDSLIITobMV 1440 YQRPEfDVTTFIGVIIAACLAS....NI........K IKGAFcGfClSLLYAl fal faMV585 FQRM=DALZYAWUIVTLACLCH .... ........ VYDLMDPNFVVASMSLIGCowpeaMV1492 CGIPEFP !IVHIFNEILIRYHYK. 9 aa. ELNVQSFDKLIGLVTYgIDNLISMdlbrg 804 AMNKKW4FLILFVML3F4IAS .... ...... RVLEERLTNSKCAAFIERRNIVHWNileF 3120 DQRGMQVVTYALIFTNIAVQKV.25 aa.RTWLFENGEERLSRMAVSgRRCVVKYellowF 3100 DQRGDIQVVXYALifITNLKVQLI.25 aa.EAWLTEHGCDRLKRNAVSfRCVVRSINDBIS 2327 AMKKMXFLTLFVZVUVVIASRVL..........EERLKTSRCAAFIMflRNIIHDengueV 3088 DQRGEQVG!YGLMFTNEUAQLIRQ. 23 aa.WLARVGRERLSRNAISfIRRCVVKFootMDV 2156 GGMPflSCSATSII=ILNINIYVLYAL ...... RRHYEGVELDTYTNISYRIVVAEMC 2121 GGLICAATSNIflIMNNIIIRAGLY ...... LTYKNFEFDDVKVLSYflRLLVARhinol4 2006 GGMPECSGXSIFIXINNIIIRT ..... LILDAYKGIDLD. KLKILAYflRLIVSHepat A 1191 GSMPMfSPC!AILMIIPNVNLYY... VFSKIFGKSPVFFCQALKILCY MVLIFPolio 2025 GGNPM2CSGZSIFNMNINNLIIRT..... LLLKTYKGIDLD.HLKMIAYgRVIASCoxsac B 300 GGMPIfCSGZSIFIMMINNIIIRT..... LMLKVYKGIDLD.QFRMIAYEDDVIASdURNA Viruses:L-A 544 GTLLQWRLTTFMtVLNWAYMKLAGV.......... FDLDDVQDSVHNflDVMISBTV 715 DTHILENSZLIANIMHXNAIGTLIQRA.... VGREQPGILTFLSEQYVMUfTLFYROTA V 586 GAVAMEKQZKAA]gIAXLULIKTVLSRISN...... KYSFATKIIRVDgBMYAVREOVIRUS 676 TTFPSSTAZSTEUANNSTMMETFLTV... 20 aa. QRNYVCQOfWiGLMIPHI6 390 VGLSMQGAZDlIMg!LLSITYLVMQLD.24 aa.... QGHEEIRQISKARAILGIBDV 476 YGQGUNAA!FINULIETLVLDQWNL... 20 aa. NFKIERSIDDIRGK

kit (Perkin-Elmer/Cetus). The first reaction mixture con-tained 100 mM Tris HCI at pH 8.3, 50 mM KCl, 1.5 mMMgCl2, 0.001% gelatin, each dNTP at 0.2 mM, 2.5 units ofTaq DNA polymerase, 10-4 pmol of CsCl-purified DNA ofSindbis virus clone TotollOl (30) or a clone of reovirussegment Li in pBR322 (5), and 100 pmol each of twosynthetic 55-mer primers (with the 5' 30 bases of eachcomplementary to the L-A flanking sequences and the 3' 25bases of each complementary to the ends of the reovirus orSindbis virus sequence to be amplified). The samples weresubjected to 35 cycles of amplification in a DNA thermalcycler (Perkin-Elmer/Cetus) as follows: 94°C for 1.5 min,37°C for 1.5 min, and 72°C for 2 min. Then, 2.5-10% of theproduct was used as template for a second identical PCRexcept that only the 55-mer primer having the L-A (-) strandsequence was included. This generated an excess of thestrand to be used as a mutagenic oligonucleotide. Then, 40%of the reaction products in 8 ,ul of H20 were 5' phosphory-lated with T4 polynucleotide kinase (BRL) and used forsite-directed mutagenesis of the L-A cDNA expression plas-mid pI2L2 as described above.Assay of Altered L-A Expression Vectors. The L-A expres-

