JOURNAL OF VIROLOGY, Sept. 1994, p. 6014-6020 Vol. 68, No. 90022-538X/94/$04.00+0Copyright © 1994, American Society for Microbiology

A Cryptic RNA-Binding Domain in the Pol Region of the L-ADouble-Stranded RNA Virus Gag-Pol Fusion Protein

JUAN CARLOS RIBAS,* TSUTOMU FUJIMURA,t AND REED B. WICKNER

Genetics of Simple Eukaryotes, Laboratory of Biochemical Pharmacology, National Institute of Diabetesand Digestive and Kidney Diseases, Bethesda, Maryland 20892

Received 8 April 1994/Accepted 17 June 1994

The Pol region of the Gag-Pol fusion protein of the L-A double-stranded (ds) RNA virus of Saccharomycescerevisiae has (i) a domain essential for packaging viral positive strands, (ii) consensus amino acid sequencepatterns typical of RNA-dependent RNA polymerases, and (iii) two single-stranded RNA binding domains. Wedescribe here a third single-stranded RNA binding domain (Pol residues 374 to 432), which is unique in beingcryptic. Its activity is revealed only after deletion of an inhibitory region C terminal to the binding domainitself. This cryptic RNA binding domain is necessary for propagation of M1 satellite dsRNA, but it is notnecessary for viral particle assembly or for packaging of viral positive-strand single-stranded RNA. The crypticRNA binding domain includes a sequence pattern common among positive-strand single-stranded RNA anddsRNA viral RNA-dependent RNA polymerases, suggesting that it has a role in RNA polymerase activity.

Multiple RNA-protein interactions must occur in the pro-cesses of replication and packaging of an RNA virus. Accord-ingly, many RNA-binding viral proteins have been described,but in only a relatively few cases has the function of the RNAbinding been defined.L-A is a double-stranded (ds) RNA virus of Saccharomyces

cerevisiae whose genome consists of a single 4.6-kb segment(reviewed in reference 25). L-A's cis-acting packaging signal, astem-loop structure with a protruding A residue, is located 400nucleotides from the 3' end of the viral positive strand (8).Replication (negative-strand synthesis) requires two sites, aninternal enhancer and the special 3'-end structure and se-quence where replication initiates (6) (Fig. 1A). Viral positive-strand single-stranded (ss) RNA has two open reading frames(ORFs), the 5' gag encoding the major coat protein (Gag) andthe 3' pol expressed only as a Gag-Pol fusion protein formed bya -1 ribosomal frameshift event in the region of overlap ofgagand pol (5, 10, 13). Gag covalently binds the 5' cap structuresof cellular mRNAs, a reaction that appears necessary forexpression of viral information (1, 2). Pol is multifunctionalwith an N-terminal domain necessary for packaging of viralpositive strands (9, 18), amino acid sequence patterns typical ofRNA-dependent RNA polymerases of positive-strand ssRNAand dsRNA viruses (13, 19), and in vitro ssRNA bindingactivity (10, 18) (Fig. 1A).Ml is a 1.8-kb dsRNA satellite of the L-A virus that needs

the L-A proteins for its propagation. Ml is responsible for theS. cerevisiae killer phenotype, encoding the secreted killer toxinand immunity to the toxin (reviewed in reference 3). ThisdsRNA satellite has been a useful tool in studying in vivo therequirements of L-A proteins for RNA propagation. XdsRNA, a deletion mutant of L-A, has likewise been useful instudying the cis signals necessary for several steps in the virallife cycle (7). The availability of in vitro replication andtranscription systems for this virus and the isolation of host

* Corresponding author. Mailing address: Bldg. 8, Room 225,National Institutes of Health, Bethesda, MD 20892. Phone: (301)496-3452. Fax: (301) 402-0240. Electronic mail address: wickner@helix.nih.gov.

t Present address: Department of Microbiology, Facultad de Biolo-gia, University of Salamanca, Salamanca, 37008 Spain.

mutants in genes necessary for viral propagation and in genesrepressing viral copy number, along with the molecular analysisof these genes and the viral proteins, have led to the under-standing of many aspects of the functions of viral and hostcomponents in this system (reviewed in reference 25).

Proteins expressed from an L-A cDNA clone can propagatethe M1 or X satellite dsRNAs in S. cerevisiae, and the effects ofmodifications of the cDNA clone can be used to study thefunctions of the proteins. The regions surrounding the twomost conserved RNA-dependent RNA polymerase motifshave been studied by alanine scanning mutagenesis, resultingin definition of two critical domains for viral propagation (19).In vivo packaging studies showed that Pol residues 67 to 213are necessary for encapsidation of transcripts carrying thepackaging signal (9, 18). However, Gag alone is sufficient toform morphologically intact (albeit empty) virus particles (9).In vitro blotting experiments with fragments of Pol expressedin Escherichia coli have already defined two RNA bindingdomains. The N-terminal domain (residues 172 to 190) islocated within the packaging domain, and its deletion preventspackaging. The C-terminal RNA binding domain (residues 770to 819) is not essential for packaging but is required for Mlpropagation (18) (Fig. 1A).

