Bovine leukemia virus RNA sequences involved in dimerization and specific gag protein binding: close relation to the packaging sites of avian, murine, and human retroviruses.

CopyrightC)1993, AmericanSocietyforMicrobiology

Bovine

Leukemia

Virus RNA

Sequences

Involved in

Dimerization

and

Specific

gag

Protein

Binding:

Close Relation

to

the

Packaging

Sites of

Avian, Murine,

and

Human

Retroviruses

IYOKOKATOH,1* TERUOYASUNAGA,2t ANDYOSHIYUKIYOSHINAKA'

Microbiological Research Institute, Otsuka Pharmaceutical Co., Ltd., 463-10Kagasuno, Kawauchi, Tokushima 771-01,1and Computation Center, TheInstituteof Chemical and

Physical Research, 2-1 Hirosawa, Saitama 351_01,2Japan

Received 1September1992/Accepted 15 December1992

In vitro detection ofa specific complex of the bovine leukemia virus (BLV) MA(p15) protein and the 5'-terminalRNAdimer led to thehypothesisthat theNH2-terminaldomain ofretrovirusgagproteinprecursor

isinvolved intheselective viral RNApackagingmechanism. Herewedescribemappingof the BLV RNA for

dimer-formingandMA(plS)-bindingabilitiesbyasimplecDNAprobingmethodfollowedbymutationanalyses

with thereactiveU5-5'gagRNA. The RNAdimerization is mediated bytheregionharboringU5,theprimer binding site (PBS), and the 30 bases immediately downstream of PBS. This conclusion is supported by computer-assisted RNA secondary-structure analysis which predicted a multibranched stem-loop folding

throughoutthe dimerregiondetermined. Anotherregionfrom PBS tothe 5'-terminal 60 residues ofthegag gene, partially overlapping the dimer region, likely provides essential elements for the MA(pl5) binding reaction, although the presence of either the 3' or5' neighboring sequences increases the complex-forming efficiencysignificantly, and each of thesubstructures predicted withinthecore region has,ifany,onlyvery

weakaffinitytoMA(p15).Thesein vitro characterizations of the BLV RNAmayreflectgeneralfeatures of the specific protein-RNAinteraction in thepackagingeventsof variousretroviruses.5'-terminal folded structures ofretroviral RNA molecules and theirbiologicalactivitiesarediscussed.

Although the molecular mechanism ofretrovirus replica-tion has been characterizedtoagreatextent,several critical

processesremain tobeexplained. Amongthem isthestage

of virusparticle formation, the stage atwhich assemblyof thegagprecursor polyprotein andspecific association ofa

domain of theprecursor with viral RNAare thought to be primary biochemical events leading to productive virus morphogenesis.

Two schools of evidence have beenpresentedin thestudy of the virus protein-RNA interaction. Results of in vitro mutagenesis experiments emphasizedthatthe COOH-termi-nalnucleocapsid (NC) domain of thegagprecursorand its zincfinger-like motifsarerequiredfor viral RNApackaging (3, 9, 10, 15, 29-31). Direct RNAbinding experimentswith purifiedviral proteins suggested that the NH2-terminal do-main ofthe matrix-associated (MA) proteinorits adjacent

polypeptide plays the key role (17, 21, 39, 40). The NC proteinwasshowntobeanonspecificRNA-binding protein

(16, 17, 20) thatformstheviralnucleocapsid complex. On the otherhand,locationof thepackaging signalin the

genomeRNAhas been assessedbydeletion and recombina-tion analyses, andno significant difference amongthe rela-tivepositionsdetermined for retroviruses of variousspecies has been found(3, 18,23, 28, 33, 41, 43, 45). Recent results ofdetailed examinations ofmurine leukemiavirus(MuLV) and Roussarcomavirus, however, indicate that packaging

sequences should be in a more extended region than was

originallydetermined(1,5-7, 24). Another intriguing feature of retroviralgenome encapsidation isthat theviral RNA is incorporated into the particles as a dimer of the identical