sion vector pI2L2 (21) has the PGKI promoter (32), the L-Asequences from one of our full-length clones (4), the origin ofreplication of the 2-,um DNA plasmid, the yeast TRPI genefor selection in yeast, the fl phage replication origin forproduction of ssDNA, and pBR322 sequences (Fig. 2).Amino acid residues in L-A's pol ORF are numbered startingwith the Arg residue at base 1964, which is the first amino acidafter the -1 ribosomal frameshift (17). The L-A expressionvector pI2L2, specifically mutated in the L-A coding se-quences, was introduced by transformation (33) into yeaststrain 2955p°. Into this strain was then introduced, by cyto-plasmic mixing [cytoduction (34)], M1 and L-A viruses.Because this strain is defective in MAKIO, it is unable tomaintain L-A dsRNA, even though the normal L-A proteinsare supplied from the cDNA vector. M1 is stably maintained,however, and is present in viral particles (21). This provides

L-A

Yeast PGK ORFIPromoter Major Cost Prenin>

I I

FIG. 1. Consensus sequence patterns of viral RNA-dependent RNA polymerases of (+) single-strandedRNA (ssRNA) viruses (1, 3) and double-stranded RNA(dsRNA) viruses (4-9). The region shown here has themost extensive homology among these proteins, butconserved patterns extend well beyond these in bothN-terminal and C-terminal directions (1, 3). The mosthighly conserved residues are shown by asterisks. Theviral sequences shown here are not more homologousto L-A than those of other (+) ssRNA and dsRNAviruses. Carnat, carnation mottle virus; TobEch, to-bacco etch virus; MV, mosaic virus; TobMV, tobaccomosaic virus; Mdlbrg, Middleburg virus; F, fever virus;FootMDV, foot-and-mouth disease virus; EMC, en-cephalomyocarditis; BTV, bluetongue virus; IBDV,infectious bursal disease virus; aa, amino acids. Aminoacid symbols shown in lowercase do not agree with theconsensus.

an in vivo assay of the activity of the proteins encoded by theL-A cDNA clone in the absence of the L-A virus itself. Thestable maintenance of M1 by the L-A cDNA clone requiresboth the major coat protein (gag) and the gag-pol fusionprotein (21).

Semiquantitative Killer Activity Assay. To determine notonly ifa specific L-AcDNA mutation permits M1 propagationbut also different degrees of activity when M1 is ultimatelylost (see Results), we designed a semiquantitative method.The 2955p° transformants were grown for 4 days at 30°C,pooled and grown overnight, mixed in 0.1 ml of water withapproximately equal amounts of the freshly grown donorstrain, 2629, and poured onto a YPAD plate to allow cyto-duction. After 6-8 hr at 30°C, the mating mixture wasstreaked for single colonies on synthetic complete mediumlacking leucine and tryptophan, to select against the donorstrain and for cells carrying the L-A cDNA expressionplasmid. After 4 days, colonies were replica plated to SDmedium, on which only diploids can grow, to YPG, on whichonly diploids and cytoductants can grow, and on 4.7MBplates having a lawn of strain 5X47, to test (3 days, 22°C) forthe maintenance of M1 dsRNA by the killer phenotype. Thepresence or absence ofM1 in the cytoductants is the indicatorof whether the L-A proteins made from the mutant cDNAclone were active or inactive.For each mutant cDNA, 50-100 cytoductants were exam-

ined and their killer zones were estimated as wild-type(100%o), 1/2-size (50%), ¼/4-size (25%), dark blue color aroundthe colony but no clear zone (10%1), or no sign of killing (0%1).The results obtained for different mutants having the sameamino acid change were quite similar.

RESULTSIt has been argued that for scanning a protein with amino acidchanges, mutagenesis to alanine is the most "neutral" changethat one can make (35). This amino acid is neither bulky norparticularly small. It is not a helix-breaker and is neither

ftor

aI ~~ ~~~~~~~II

FIG. 2. The L-A expression vector pI2L2, which was the starting point for the mutations described here. PGK, phosphoglycerate kinase;ORF, open reading frame; ori, replication origin.

RNA polywras.