In this work, we show that a third RNA-binding domain ispresent in Pol. This domain is cryptic, revealed only when aninhibitory region C terminal to the binding domain is deleted.We show that this cryptic RNA binding domain is essential forMl propagation in vivo but not for particle assembly or RNApackaging. The location of a conserved RNA polymerase motifin this cryptic RNA binding domain suggests that its function isrelated to RNA polymerization.

MATERUILS AND METHODS

Strains and media. 4.7MB (27), YPAD, YPG, SD, andsynthetic complete medium lacking specific amino acids (22)and LB and TB media (20) have been described previously. S.cerevisiae JR1 (AL4Ta trpl ura3 leu2 his3 pep3::HIS3nucl::LEU2 L-A-o L-BC), JR6 (JRlp°), JR5 pI2L2 K+(MATot karl-i ura2 leu2 trpl L-A-o pI2L2 M1), and 5X47(MA4Ta/MATot hisl/+ trpl/+ ura3/+ M-o) were used. E. coli

6014

CRYPTIC RNA-BINDING DOMAIN OF L-A'S Pol 6015

APseuidoknot

Slippery site 3...I8GUUUAGGgag

9ORF

-1 Ribosomalframeshift site

L-A (+) strand

.qaq ORF

7mGp

major coat protein

AX:C- u~

..GAAuAcUACCAUCGG... U-A

Packgig lAUjAA-tUAUG

Pol ORFReplicationsites

gag-pol fusion protein

RNA / \ SG...T...NT..N-- GDD Xackagin RNA-dependent

binding RNA polymerase bindingconsensus

B

Apal RNA binding\Nr I

SnaBI SnaBI(Noti Nrul Mscl

(1) (1o) (369) 614)(670) (914)57-58 205-6 223-4 305-6

po1RNA binding

Sa Afl 11 Xhol

(1241) nucleotlde (1910) 12580)

413-4 amino acWi 637-8 860

FIG. 1. (A) Sites and encoded proteins of the L-A positive strand.The gag ORF encodes the major coat protein (Gag), and the pol ORFis synthesized as a Gag-Pol fusion protein by a -1 ribosomal frame-shift. The specific encapsidation of viral positive-strand ssRNA isdriven by the stem-loop packaging signal, and the conversion of viralpositive-strand ssRNA to ds form (replication) depends on an internalsite and a 3'-end RNA replicase recognition site. The Gag proteinHis-154 is the site of covalent attachment of cap structures. The Polregion has two ssRNA binding domains, both necessary for viralpropagation and the N-terminal domain essential for packaging viralpositive-strand ssRNA. The in vivo packaging domain consists of Polresidues 67 to 213 and contains the N-terminal in vitro RNA bindingdomain. Pol also contains the consensus patterns typical of RNA-dependent RNA polymerases of positive-strand ssRNA and dsRNAviruses. (B) Restriction map of the pol ORF. Dashed lines are outsidethe ORF. Nucleotide and amino acid sequence numbers refer to poland begin with the first amino acid after the frameshift. Boxes show thelocations of the previously defined N- and C-terminal RNA bindingdomains.

MV1190 and CJ236 (Bio-Rad), DH5o (Bethesda ResearchLaboratories), and WM6 (17) were used.

Plasmids and DNA techniques. pJR7 (18) contains the L-Apol region inserted in frame with the 29-residue OmpA signalpeptide-FLAG peptide of the pFLAG E. coli expression vector(IBI). pJR7D (18) is pJR7 with a deletion from amino acid 415of Pol to the C terminus. From pJR7 were derived deletionsp3'1 to p3'56 with exonuclease III digestion, starting from the3' end of pol and leaving intact its 5' end, while p5'1 to p5'56delete from the 5' end of pol, likewise with exonuclease IIIdigestion, but remaining in frame with the AUG start codonand leaving its 3' end intact (18). pDl to pD42 were 5'-enddeletions of pJR7D (18).pJR14 was made from pJR7 by cutting with NruI (Fig. 1B)

and religating, removing the last 22 amino acids of theOmpA-FLAG peptide and the first 223 amino acids of the Polregion, and leaving the remaining residues, 224 to 860, inframe. Similarly, pJR15, pJR16, pJR17, pJR18, pJR19, pJR20,pJR30, pJR31, pJR32, and pJR54 were made by NruI restric-tion and religation of p3'2, p3'4, p3'3, p3'9, p3'13, p3'18,p3'11, p3'14, p3'17, and pJR7D, leaving in frame Pol residues