*Correspondingauthor.

tPresentaddress: GenomeInformationResearchCenter,Osaka

University, 3-1Yamadaoka, SuitaCity,Osaka565, Japan.

molecule. It has longbeenpostulated that the dimer struc-ture is closely related to the packaging events. Although hypotheticaldimermodelswerededuced from possible base

pairingswithin the 5'-terminal RNAsequences, their

prob-abilities have not been experimentally validated (8, 13). Recently, thepresence ofafoldedstructureofa stem-loop

with multiplebranches has been reported for thecis pack-aging element of human immunodeficiency virus type 1 (HIV-1) on the basis of computer analysis in combination withenzymeand chemicalprobing experiments (12). Asyet,

however, no direct in vitro evidence points to the folded structureasthesignal recognized by the HIVgagprecursor

protein (26).

Hereweshowthatminimum elements required for bovine

leukemia virus(BLV) RNA dimerization and complex for-mation with MA(p15)arelocated inaregion corresponding to the packaging site determined for various retroviruses. Ourelectrophoretic analysisoftheprotein-RNAinteraction

may account for one of the critical steps of retrovirus packagingofdimericviralRNA.

MATERUILSANDMETHODS

Purification ofMA(pl5). BLVwasprepared from culture

fluid of Bat2Cll cells (36) as described previously (17).

MA(p15)waspurifiedfromthe BLVpreparations byacetone

treatmentfollowed by Sephacryl S-200 column chromatog-raphy and reverse-phase high-performance liquid chroma-tography (RP-HPLC) as described previously (17). The purified protein fractionwaslyophilized, solubilized in H20

with 0.025% Nonidet P-40, and quantitated by the Lowry method (25). TheMA(pl5) solutionwas adjustedto 50 ,uM andstoredat4°C.

Synthesis of RNA fragments. Viral RNA fragments and

1830

on November 9, 2019 by guest

http://jvi.asm.org/

DIMERIZATION AND gag PROTEIN BINDING IN BLV 1831 their mutants weresynthesizedby the SP6 RNApolymerase

reaction as describedpreviously (17). 32P-labeledU5-5'gag and mutant RNAs wereproduced in reaction buffer (40 mM Tris-HCl [pH 7.5], 6 mMMgCl2, 2 mM spermidine, 10 mM dithiothreitol, 1 U of human placenta RNase inhibitor (Takara) per

p1)

containing 0.5 mM each GTP, UTP, and ATP, 0.05 mM CTP, 0.025 mmol [-32P]CTP, 2 ,g of linearized template DNA, and 20 U of SP6RNApolymerase (Takara)inatotal volume of 20

P,l.

Incubationwasfor 1 hat

37°C. Unlabeled U5-5'gag was synthesized by incubating for 16 h at 37°C the followingreaction mixture(50

jlI):

the same buffer containing 0.5 mM all four ribonucleoside triphosphates, 1 ,ug of pAM5G cleaved by SailI (pAM5G/ Sall; SalI from Boehringer Mannheim) per

PIu,

and30 U of the enzyme. Afterthepolymerase reaction, the templatewas removed by DNase I (8 U) digestion for 10 min. Products were deproteinized bytwo extractions with phenol-chloro-form-isoamyl alcohol (25:24:1) followed by one extraction withchloroform-isoamyl alcohol(24:1)and thenprecipitated with ethanol.Dried RNAwasdissolved inH20 and quanti-tated. Aliquots were stored at -70°C. For the MA(p15) bindingassay, RNAsolutionwasmadeonice in 2x binding buffercontaining 20 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonicacid(HEPES;pH7.9), 12mMMgCl2,20 mM NaCl, 2 mM dithiothreitol, and 1 U of RNase inhibitor per

PI.