SG...T...NT..NGDD YestORF2 2 N TRPI,%- 2 ADNAOr! Amp OilRNJ

Proc. Natl. Acad. Sci. USA 89 (1992)

Dow

nloa

ded

by g

uest

on

Dec

embe

r 21

, 202

0

Biochemistry: Ribas and Wickner

particularly hydrophobic nor polar. These ideas have beenencapsulated in the expression, "alanine-scanning mutagen-esis," and we have adopted this approach to simplify the taskof defining essential regions in the L-A pol ORF.L-A pol Domains. We substituted alanine for pairs ofamino

acids or single residues as shown in Fig. 3. The most highlyconserved residues [544SG, 549T, 579GDD (the numberalways indicates the first residue)] were all essential foractivity (<1%), but the less strictly conserved 553NT (whichis NS, HT, or GT in some viruses) was 4% active whenchanged to AA, and 557N (which is T, M, S, . . . in otherviruses) was 17% when changed to A. 554T is often S in otherviruses, and this works in L-A as well, but 557N cannot bechanged to T (Fig. 4B). The eight nonconserved residuesinside the SG...T...NT..N motif were essential, but the ratesof loss of M1 appeared to be slower than for the adjacentconserved residues (except for 557N).We sought to establish borders of putative functional

domains. The SG ...T. .. NT..N domain comprised 22 aminoacids from residue 540 to residue 561, and the 579GDDdomain was 29 amino acids from residue 565 to residue 593.Having localized the limits for both domains, we constructedsingle-residue changes of each couple to ascribe this limit toone specific amino acid. In the couple 54OGT, only the T waslethal, while in 560YM, either alone was lethal. Thus, theSG...T...NT..N domain (domain 1) comprises 21 amino acidsfrom 541T to 561M. Although 565GV -k AA was less than25% active, both 565G -* A and 566V -- A were fully active,suggesting that each is partially necessary. Thus, the GDDdomain is 29 residues from 565G to 593V (domain 2).There are two GDD sequences in the L-A pol ORF, one

starting at residue 579 and the other at residue 707. While anyof the substitutions in the GDD starting at 579 were lethal,there was no effect of 707GDD -> AEE (Fig. 4A). Betweenthe two domains, there are 3 to 5 residues that were notessential for activity (at least 562KLA). Since the number ofresidues between the domains is quite variable among variousviruses, we sought to determine if this distance is critical byinserting 6 alanine residues at this point (Fig. 4C). Thismutation largely, but not completely, inactivated the plas-

Proc. Natl. Acad. Sci. USA 89 (1992) 2187

APlasmidWild-type = pl2L2M4M5

Domain2 Downstream

aa 579 G D D aa 707 G D DA E E G D DG D D A £ F

Domain 1

BWild-type = pl2L2 S GWR L T T F M N T V L NM53 SGWRLTTFMNSVLNM54 SGWRLTTFMNTVL_poliolC S G.QC Q T 2 F NTV L NreolC SG TT I f N T VL Nmdl SGMFLTL F V N T M L N

CM55 LNWADLDD

L-A domain I L-A domain 2AMAAAAA

Activity100%

0.4100

100%100

20.30.42

18%

FIG. 4. (A) Effect of changing 579GDD to AEE and 707GDD toAEE. (B) Changes in domain 1: Effect of changing 554T to S and557N to T and replacement of residues within L-A's RNA polymer-ase SG...T...NT..N motif with homologous domains from poliovirus(polio 1C), reovirus (reolC), and Middleburg virus (mdl). Changedresidues are indicated by underlining. (C) Expansion of the interdo-main region of L-A with six alanine residues inactivates it.

mid, though less dramatically than a conserved residuemutation, indicating that this distance is important.When full killer colonies from 538LQ -- AA and 562KL

AA, and weak killer colonies from 562GV -- AA, 573QDAA, 575SV -- AA, 584IS -* AA, and 586LN -- AA wererestreaked for single colonies on media selecting retention ofthe L-A expression plasmid, all of the colonies from mutantgroups that did not affect the killer activity in the first assayremained full killers, but those from mutant groups withreduced activity in the first assay now had no killer activityat all. Thus, the former maintain M1 as stably as the wild type,and the latter will ultimately completely lose M1, though theymay vary in their rate of its loss.