224 to 755, 685, 629, 546, 432, 366, 506, 449, 437, and 414,respectively.pJR21 to pJR26 were made from p5'28, p5'32, p5'27, p5'38,

p5'42, and p5'46, deleting the AfllI-XhoI fragment (Fig. 1B),filling its ends with Klenow polymerase, and religating. Theseplasmids express Pol amino acids 290 to 637, 341 to 637, 374 to637, 399 to 637, 457 to 637, and 500 to 637, respectively. pJR34to pJR38 were made from pD40, pJR21, pJR22, pJR23 andpJR24, respectively, by cutting with Sall in pol, and with ApaIin the pFLAG sequence and transferring the resulting segmentinto pJR30 also cut with the same enzymes. They expressed Polresidues 201 to 506, 209 to 506, 341 to 506, 374 to 506, and 399to 506, respectively.pJR39 is pJR34 with a Notl-Sall deletion (Klenow filled) of

pol and expressing Pol residues 201 to 205 and 414 to 506.pJR40 is pJR14 with the termination codon UAG introducedby site-directed mutagenesis at amino acid 438, expressing Polresidues 224 to 437.pJR47, pJR48, pJR49, pJR50, and pJR51 are p3'3, p3'9,

p3'11, p33'14, and p3'17 after deletion of the SnaBI-MscIfragment, expressing the Pol amino acids 1 to 57 and 306 to629, 1 to 57 and 306 to 546, 1 to 57 and 306 to 506, 1 to 57 and306 to 449, and 1 to 57 and 306 to 437, respectively. pJR58 hasthe X sequence from pRE76-2-14 (6) cloned in the S. cerevisiaeexpression vector pVT1O1U (URA3 selection, ADHI pro-moter [18, 24]). pJR61 is pJR30 with a deletion of Pol residues374 to 432 made by site-directed mutagenesis. pJR62 is theL-A cDNA expression plasmid pI2L2 (TRP1 selection, PGK1promoter [26]) with Pol amino acids 386 to 425 deleted bysite-directed mutagenesis. pJR79 is pI2L2 with the substitution416KYEWGKQR423-*416AAEWGAQA423 made by site-di-rected mutagenesis.

Site-directed mutagenesis (16) was done with the Bio-RadMuta-Gene kit, and DNA sequencing (21) was done with aSequenase kit (U.S. Biochemicals). Plasmid DNA was intro-duced into S. cerevisiae as previously described (14) with somemodifications (23). Single-stranded RNA transcripts weremade with [a-32P]UTP, T3 (for the negative strand) or T7 (forthe positive strand) RNA polymerase, and pLM1 DNA (8, 12)as recommended by the supplier (U.S. Biochemicals). pLM1contains the L-A sequences of the natural deletion mutant, X(12). Other DNA manipulations were as described previously(20).

Heterologous protein expression in E. coli. Plasmids for E.coli expression were maintained in media supplemented with0.5% glucose in order to avoid the toxicity of basal Polexpression. Heterologous proteins were prepared as describedelsewhere (18). Briefly, a 2-ml culture of strain WM6 carryingdifferent plasmids was grown with IPTG (isopropyl-3-D-thio-galactopyranoside) for 5 h to induce expression, and cells werecollected and lysed by several freeze-thaw cycles. After differ-ential centrifugation and detergent washing, the heterologousproteins remained in the pellet, which was resuspended in 150,ul of 50 mM Tris HCl (pH 7.5)-150 mM NaCl-10 mMEDTA-1.5 mM phenylmethylsulfonyl fluoride-10% glycerol,and 50 ,ul of 4x sodium dodecyl sulfate (SDS) buffer (1 x = 10mM Tris HCl [pH 7.5], 1% SDS, 0.005% bromophenol blue,1% mercaptoethanol, 10% glycerol) was added.

Electrophoresis and RNA binding blot assay. Equalamounts of proteins solubilized in lx SDS buffer were ana-lyzed by SDS-12% polyacrylamide gel electrophoresis(PAGE) and Coomassie blue stained or electroblotted to anitrocellulose sheet (18). The sheet was washed in the presenceof 4 M urea, and the binding reaction was carried out at roomtemperature for 30 min in 45 ml of 10 mM morpholineethane-sulfonic acid (MES)-Tris (pH 7.0)-50 mM NaCl-1 mM

VOL. 68, 1994

6016 RIBAS ET AL.

EDTA-1 x Denhardt's solution-0.01 ,ug of heparin per ml-1.5mg of denatured calf thymus DNA and a 32P-X positive-strandprobe (100 ng, 4.5 x 10' to 5.0 x 106 cpm of T7 transcript ofpLM1). Although the probe includes the L-A packaging signal,the in vitro RNA binding was not site specific. In place of calfthymus DNA, equal amounts of tRNA or Torula total RNAwere also used, and similar Pol binding was observed, but withhigher background binding by E. coli proteins. Binding com-plexes were detected by autoradiography.

Preparation of particles made from an L-A cDNA cloneexpressed in S. cerevisiae. Strain JR1, carrying different L-AcDNA constructs, was grown in H-Trp or H-Trp-Ura mediumat 30°C for 3 days, and cDNA-derived particles (called cDNAparticles below) were purified as described elsewhere (18).Cells were harvested (10 g [wet weight]) and after treatmentswith Zymolyase 20T in the presence of osmotic stabilizer werelysed by osmotic difference. cDNA particles were partiallypurified by differential centrifugation and separated by CsClgradient centrifugation (130,000 x g, 20 h, 4°C, p = 1.32 g/ml,13-ml total volume). Fractions of 0.5 ml were collected andstored at -70°C. For immunological analysis of viral proteins,SDS-7.5% polyacrylamide gels of gradient fractions wereeither stained with Coomassie blue or transferred to a nitro-cellulose sheet, which was cut along the 97-kDa prestainedmolecular mass marker and treated as described previously(10). The upper part was incubated with anti-Pol, the lowerpart was incubated with anti-Gag polyclonal antiserum, andthe detecting reaction was with an alkaline phosphatase-conjugated second antibody (Promega). For Northern (RNA)blot hybridization, gradient samples were prepared and pro-cessed as described elsewhere (8, 18).L-A cDNA opened empty particles. Particles were isolated by