cDNA probing. DNAs consisting of 18 or 30 bases com-plementary tothe U5-5' gag RNA sequences were synthe-sized by a DNA synthesizer (Applied Biosystems model 381A) and purified by RP-HPLC through a C18 column

(Waters)

withanacetonitrile

gradient

from 5to20%in 0.1 M triethylamine acetate (pH 7.0). In one set of experiments, syntheticcDNAswere5' end labeled with

[_y-32P]ATP

(New England Nuclear) and T4 polynucleotide kinase (New En-glandBioLabs), phenol extracted, and ethanolprecipitated. Each labeled cDNAwas addedtotheSP6 RNApolymerase reaction mixture (20

P,)

containing pAM5G/SalI (2

P,g)

and fourunlabeled ribonucleosidetriphosphates priorto incuba-tion. In another set of cDNA probing assays, unlabeled syntheticcDNAwasaddedtothe RNApolymerase reaction

to produce U5-5' gag labeled with [a-

2P]CTP.

Unlabeled ATP, GTP, andUTPwereused at0.5 mM each; CTPwas

adjusted to make a final concentration of0.075 mM. Incu-bationwasfor 1 hat37°C.DNasetreatment waseliminated. Construction ofplasmidsfor theexpressionofmutantRNA segments. The wild-type plasmid previously referred to as pAMGhas been renamed

pAM5G

for easieridentification in comparisonwith various DNAs used in this work. Deletion and fusion mutantswere newly constructed from

parental

plasmids

pAM5G

and pAMGPE

(17).

Their identities and construction methods are described below. Nucleotide po-sitionswere numbered in relation to the

transcription

start site, +1. DNA fragments generated by restriction enzyme digestionsarealways indicated belowasA-B,

representing

thesequencesfrom the5'end

(A site)

tothe 3' end

(B site)

intheorientation of theBLV

coding

sequences

(38).

Inthe

case of noncohesive

ligation,

ends were blunted with T4 DNApolymerase priortothe

ligation

reaction unless spec-ified otherwise. Plasmids obtained from transformants of EscherichiacoliHB101werescreened

by

restrictionenzyme mapping.

Construction of deletion mutants. (i)

pAM5GdP.

Plasmid pAM5G was cleaved into twofragments by

digestion

with BclI (at position

+340)

and XmnI

(at

a

unique

site in the vector, pAM18

[Pharmacia]).

One

fragment,

XmnI-BclI

(+340),

was

digested

with BstNI

(+301)

to remove the

primer bindingsite

(PBS)

and

ligated

totheother

fragment,

BclI

(+341)-XmnI.

(ii)

pAM5Gda.

The BLV U5-5' gag sequences, +147 to

+940,

were excised from

pAM5G by

digestion

withEcoRI and

SphI (at

the

unique

cleavage

sites derived from the

multiple

cloning

site of the

vector)

andfurther

digested

with AvaIl

(+369)

to remove the 5'-terminal sequences. The deleted

fragment,

AvaII

(+370)-SphI,

waslinkedtotheSall

(in

the

multiple

cloning

site)-BclI (+340) fragment

which containedthevectorand the BLV sequencesfrom +147to +340.

(iii)

pAM5GdPdK.

pAM5GdPdao

was constructed

by

the cohesive

ligatipns

of four DNA

fragments,

Avall

(+370)-SphI,

SphI-XmnI,

XmnI-BstNI

(+321),

and aBstNI-AvaII

adaptor

made

by annealing

two strandsof

synthetic

oligo-nucleotides,

5'AGGAGA3' and3'CCTCTCTG5'.

(iv)pAM5Gd,3.

pAM5G

wasincubatedwithAvaIItomake a

partial

digestion.

Linear forms were

isolated,

further cleaved

by

BssHII

(+487)

to removeAvaII

(+373)-BssHII

(+487),

andrecircularized.

(v) pAM5Gded3. BclI

(+345)-BssHII (+487)

sequences wereremovedby

digestion

of

pAMSG

with thetwoenzymes and recircularized.