Effect of Partial and Full Domain Substitutions. We re-placed the L-A domains with homologous, but nonidentical,regions from other viruses. We first substituted the noncon-served residues within the SG...T...NT..N motif with the

Domain laDomain lb

* * * ***

Domain 2

t 1v RM~~~~~~~nX~

-I1

WVLDPDTKEWYRLOGTLLSGWRLTTFMNITVLNWAYMKLAGVFDLDDVODSVHNGDDVMI SLNRVSTAVRIMDAMHRI L-A seauence

_ _ _ _ _ _ _ _ _AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAEEAAAAAAAAAAAAAAAAhAAAAAC mutant7rlt- 3 -IT 4

2 1 2 30LO4251 2X 3 ER 3 1 4 4 0 ° No. of mutants2 2 3 2 2 3 2 4 1 2 2 3 3 3 3 2 3 2 2 2 2 3 3 2 1 1 3 3 2 2 2 6 3 2 3 3 tested

FIG. 3. Alanine-scanning mutagenesis of the L-A RNA polymerase region. The regions necessary to efficiently replicate the satellite virusM1 extend substantially beyond the conserved residues. Percent killer activity reflects the rate of loss of M1. Only mutants with 100%t killeractivity here can stably maintain M1.

100-

>,75-

0)

25-

I..

I Ir

.

1:11

- -----

taa

0 .

-41- .-.ri F-.. 41

Dow

nloa

ded

by g

uest

on

Dec

embe

r 21

, 202

0

2188 Biochemistry: Ribas and Wickner

Reovirus Ll domain 121 aa (+ 3 aa on each end)

MR1 ¶FPfQSTAISTEHIANNSTMMETFYRLQl3TLLMWRLIFMIVLNWAY GV

L-A domainl21 aa

Reovirus LI domain 229 as (+ 3aaon each end)

MR2 LMLRKMIKSL1QRNWCQ GLRG IIDG1rAGK>JNYMKL>GVFDLDDV DSVHNGDDVMISLNRVSTAV IMDA

L-A domain 229 aa

Reovirus Li domains 1 +269 (+3 aa on each end)

YHL TLJ ............................VSTA .I MDAL-A domains 1 + 2

53 aa

MS1a+2

MS1b+2

Sindbis domains la + 26O aa

PTGRFKFGAMMKSI ............SDKEMAqVLIrUJI rWcYLU V..............VtSjAVfIMUAMH

L-A domains Ia +260 aa

Sindbis domains lb + 253 aa

TGTRFKFU MK1% ..................SDKEMIAE

VLDPDTKEWYRIQGMI5.VSTAV,I DAMHL-A domains lb +2

53 aa

Sindbis domain 229 aa

RVLEERLKTSRCMFIGNIIHGWSDKEMAEMS2 12%

YM=GVFDLDDVODSVHNVMISLNRVSTAVRv DAMHL-A domain 2

29 aa

FIG. 5. Substitutions of reovirus (MR) and Sindbis virus (MS)domains for RNA polymerase consensus domains in L-A and effecton activity. wt, Wild type.

corresponding residues from poliovirus, reovirus, and Mid-dleburg virus (Fig. 4B). None of these substitutions left anysubstantial activity. We then constructed substitution mu-tants in which all of domain 1, all of domain 2, or both weresubstituted from homologous regions of other viruses intoL-A to determine if they were interchangeable. While theborders of domain 2 in Fig. 3 seem clear, the left border ofdomain 1 appeared ambiguous, and so in some experiments,we made substitutions of domain la and of lb (MS). We alsoreplaced residues that did not seem to be essential, at the

border ofa domain in L-A, with residues from the other viruswhich might be involved in the function ofits (possibly larger)domain (MS). In some cases, we inserted extra amino acidresidues from the other virus at the borders of the L-Adomains (MR). In none of these substitutions did we obtainsubstantial activity (Fig. 5). For reovirus the distance be-tween its SG...T...NT..N and GDD motifs is 16 amino acidresidues greater than for L-A. If this distance were important,as it seems to be for L-A, then we may have failed to seeactivity for this reason. Sindbis virus has the same distancebetween the SG..T. ..NT..N and GDD motifs as does L-A.The failure of these substitutions to work may be becauseinteractions among more than just the two domains arecritical or because one or both domains are involved insite-specific binding to the template.Consensus Secondary Structures. The presence of consen-