a CsCl gradient as described above but with a total volume of40 ml to give better separation of particles (p = 1.32) fromcontaminant proteins (p = 1.31). Fractions (1 ml) were col-lected, and particle-containing fractions were pooled anddiluted to 25 ml with buffer A (50 mM Tris HCl [pH 7.5], 150mM NaCl, 10 mM EDTA, 1 mM dithiothreitol). Particles werecollected by centrifugation (130,000 x g, 1 h, 4°C) andresuspended in 2 ml of buffer A. Samples were dialyzed against700 ml of 20 mM Tris HCl (pH 7.5)-i mM EDTA, clarified bycentrifugation for 10 min at 4°C in a microcentrifuge, concen-trated with a Centricon-100 (Amicon) (final volume, 100 to 120RI), and stored at -70°C without glycerol. Perhaps becausethese particles are closed and empty, dialysis against low-ionic-strength buffer is not sufficient to open them as it is in the caseof normal L-A viral particles (11), and a cycle of freeze-thaw inthe absence of glycerol is critical to open them and detect theirPol region activity. After the particles were thawed, glycerolwas added to 20% to stabilize their activity, and they werestored at -70°C.RNA packaging site-specific RNA binding measured by gel

retardation. The standard reaction mixture (15 ,ul) contained50 mM Tris HCl (pH 9.5), 5 mM EDTA, 1 mM dithiothreitol,opened empty cDNA particles (100 ,ug of protein), 100 mMNaCl, 0.7 mg of bentonite per ml, 40 U of RNasin (Promega),1 to 2 ,ug of RNA (from T7 RNA polymerase transcription ofpBluescript [Stratagene]), and 32P-labeled X positive-strandssRNA (1 ng, 5 x 10' cpm). The mixture was incubated at 30°Cfor 20 min, and the binding complex was separated by electro-phoresis on a 1.5% agarose gel and detected by autoradiogra-phy. While the cDNA opened empty particles bind X positive-strand ssRNA under the conditions previously described forL-A particles (11), their more narrow separation from contam-inating proteins on the CsCl gradients necessitated these

-! Lo AC C2 cD O; N TC C: Cl: CC)C) -f C

-D -1 --,) --D --3 ---) n1a LI)

-I~- C- -,) ) -) C-) -t C-)

9768

_

43 _! m

29_---F

18

14

Coomassie Blue

gag 780 fN1~pcig172-190 432 506-546

.... X W

224 . 755224 685224 629224 546224 506224 449224 437224 432224 414224 366

ssRNA Binding

770-8 19........i...- 860

-860 + JR14- JR15- JR16- JR17- JR18+ JR30+ JR31+ JR32+ JR19- JR54- JR20

FIG. 2. Deletion of Pol residues 546 to 506 uncovers a centralcryptic RNA binding domain whose C terminus is residue 432. Proteinsproduced in E. coli from pJR14 and a series of 3'-end deletion mutantswere analyzed by SDS-12% PAGE. Gels were either stained withCoomassie blue (numbers to the left refer to the size of markerproteins in kilodaltons) or transferred to a nitrocellulose membrane,treated with urea, and probed with 32P-X positive-strand ssRNA. "+"and "-" denote ssRNA binding activity. In the figures and the text, theprotein expressed from pJRxx is designated JRxx. The graded darkchecked box shows the C-terminal end of the C-terminal inhibitoryregion, and the light hatched boxes show the RNA binding domains.

modified conditions to limit background binding by contami-nants.

In vivo killer assay of altered L-A expression vectors. Bothcytoduction and the killer assay were done as describedelsewhere (18). Ml dsRNA satellite was introduced by cyto-duction (4) into the recipient strain, JR6, carrying the mutantsof the L-A expression plasmid. The donor strain, JR5 pI2L2K+, has no L-A virus, and Ml is supported by the L-A cDNAplasmid, pI2L2, which is almost never transferred by cytoduc-tion. Thus, Ml in the recipient strain depends on the proteinsexpressed from the mutant cDNA clone.