(vi)

pAM5GdPdcKdI.

TheXmnI-BstNI

(+301) fragment

was linkedto the BssHII

(+488)-XmnI fragment.

(vii) pAMLG.The BstNI

(+302)-SailI

(+936) fragment

was isolated from

pAM5G

andinserted intothe

pAM18

vector at theSmaI

cloning

site.

(viii)

pAMgag.

TheSacl

(+147)-DraI

(+535)

region

was

deleted from

pAM5G by

digestion

of thetwoenzymes. Construction of recombinantmutants.

HindIII

(in

the

mul-tiple cloning

site)-BamHI (+5015)

sequences

encoding

the3' end of gagtothe5' end ofenv wereremovedfrom

pAMGPE

to obtain

fragment

BamHI

(+5016)-HindIII (AME),

which

corresponds

tothe linear formof

pAME

(17).

Recombinants between U5-5' gag and the env gene were

generated by

inserting

segmentsfromU5-5' gag betweenthe

HindlIl

and BamHIendsof the AME

fragment.

(i)

pAMSLE

and

pAM(5L)E.

The SacI

(+ 147)-BssHII

(+488) fragment

wasisolated from

pAM5G

andlinkedtothe AME

fragment.

The

resulting plasmids

with sense and antisense insertions werecalled

pAM5LE

and

pAM(5L)E,

respectively.

(ii)

pAMLE. The HaeII

(+286)-BssHII (+487) fragment

was

ligated

toAME.

(iii)

pAM5E.

Similarly,

the Sacl

(+147)-HaeII

(+290)

sequenceswereinserted 5' to theenv

region.

(iv) pAMa+E.

Synthetic

double-stranded DNAfrom po-sitions +322to+386 with 5'HindIIIand 3' BamHIcohesive endswas linkedtoAME.

(v)

pAMa+,3E.

The

BamHI(+5016)-ApaI

(+5206)

region

of

pAMa+E

was

replaced by synthetic

DNAfrom

positions

+410to +471with 5' BamHI and 3' ApaI

sticky

ends.

(vi)

pAMPE.

pAMa+,BE

was cleaved

by

HindIII

and BamHIto remove the

at

sequencesandwasthen recircular-ized.

Plasmids

containing

the gag sequences were cleaved

by

SalI,

and those with the env sequences were cleaved

by

AvaII;

the

plasmids

werethen

phenol

extractedandethanol

precipitated

asdescribedaboveto serveasthe

templates

for the RNA

synthesis

reaction.

Gelmobility shift assay. The RNA

gel

mobility

shift assay was

performed

by

modificationof the methodsdescribed for the HIV Rev

RNA-binding

protein

(14,

27)

and those for various

transcription

factors that bind to their

target

DNA sequences

(22). Briefly,

RNAsolution in 2x

binding

buffer VOL.67, 1993

on November 9, 2019 by guest

http://jvi.asm.org/

BLV

A U5-5gag

(Sac e

+147

r

PBS leader

L--~~~~~~~~~~~

---B cDNAssynthesized

4A4 9A4 9.2 4I1

tt:

3. 3.3

9.3 9.1 1.2 72 1.4 3.2 34

13

C Dimer formation

D MA(p15)-binding

.. :] 3~~~.

FIG. 1. Summaryof cDNAprobingassays.(A)Structure of the BLV U5-5'gagRNAfragment.RNAencompassingtheSacIsiteat +147

totheSalI siteat +936wasinvitrosynthesized. R, U5, PBS, leader,andgagregionsaremapped. The 5'splicedonorsite inthegenome RNA is 47 basesupstreamfrom the 5' end of the U5-5'gagsegment.(B)cDNAprobes. DNAssynthesizedforprobingassaysareindicated