sus patterns for viral RNA-dependent RNA polymerases insuch a large group of animal, plant, and fungal virusesindicates that these enzymes must have a common secondarystructure in these regions. Although the secondary structureprediction programs currently available are only about 60%6accurate when predicting the structure of a single residue ofa single protein (36), if used on a homologous set of proteinswith the assumption that they have a common secondarystructure in the region under consideration, the consensusprediction should be far more accurate. We have applied themethod ofRobson and Garnier (37) to 42 known or presumedviral RNA-dependent RNA polymerases, aligning the struc-tures by using the regions of homology (Fig. 6). The resultsshow a remarkable degree ofagreement among the structurespredicted for various viruses, suggesting that this approachmay be valid. The most conserved regions are predicted tohave 3-sheet structure with turns at the most conservedresidues, a result similar to that reported previously for aslightly different domain for a set of RNA-dependent poly-merases that overlaps with those we have studied (3).

DISCUSSIONRNA-dependent RNA polymerases are unique to viruses.This means that such enzymes should be possible targets forantiviral drugs. Fortunately, the RNA-dependent RNA poly-merases of (+) strand ssRNA viruses and dsRNA virusesshare substantial common sequence patterns (1-3, 38). If onecould determine the specific function of these residues, itmight be possible to design drugs, based on the substrate, thatwould act to block the replication of whole classes of animaland plant viruses. It was with this admittedly ambitious goal

L-A ORF2: E W Y R L Q G T L L S G W R LT T F M N T V L E W A Y M K LL-A Predicted H H E E E E T E E E C T E E E E E E E E E H H H H H H H H E

Helix 151210 8 7 8 5 4 4 4 5 5 5 6 9 8 9 9111216211921222424262219Sheet 192629313229101714 5 3 9162827272626232217181614181514121215Turn 5 1 0 0 1 3251915161612 6 0 1 0 2 2 2 2 1 0 0 0 0 0 0 0 0 0Coil 0 1 1 1 0 0 0 0 714161413 6 3 5 3 3 4 4 6 1 5 4 0 0 0 0 0 0

Consensus conformation:E E E E E E T T T T C C E E E E E E E E E H H H H H H H H H

L-A ORF2L-A Predicted

HelixSheetTurnCoil

A G V F D L D D V 0 D S V H N G D V M I S L N R V S T A VE E E E H H H H H H E E E C C T T E E E E E E E H H H E E E* .** ** * ** ** * ** * ** ** .***

1715191724303431292516 8 911101112171512 912132026262325282912 9111410 4 4 7 9142323128j11 3 81825293128311812 8101411103 4 5 6 6 3 3 3 2 2 2 0 1 2172620 6 0 0 1 1 0 1 2 2 3 0 0 05 9 5 2 0 3 1 0 1 0 0 0 0 1 3 1 1 1 2 0 0 0 0 1 0 4 4 1 1 1

Consensus conformation:H H H H H H H H H H E E E E T T T E E E E E E H H H H H H H

FIG. 6. Conservation of secondary structure among many RNA-dependent RNA polymerases from diverse viruses. The program of Robsonand Gamier (37) was used to predict the conformations of 42 known and presumed RNA-dependent RNA polymerases of (+) ssRNA virusesand dsRNA viruses. The output was aligned by using the conserved residues (double underlining), and the number of sequences predicted tohave the indicated conformation is tabulated. Asterisk, L-A prediction agrees with consensus; dot, almost agrees; E, ,-sheet; H, a-helix; T,turn; and C, coil. Domains 1 and 2 are underlined.

tjQ7

Proc. Natl. Acad. Sci. USA 89 (1992)