RESULTS

Pol has a cryptic, third in vitro ssRNA binding domain. Wehave shown that Pol has an N-terminal ssRNA binding domainfrom residues 172 to 190 and a C-terminal ssRNA bindingdomain from residues 770 to 819 (18). Starting with pJR14,which lacks the N-terminal domain and expresses Pol residues224 to 860, we made deletions from the 3' end and analyzedtheir in vitro ssRNA binding activity (Fig. 2). The first deletionderivative (pJR15) has already lost ssRNA binding activity, asexpected from its ending at residue 755 and thus lacking theC-terminal RNA binding domain.Longer C-terminal deletions of Pol also failed to bind RNA,

but surprisingly, a deletion lacking 354 amino acids from thePol C terminus (JR30, C terminus at residue 506) again

J. VIROL.

CRYPTIC RNA-BINDING DOMAIN OF L-A'S Pol 6017

Co C: Cl

9768-

43

29 __

18

14 -

a m O

C}~I'll

n

- aooOv v vq en LO

5n-

m m m m

9768_-

43 -

ma29

16

14

Coomassie Blue ssRNA Binding

gag 780 i-\ pol172-190 432 506-546 770-819

1 - --- - -- 860224 506224 437224 437

IUAG)

+ JR30+ JR32+ JR40

Coomassie Blue ssRNA Bindinggag __ 780 Z Pof-_-__ ---172190 432 506-546 770-619-1 ... 860

1 -57 306 6291 -57 306 5461 -57 306 5061 -57 306- 4491 -57 306- 437

+ 3'3+/-JR47- JR48+ JR49+ JR50+ JR51

FIG. 3. The cryptic central in vitro RNA binding domain is not due to fusion C-terminal sequences or a unique N terminus. At left, atermination codon at position 438 of JR14 makes the same fragment as JR32 but without the extra residues due to fusion with downstreamsequences. At right is shown a series of 3'-end deletions lacking the N- and C-terminal in vitro RNA binding domains and having a different Nterminus than those of Fig. 2. Fragments were tested as in Fig. 2.

showed binding activity. This fragment lacks both N- andC-terminal RNA binding domains. Three shorter fragments(JR31, JR32, and JR19) also showed this activity. This impliesthe existence of a third RNA binding domain, distinct from theN- and C-terminal domains, and normally masked by othersequences in Pol C terminal to the binding domain itself. TheC-terminal end of this inhibitory region is localized betweenresidues 546 and 506 of Pol (from the last deletion withoutbinding [JR18] to the first with binding [JR30]). The C-terminal end of the cryptic binding domain itself is betweenresidues 432 and 414 (Fig. 2).

Cryptic binding is not an artifact of C-terminal sequences orof the special N-terminal sequence. Random 3'-end deletionsproduce extra C-terminal amino acids until the first termina-tion codon is reached. We introduced into pJR14 a UAGtermination codon in place of residue 438 of Pol to produce aprotein (JR40) with the same Pol sequence as JR32 but lackingany extra C-terminal residues (Fig. 3, left). This protein boundssRNA as well as did JR32 or JR30, showing that the extraC-terminal tail was not responsible for the binding. In addition,sequence analysis of the deletion constructs shown in Fig. 2revealed that JR17, JR18, and JR20 had the three possible tailsformed by fusion in the three possible frames with the down-stream sequence, and none of these proteins bound RNA.

Since all of the constructs in Fig. 2 had the same N-terminalstructure starting at Pol residue 224, we constructed a similarseries of C-terminal deletions with a different N-terminalsequence (Pol 1 to 57 joined in frame to Pol 306 and thusavoiding the N-terminal RNA binding domain) and ending atthe same points as in the constructs in Fig. 2. These proteins(Fig. 3, right) showed essentially the same binding as did thosein Fig. 2, except for JR47, which showed weak activity, whilethe comparable JR17 had no activity.The N-terminal end of the cryptic RNA binding domain is at

Pol 374. Starting with a construct (JR34) lacking the C-terminal inhibitory domain, we made deletion mutations fromthe N terminus and found that the N terminus of the Pol

fragment is important for whether RNA binding activity isobserved. Fragments starting with residue 1, 224, or 374 (Fig.2, 3, and 4) show binding activity in the absence of theC-terminal inhibitory domain, but those starting at residue 201,209, or 341 did not bind (Fig. 4, left). The binding of JR37shows that the N terminus of the central binding domain lies Cterminal to residue 374 (Fig. 4, left). Further removal ofresidues 374 to 398 again eliminated RNA binding activity,indicating that the N terminus of the cryptic binding domain isin this interval (Fig. 4, left). In contrast, a construct similar toJR37 in which the C-terminal inhibitory domain was left intacthad no RNA binding activity (Fig. 4, right).The variation of RNA binding depending on the N terminus

of the expressed Pol fragment suggests that binding maydepend on the folding of the fragment, a conclusion consistentwith its being cryptic.

Deletion of residues 374 to 432 eliminates the cryptic RNAbinding activity. Starting with the largest fragment that has thecryptic RNA binding activity (JR30), we deleted residues 374to 432, fusing the remainder in frame (Fig. 5). This construct(JR61) had no RNA binding activity remaining, indicating thatthe segment deleted was the only one responsible for thecryptic RNA binding activity.

Basic regions and RNA binding. The two most basic clustersin the area of the central RNA binding domain are461KVHKR465, which lies outside the cryptic (central) RNAbinding domain, and 393KMPKHKISR401. The latter is withinthe cryptic RNA binding domain but is not sufficient for RNAbinding: JR54 (224 to 414, Fig. 2) lacks the inhibitory regionand has an N terminus that is consistent with RNA binding butlacks binding activity. These results indicate that, while RNAbinding domains in this and other proteins include basicresidues, the most basic regions are not necessarily sufficientfor the binding of RNA.