by bars placedintheir targetsequences.cDNA 4.4wascomplementaryto+95to +124,whichwerenotcontainedinthe RNAfragment. (C) Inhibitionof RNA dimer formation. cDNAs that interferedwiththeRNAdimer formation in at leastoneofthetwo sets ofprobingassays areshownbyfilled bars. Probes9.1, 4.1, 1.1, 7.2,and1.3hadsequencescomplementaryto +245 to+274,+275 to+304,+322to+339,+340

to+369,and +350to+379, respectively. (D)Alteration ofthe RNAmobilityshift. cDNAs1.1,7.2, 3.1,and3.2,which influencedcomplex formation with MA(p15), are indicatedby dotted bars. Probes 3.1 and 3.2 were complementary to +410 to +439 and +440 to +469, respectively.

wasmixedwithMA(pl5) proteinsolutionataratio of 1:1on

ice and incubated for 10 min at 20°C. One volume of the loading buffer (30% glycerol, 0.02% xylene cyanol, 0.02% bromophenol blue)wasaddedper5 volumes of the reaction mixture,from which 3 plwasloadedperslot. Sampleswere runfor 3 hat120 Vat4°C.Theelectrophoresis gelwasmade of2.5%polyacrylamide and0.5%agaroseinTBEbuffer(90 mMTris-borate, 2 mM EDTA [pH 8.0]).

AssayresultswerenotaffectedeitherbyNaCl concentra-tion in thebindingreaction withinarangefrom10to50mM

orby pHbetween 7.5 and 8.0. Thepresenceof100mMNaCl reduced the band shift efficiency of U5-5' gag by 30% without relaxingthebinding specificityofMA(pl5).

Prediction of theBLVRNAsecondarystructure.

Minimum-energysecondarystructureswereconstructedbyoperating a SUN computer with programs FOLDRNA and SQUIG-GLES supplied by the University of Wisconsin Genetics Computer Group. BLVsequence datawere fromreference 38.

RESULTS

cDNAprobingassays.To determine thefunctionalmapof the U5-5'gagRNA for dimerization and MA(p15)-binding activities,weappliedasimplecDNAprobingmethodonthe basis of the following assumptions: that short single-stranded DNA probes complementary to the U5-5' gag

sequencescan hybridizeto the nascent RNA chains

gener-atedbythepolymerasereaction,which results inmasking of the activesequences and/or disruption of theinherent

fold-ing by imposition of an RNA/DNA double helix; and that subsequentelectrophoretic analysisincombination with the proteinbindingreactionpermits theestimation of the active regions.TwosetsofcDNAprobingassays werecarriedout asdescribed in Materials andMethods. cDNAprobes of18 or30basesweresynthesizedtoscantheRNAsequences as

shown inFig. 1. Probe 4.4wasdirectedto30basesin the3' partofthe Rregion,whichisnotcontained intheU5-5'gag

RNA andtherefore usedas aninactive cDNAprobe.

In thefirst assay, eachcDNAwas 32Pphosphorylated at

the 5' terminus and added to the SP6 RNA polymerase reaction. Recovered RNA was electrophoresed through a

nondenaturing gel to monitor dimerization and complex formation efficiency. Typical results are shown in Fig. 2A. The cold U5-5' gag RNA probed with various 32PcDNAs (odd-numbered lanes with the exception of lanes 1 and 9) appeared as dimer and monomer bands at the positions corresponding to those of the internally labeled U5-5'gag

RNA (lanes 1 and 9). Probe 4.4 failed to forma detectable

band(lane17). Unhybridized probemolecules madeabroad bandatthe bottom of eachlane.

As described previously, the U5-5'gag RNA efficiently

dimerizedtoform the dimer andmonomerbands inaratio of

4:1onthe basis of theradioactivitycountineachband(lane 1). Hybridization ofprobes 1.1, 7.2, and 1.3 (lanes 21, 23, and 13) affected the RNAdimerization to alter theratio of dimer/monomer populations to 1:4, 1:5, and 1:1,

respec-tively. Other cDNAs didnotappearto impair dimer forma-tion, althoughtheir hybridization efficiencies were not sig-nificantly differentfromthoseofthethreeinhibitory probes,

ascompared bythebandintensities.