MMl +ZEI

KADI L 13

Dow

nloa

ded

by g

uest

on

Dec

embe

r 21

, 202

0

Proc. Natl. Acad. Sci. USA 89 (1992) 2189

that we began to define domains in the L-A virus pol ORFsurrounding the conserved RNA-dependent RNA polymer-ase sequence patterns that are essential for viral propagation.Previous studies of these regions in other viruses have beenlimited to mutation of the G residue of the GDD motif inpoliovirus and Q83 phage (22, 23) and mutation in bromemosaic virus of several residues in and around the D... [YF]Dconsensus pattern that is N-terminal to the SG...T...NT..Nmotif (24), all resulting in decreased or temperature-sensitiveactivity. One study of QJ3 replicase found that insertions inmost parts of the molecule inactivate it (39). In the studies ofpoliovirus (23), polymerase was produced in Escherichia coliin soluble form and the mutants produced were examined forsome of their enzymatic properties. Although some substi-tutions resulted in decreased activity, and not simply a"dead" enzyme, no qualitative changes in the enzyme werefound that would suggest the nature of the defect. Somemutations of the human immunodeficiency virus reversetranscriptase domains have also been analyzed (40). Al-though in vitro systems for L-A replication, transcription,and a partial packaging reaction have been developed (11,12), we were unable to assay these activities in particlesproduced from the L-A expression vector. Thus our studieshave been initially restricted to measuring the effects of themutations on overall viral propagation, assuming thatchanges in this region of pol are affecting mostly RNA-dependent RNA polymerase activity.

This is the most extensive study made to date of thesedomains. We found that all of the consensus residues exam-ined were essential. Moreover, residues in the surroundingareas were also necessary, in various degrees, for viralpropagation. Our experiments confirm part of the "consen-sus" approach in that the most stringently required residueswere, in fact, those which are most conserved ones, exceptperhaps for the NT and N residues of the SG...T...NT..Nmotif. However, the domain sizes predicted (3) were 28 and11 residues for the SG...T...NT..N and GDD motifs, respec-tively, but we found 21 and 29 residues as their actual extents.Of course, our work assumes that failure of viral propagationis due to failure ofRNA polymerase activity, and we have notyet been able to measure this directly. If sequences with otherfunctions are interspersed among the RNA polymerase mo-tifs, it would confuse our results. It will be of interest tocompare the results obtained here with similar studies fromother systems when they become available.We could locate a three- to five-amino acid region between

the SG.. .T...NT..N and GDD motifs that was not essential,but even this region could not be freely changed in length.Insertion of six alanine residues prevented viral propagation.The importance of this interdomain distance may, in somecases, have prevented our seeing activity with substitutionsof sequences from other viruses. Although we focused on thetwo most conserved domains, less conserved motifs in otherparts of the molecule have been recognized (1, 3), andspecific interactions of these with the two domains westudied could have made substitution mutants inactive. Alsopotentially critical may be some template sequence-specificinformation encoded in these domains. Although we argue(above) that this is unlikely, it is not impossible, and exam-ination, in the L-A system, of where such specificity resides,may be enlightening.

We thank Charles Rice, Wolfgang Joklik, and Bert Semler forgenerously providing clones of Sindbis virus, reovirus, and poliovi-rus RNA polymerases, respectively. J.C.R. acknowledges postdoc-toral fellowship support from the Ministerio de Educaci6n y Ciencia,Spain.

1. Kamer, G. & Argos, P. (1984) Nucleic Acids Res. 12, 7269-7282.

2. Argos, P. (1988) Nucleic Acids Res. 16, 9909-9916.3. Poch, O., Sauvaget, I., Delarue, M. & Tordo, N. (1989) EMBO

J. 8, 3867-3874.4. Icho, T. & Wickner, R. B. (1989) J. Biol. Chem. 264, 6716-

6723.5. Wiener, J. R. & Joklik, W. K. (1989) Virology 169, 194-203.6. Roy, P., Fukusho, A., Ritter, G. D. & Lyon, D. (1988) Nucleic

Acids Res. 16, 11759-11767.7. Cohen, J., Charpilienne, A., Chilmonczyk, S. & Estes, M. K.

(1989) Virology 171, 131-140.8. Mindich, L., Nemhauser, I., Gottlieb, P., Romantschuk, M.,

Carton, J., Frucht, S., Strassman, J., Bamford, D. H. &Kalkkinen, N. (1988) J. Virol. 62, 1180-1185.