Evidence that the cryptic RNA binding domain is necessaryfor M1 propagation in vivo. The L-A cDNA expression plas

VOL. 68, 1994

6018 RIBAS ET AL.

v cn r- O cn r LO CO rs CO mC CO - mm m m m m

rm m E r7 m

) )I C)C-) C-) O-) - ) C-) 3

9768

43_M . ....

2-

18

1es

CO - CU C) ~t LO (cNl CN N N N ("iC\lN-U CU CU: CU CU Cl: a:OCmcor m m m

9768 -

43 - _

A......_

29

18

14

Coomassie Blue ssRNA Binding

gag 780 po/172-190 374-432506-546 77C 819

*0. ' 860

3T11 + 1 506JR34 - 201 506JR35 - 209 506JR36 - 341 -506JR37 + 374--- 506JR38 - 399 506JR39 - 201 414 -506

Coomassie Blue ssRNA Bindinggag 780 po/

172-190 374-432505 46 770 819

-.

5'28JR21JR22JR23JR24JR25JR26

29 P---- 37341 - - 637374---- ---- .637399 ---637

45 7-- 637

5 o ---637

FIG. 4. At left, certain N termini of the expressed Pol fragment inhibit binding, and the N-terminal end of the cryptic binding domain is residue374. At right, the C-terminal inhibitory region is sufficient to prevent RNA binding.

mid can support the propagation of Ml in S. cerevisiae, but a

deletion within the cryptic binding domain (residues 386 to425, pJR62) or the substitution 416KYEWGKQR423-*416AAEWGAQA423 (pJR79) resulted in inability to support Ml

o _-CY) Cocc a:

co coa: cr

97 -68 -

43

29A.

18 --

14

Coomassie Blue ssRNA Bindinggag 780 Pol

9l7542-I9O374843546 770-819

11 --- - --- l- --IIIA--860

JR30 + 224 506

JR61 224 --2 I 506

FIG. 5. Deletion of the central RNA binding domain in the largestfragment with RNA binding activity (due to the central domain)eliminates the activity. Protein fragments were expressed and analyzedas explained in Fig. 2.

propagation. Thus, this region, and presumably its RNAbinding ability, has some function in the viral replication cycle.

Viral particles produced from pJR62, the deletion within thecryptic RNA binding site, were indistinguishable from thoseproduced from the wild-type cDNA in their amount and intheir incorporation of Gag-Pol fusion protein into particles(data not shown). Expression in S. cerevisiae of deletion mutantor wild-type cDNA clones together with another plasmidmaking heterologous transcripts carrying the L-A packagingsite (9, 18) showed that the mutant was as able as the wild typeto package (data not shown).When opened, purified cDNA particles specifically recog-

nize and bind RNA transcripts carrying the L-A packaging site,a reaction detected in a gel retardation assay. Such cDNAparticles made from the wild-type clone, pI2L2, bind labeledtranscripts (from pLM1) but not when the packaging site of thetranscripts has been destroyed (data not shown). Particlesmade from the substitution mutant pJR79 bind packagingsite-containing transcripts equally well, as expected from theknown location of the Pol packaging domain (residues 67 to213) (18).

DISCUSSION

During the viral propagation cycle of an RNA virus, differ-ent proteins (or different domains of the same protein) mustinteract with specific sites on the genomic RNA, in its variousforms: positive-strand and/or negative-strand ssRNA and/ordsRNA, depending on the virus. These various interactionsmust be turned on or off depending on the stage of the viralreplication cycle. For example, the RNA packaging domain ofthe Pol region of the L-A dsRNA virus Gag-Pol fusion proteinfirst recognizes the RNA packaging signal, located about 400nucleotides from the 3' end of the positive strands. At thisstage, negative-strand synthesis to form dsRNA is probablyprevented until the Gag region of the Gag-Pol fusion proteinhas associated with free Gag protein to enclose the positivestrands in the virus particle. Then, the RNA polymeraseinteracts with the internal replication enhancer, overlapping

CD - CN C tO(CDCM CUCl CUN CU CMinzmmmmmm

J. VIROL.

CRYPTIC RNA-BINDING DOMAIN OF L-A'S Pol 6019

ARSGRSYDASCNAYRRTYSWRCPRRVDRTGGQCFSRVNVIEPSHGPRPTRYILQEPGTYPA

WIRFRN.RVQAVSRQKATHFL FDIVPAAVISDF T TSDTSSFAYKSHTYAVNVTALRFSDTY________ _ __ __ ___ __ ___ __ ___ __ ___ ______--__ ____--------gdomain IJ l4ii-Aerm. Abiiidi ad6r iin

A.LY).TDTNMT I PAARRQA ATY_QQyAriFQQNTPTVMDSLANILDVDANIIRPKHFKGLRLYTRSKVTA EKHTHLRPDEILVEAAKVSPRRIYYLMCVVELLANLQVDLEAAVATILAYVLTLSEKFVPIFLDSRAIWVGEPGPDALTARLKASSGOIKSIHTADYEPLTELFELAVLM