After incubation with the MA(p15) protein, U5-5' gag

formed a shifted band presumed to contain the U5-5'gag

dimer(Fig. 2A, lanes 2 and10). Itis thought that MA(pl5) hasanRNA-annealing activity concomitant with the binding

activity (17). Bymild heattreatmentofthe RNA priortothe proteinbinding reaction, both the dimer andmonomershifts

becamedetectable in thesamelane,asdescribed previously.

On a 3.5% polyacrylamide gel, the two band shifts were

detectedby 30% reductionin theirmobilities. In aparallel

experiment on a 2.5% polyacrylamide-0.5% agarose

com-posite gel, their shifts were found as 60 to 70% slower migration than for the free dimers and monomers. From

theseobservations, shifted bands detected in thisstudywere

judged bytheir relativemobilitiestorepresenteither dimer

ormonomercomplexes.

Many of the

[32P]cDNAIRNA

hybrids normally reacted withtheprotein,asjudged by theappearanceof the shifted

(SaI)

+940

I

|~ R _U5

₇₁

_gag

on November 9, 2019 by guest

http://jvi.asm.org/

[image:3.612.77.536.71.233.2]

DIMERIZATION AND gag PROTEIN BINDING IN BLV 1833

A

SC-U5-5g;g 9 2 91 12 -3.J3; 933 :,

MA(D15) - + + + +

sd- g g

d- _

m-UUI

--'P WNA

2 3 4 5 6 7 t1

4.4 4.1 1.1 7.2

MA(p15)- + -+ + +

id-g

17 11 2 13 1. '

73 3.1 3.2

_ _

* .

._

B

4 9.1 92 93

-MA(p15 - + _ +

-bd- _S_J*

m-4

2 3 4

FIG. 2. Resultsof cDNA probingassays.(A) [32P]cDNA probing

assay. Unlabeled US-S' gag synthesized in the presence of

32P-labeledcDNAwasanalyzed bypolyacrylamide-agarose gel

electro-phoresis after incubation with (lanes +) or without (lanes -)

MA(p15). Control experiments with 32P-labeled US-S' gag are

shownin lanes 1, 2, 9, and10. Input cDNAis indicated for eachset

ofexperiments. Positions of dimers (d),monomers(m), andshifted

dimers (sd) ofU5-5'gagRNA are indicated. The position offree

[32PJcDNA is also indicated. Lanes are numbered at the bottom.

Binding of 9.3,9.4, 3.3, and3.4did notinterfere with either dimer

formation or protein binding. (B) Probing assay with unlabeled

cDNA.Resultswith probes4.1, 9.1, 9.2, and 9.3aswellas results

ofacontrol experiment(-)are shown. Probes 4.1and 9.1 impaired

thedimerformationof32P-labeledUS-S'gagand also affectedband

retardation. Results with probes 9.2, 9.3, and other cDNAswere

similartothose shown in panelA.

dimer band (Fig. 2A, lanes 4, 6, 8, 12, 16, 20, and 26). In contrast, U5-5' gag hybridized with probe 1.1 or 7.2 exhib-ited unusual electrophoretic features after incubation with MA(p15). As observed in lane 22, the dimers of the RNA/1.1 hybrid shifted normally, while the monomers stayed at the freeposition. The

RNA/7.2

hybrids made noclear bandshift and formed a diffusion through lane 24. Another notable profile was observed with probes 3.1 and 3.2 (lanes 27 to 30). Although dimers were formed in the same efficiency as in the control

U5-5'

gag RNA, disruption of the shifted band was detected (lanes 28 and 30). An extended smear was formed along lane 30.