9. Morgan, M. M., Macreadie, I. G., Harley, V. R., Hudson,P. J. & Azad, A. A. (1988) Virology 163, 240-242.

10. Wickner, R. B. (1991) in Molecular and Cellular Biology oftheYeast Saccharomyces, eds. Jones, E. W., Pringle, J. & Broach,J. R. (Cold Spring Harbor Lab., Cold Spring Harbor, NY), pp.263-2%.

11. Fujimura, T. & Wickner, R. B. (1988) J. Biol. Chem. 263,454-460.

12. Fujimura, T. & Wickner, R. B. (1989) J. Biol. Chem. 264,10872-10877.

13. Esteban, R., Fujimura, T. & Wickner, R. B. (1989) EMBO J. 8,947-954.

14. Fujimura, T., Esteban, R., Esteban, L. M. & Wickner, R. B.(1990) Cell 62, 819-828.

15. Hopper, J. E., Bostian, K. A., Rowe, L. B. & Tipper, D. J.(1977) J. Biol. Chem. 252, 9010-9017.

16. Fujimura, T. & Wickner, R. (1988) Cell 55, 663-671.17. Dinman, J. D., Icho, T. & Wickner, R. B. (1991) Proc. Natl.

Acad. Sci. USA 88, 174-178.18. Bussey, H. (1988) Yeast 4, 17-26.19. Bostian, K. A., Sturgeon, J. A. & Tipper, D. J. (1980) J.

Bacteriol. 143, 463-470.20. Sommer, S. S. & Wickner, R. B. (1982) Cell 31, 429-441.21. Wickner, R. B., Icho, T., Fujimura, T. & Widner, W. R. (1991)

J. Virol. 65, 155-161.22. Inokuchi, V. & Hirashima, A. (1987) J. Virol. 61, 3946-3949.23. Jablonski, S. A., Luo, M. & Morrow, C. D. (1991) J. Virol. 65,

4564-4572.24. Kroner, P., Richards, D., Traynor, P. & Ahlquist, P. (1989) J.

Virol. 63, 5302-5309.25. Wickner, R. B. & Leibowitz, M. J. (1976) Genetics 82, 429-

442.26. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular

Cloning:A Laboratory Manual (Cold Spring Harbor Lab., ColdSpring Harbor, NY), pp. 250-251.

27. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl.Acad. Sci. USA 74, 5463-5467.

28. Tabor, S. & Richardson, C. C. (1987) Proc. Natl. Acad. Sci.USA 84, 4767-4771.

29. Kunkel, T. A. (1985) Proc. Natl. Acad. Sci. USA 82, 488-492.30. Rice, C. M., Levis, R., Strauss, J. H. & Huang, H. V. (1987)

J. Virol. 61, 3809-3819.31. Saiki, R. K., Gelfand, D. H., Stoffel, S., Scharf, S. J., Higuchi,

R., Horn, G. T., Mullis, K. B. & Erlich, H. A. (1988) Science239, 487-491.

32. Hitzeman, R. A., Leung, D. W., Perry, L. J., Kohr, W. J.,Levine, H. L. & Goeddel, D. V. (1983) Science 219, 620-625.

33. Ito, H., Fukuda, Y., Murata, K. & Kimura, A. (1983) J.Bacteriol. 153, 163-168.

34. Conde, H. & Fink, G. R. (1976) Proc. Natl. Acad. Sci. USA 73,3651-3655.

35. Cunningham, B. C. & Wells, J. A. (1989) Science 244, 1081-1085.

36. Gamier, J. (1990) Biochemie 72, 513-524.37. Robson, B. & Gamier, J. (1986) Introduction to Proteins and

Protein Engineering (Elsevier, Amsterdam).38. Delarue, M., Poch, O., Tordo, N., Moras, D. & Argos, P.

(1990) Protein Eng. 3, 461-467.39. Mills, D. R., Priano, C., DiMauro, P. & Binderow, B. D. (1988)

J. Mol. Biol. 205, 751-764.40. Larder, B. A., Kemp, S. D. & Purifoy, D. J. M. (1989) Proc.

Natl. Acad. Sci. USA 86, 4803-4807.

Biochemistry: Ribas and Wickner

Dow

nloa

ded

by g

uest

on

Dec

embe

r 21

, 202

0