NRGVGHVSWQAEKDHRLNPDVAVVDQARLYSCVRDMFEGSKOTYKYPFMTWDDYTANRWECnotmic RNA bindina domain

WVPGGSVHSQYEE_DNDYIPGYTRNKFITVNKMPKHKI RMIASPPEVRAWTSTKYEWGKOnRAIYGTDLRSIi LITNFAMA:MFRCc:E:DV:L:T:HKFPV -Q AAKHWVN LGASSF:FDF

Inhibitor of RNA bindinaRNA.de.endenU.R'. Polymerase Consensus Domainsr%RNA-dependent RNA Polymnerase Consensus Domains

_ R LTTFMNTVLNWAYMKLAGVFDLDDVQDSVHNGDDVMISLNRVSTAVRIMDAMH

RINARAQPAKCNLFSISEFLRVEHGMSGGDGLGAQYLSRSCATLVHSRIESNEPLSVVRV

MEADQARLRDLANRTRVQSAVTAIKEQLDKRVTKIFGVGDDVVRDIHTAHRVCGGISTDT

WAPVETK I ITDNEAYE IPYE IDDPSFWPGVNDYAYKVWKNFGERL E FNKIIKDAVARGSRSC-term. RNA binding domain

TIALKRKARITSKKNEFANKSEWERTMYKAYKGLAVSVYY NLSKFMSIPPMANIEFGQAR

YAMQAALDSSDPLRALQVIL

B

60

120

180

240

300

360

420

480

540

600

660

720

780

840

860

pol780172-190

I 167 213

in vivo packaging domain

...506-546

374-432 770 819I _

IW~IRAfl.1NS/ASIIIIIAl,SOSSa/ A-vv:s on-,

541-561 565-593RNA polymeraseconsensus domains

= RNA binding domain

= Inhibitor of RNA bindingFIG. 6. (A) Pol amino acid sequence with the in vivo RNA packaging, in vitro RNA binding, RNA binding-inhibitory, and RNA polymerase

consensus domains indicated. The graded stippled region (and the graded checked boxes in part B) indicate that the C-terminal border of theRNA-inhibitory region is known to be between residues 506 and 546 but its N-terminal border is not known. (B) Linear diagram of the functionaldomains of Pol. The in vivo packaging domain contains the N-terminal in vitro RNA binding domain (18). The central in vitro RNA bindingdomain is cryptic, inhibited by a region C terminal to it. This cryptic domain and the C-terminal RNA binding domain are located on either sideof the RNA-dependent RNA polymerase consensus domains and therefore may function in polymerization.

with the packaging signal and the 3'-end replication initiationsite, and begins to synthesize negative strands to form dsRNA.In synthesizing negative strands, Pol must presumably releasethe packaging signal. Once dsRNA has been formed, theenzyme initiates transcription to make the viral positive strandswhich are extruded to act as mRNA and to be packaged in newparticles. Each round of transcription must involve binding to

the one end and eventually releasing the RNA when transcrip-tion is complete.

In view of this sequence of binding and releasing of thegenomic RNA at various stages of its replication and packag-ing, we expect that the cryptic binding domain that we describehere will prove to be a part of these processes and that othersuch cryptic domains will be found in other RNA viral proteins.

gag

1

VOL. 68, 1994

m r.,55en., m %pw%o

6020 RIBAS ET AL.

We find that the cryptic binding domain is essential forpropagation of the M, genome but not for packaging of viralRNA. We have not, however, shown that the minimal changenecessary to eliminate the cryptic binding is the minimalchange necessary to eliminate Ml propagation. The proximityin Pol of the cryptic RNA binding domain to the RNA-dependent RNA polymerase consensus domains suggests thatit is involved in some step in polymerization.

Koonin's alignment of RNA-dependent RNA polymerasesof dsRNA and positive-strand ssRNA viruses (15) includes aconserved domain I which in L-A is Pol residues 391VNKMPKHK398, which is inside the cryptic binding domain (374 to432). However, we find no homology of this cryptic domainwith known RNA binding domains in other proteins.We have now defined three in vitro RNA binding domains of

Pol (Fig. 6). The N-terminal domain is part of the larger in vivoRNA packaging domain, while the central cryptic domain andthe C-terminal domain may, on the basis of their proximity toRNA polymerase motifs, function in polymerization. One goalof these studies is to define the detailed mechanisms by whichthese RNA binding domains interact with sites on the viralRNA.

REFERENCES1. Blanc, A., C. Goyer, and N. Sonenberg. 1992. The coat protein of

the yeast double-stranded RNA virus L-A attaches covalently tothe cap structure of eukaryotic mRNA. Mol. Cell. Biol. 12:3390-3398.

2. Blanc, A., J. C. Ribas, R. B. Wickner, and N. Sonenberg. 1994.His-154 is involved in the linkage of the Saccharomyces cerevisiaeL-A double-stranded RNA virus Gag protein to the cap structureof mRNAs and is essential for Ml satellite virus expression. Mol.Cell. Biol. 14:2664-2674.

3. Bussey, H. 1988. Proteases and the processing of precursors tosecreted proteins in yeast. Yeast 4:17-26.

4. Conde, J., and G. R. Fink. 1976. A mutant of Saccharomycescerevisiae defective for nuclear fusion. Proc. Natl. Acad. Sci. USA73:3651-3655.

5. Dinman, J. D., T. Icho, and R. B. Wickner. 1991. A -1 ribosomalframeshift in a double-stranded RNA virus of yeast forms agag-pol fusion protein. Proc. Natl. Acad. Sci. USA 88:174-178.

6. Esteban, R., T. Fujimura, and R. B. Wickner. 1989. Internal andterminal cis-acting sites are necessary for in vitro replication of theL-A double-stranded RNA virus of yeast. EMBO J. 8:947-954.

7. Esteban, R., and R. B. Wickner. 1988. A deletion mutant of L-Adouble-stranded RNA replicates like Ml double-stranded RNA. J.Virol. 62:1278-1285.

8. Fujimura, T., R. Esteban, L. M. Esteban, and R. B. Wickner. 1990.Portable encapsidation signal of the L-A double-stranded RNAvirus of S. cerevisiae. Cell 62:819-828.

9. Fujimura, T., J. C. Ribas, A. M. Makhov, and R. B. Wickner. 1992.Pol of gag-pol fusion protein required for encapsidation of viralRNA of yeast L-A virus. Nature (London) 359:746-749.

10. Fujimura, T., and R. B. Wickner. 1988. Gene overlap results in a

viral protein having an RNA binding domain and a major coatprotein domain. Cell 55:663-671.

11. Fujimura, T., and R. B. Wickner. 1988. Replicase of L-A virus-likeparticles of Saccharomyces cerevisiae. In vitro conversion ofexogenous L-A and Ml single-stranded RNAs to double-strandedform. J. Biol. Chem. 263:454-460.

12. Fujimura, T., and R. B. Wickner. 1992. Interaction of two cis siteswith the RNA replicase of the yeast L-A virus. J. Biol. Chem.267:2708-2713.

13. Icho, T., and R. B. Wickner. 1989. The double-stranded RNAgenome of yeast virus L-A encodes its own putative RNA poly-merase by fusing two open reading frames. J. Biol. Chem. 264:6716-6723.

14. Klebe, R. J., J. V. Harris, Z. D. Sharp, and M. G. Douglas. 1983.A general method for polyethylene glycol-induced genetic trans-formation of bacteria and yeast. Gene 25:333-341.

15. Koonin, E. V. 1992. Evolution of double-stranded RNA viruses: acase of polyphyletic origin from different groups of positive-stranded RNA viruses. Semin. Virol. 3:327-339.

16. Kunkel, T. A. 1985. Rapid and efficient site-specific mutagenesiswithout phenotypic selection. Proc. Natl. Acad. Sci. USA 82:488-492.

17. Mandecki, W., B. S. Powell, K. W. Mollison, G. W. Carter, andJ. L. Fox. 1986. High-level expression of a gene encoding thehuman complement factor CSa in Escherichia coli. Gene 43:131-138.

18. Ribas, J. C., T. Fujimura, and R. B. Wickner. Essential RNAbinding and packaging domains of the Gag-Pol fusion protein ofthe L-A double-stranded RNA virus of Saccharomyces cerevisiae.Submitted for publication.

19. Ribas, J. C., and R. B. Wickner. 1992. RNA-dependent RNApolymerase consensus sequence of the L-A double-stranded RNAvirus: definition of essential domains. Proc. Natl. Acad. Sci. USA89:2185-2189.

20. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecularcloning: a laboratory manual, 2nd ed. Cold Spring Harbor Labo-ratory Press, Cold Spring Harbor, N.Y.

21. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequencingwith chain-terminating inhibitors. Proc. NatI. Acad. Sci. USA74:5463-5467.

22. Sherman, F. 1991. Getting started with yeast, p. 3-21. In C.Guthrie and G. R. Fink (ed.), Guide to yeast genetics andmolecular biology, vol. 194. Academic Press, San Diego, Calif.

23. Valle, R. P. C., and R. B. Wickner. 1993. Elimination of L-Adouble-stranded RNA virus of Saccharomyces cerevisiae by expres-sion of gag and gag-pol from an L-A cDNA clone. J. Virol.67:2764-2771.

24. Vernet, T., D. Dignard, and D. Y. Thomas. 1987. A family of yeastexpression vectors containing the phage fI intergenic region. Gene52:225-233.

25. Wickner, R. B. In B. N. Fields, D. M. Knipe, and P. M. Howley(ed.), Fields virology, 3rd ed., vol. 1, in press. Raven Press, NewYork.

26. Wickner, R. B., T. Icho, T. Fujimura, and W. R. Widner. 1991.Expression of yeast L-A double-stranded RNA virus proteinsproduces derepressed replication: a ski- phenocopy. J. Virol.65:155-161.

27. Wickner, R. B., and M. J. Leibowitz. 1976. Two chromosomalgenes required for killing expression in killer strains of Saccharo-myces cerevisiae. Genetics 82:429-442.

J. VIROL.