In the second cDNA

probing

assay, the RNA polymerase reaction contained

[ax-

2P]CTP

as a labeled nucleoside triphosphate and unlabeled probe cDNA so that effects of the input probe were monitored by the radiolabel within the body of the RNA. Results

obtained

in this probing method were similar to those in the first experiments; however, probes 4.1 and 9.1 obviously inhibited dimerization (Fig. 2B, lanes 1 and 3). The dimerized RNA in the presence of probe 4.1 or 9.1 was observed to form complexes with MA(p15). The monomers were less reactive than the control mono-mers, and long smears were found between the shifted dimers and free monomers (lanes 2 and 4). Probes 4.1 and 9.1 had a complementary stretch within their sequences, which may have resulted in a different binding feature after a phosphorylation reaction containing spermidine.

From the results of the two cDNA probing experiments, dimer formation was suspected in the sequences covered with probes 9.1, 4.1, 1.1, and 7.2. Since probes 1.3 and 7.2 had 19 bases in common, the weak interfering activity of probe 1.3 was represented by probe 7.2. The MA(p15) binding was expected to involve the dimer-forming regions and the sequences covered with 3.1 and 3.2 (Fig.

1C

and D). For convenience, the 7.2-, 3.1 + 3.2-, and 4.1 + 9.1-directed regions are referred to as cx,

,B,

and -y, respectively.

Deletion

analysis

of U5-5' gag. To confirm that thea, 1,

y,

and PBS regions are related to dimer formation and complex formation, a deletion experiment was carried out. Structures of the mutant RNAs synthesized are illustrated in Fig. 3A. In the structure names, dP,

dot,

and

do

are used to represent specific deletions as follows: dP, deletion of the

3'-terminal

17 or 12 bases of

US

and the adjacent 18 bases of PBS; da, deletion of the 25 bases of the otregion; and

d,B,

deletion of

the sequences from +373 to +487 harboring the 1 region. The wild-type and mutant RNAs were

synthesized

with 32p label in parallel reactions and analyzed for their dimerization and MA(p15)-binding abilities.

Only 20% of the RNA molecules in the 5Gdcx, 5GdP, and 5GdPdcx preparations were dimerized (Fig. 4A), suggesting that removal of either the PBS orcxregion results in a loss of the dimerization potential ofU5-5'gag. These mutant mono-mers did not make a discrete band shift after the MA(p15) binding reaction (lanes 4, 6, and 8), while the dimers were able to make a clear shift as did the wild-type RNA (lane 2). Deletion of the 13-containing region did not alter the dimerization efficiency (lane 11). However, only 40% of the 5Gd1 dimer population participated in the complex forma-tion with MA(p15) (lane 12). Moreover, the dimer complex did not appear as a firm band. These results indicate that the 1 region is not required for RNA dimerization but may be involved in the protein binding reaction.

5Gdcxd1 and 5GdPdcxd13 also poorly dimerized (Fig. 4A, lanes 13 and 15). Their dimer shifts (in lanes 14 and16) were not solid as those of 5Gdcx, 5GdP, and 5GdPdcx.Thecontrol gag RNA made no band shift after the protein binding

VOL. 67, 1993

on November 9, 2019 by guest

http://jvi.asm.org/

[image:4.612.72.306.109.560.2]

,+147 U5-5'gag I R_

HaeIBStNI B IIAva I

rU

I U5 l--I

PBS

Bs8 Dra +940

r+417

-I-

~~

_gaggag

-

1

-i R

I

R

I

zR

I

R

R

I

R

+245 +304 +340 +369 +410 +46

r- -I r r

'Y Of0a

+302 r+341 U5

I...=~

+370 30 ...

U5 [...

U5

U5

69

gag

Wag

IHz

ITh:

+3M +488

+344,

.

.

[image:5.612.134.469.74.422.2]

+302.

...

:1

ga

"

r+302ga

,+535

ga

IPr-Bam H Aja Ava I

1 +5501

L2IZ

env

II

L

R r +491

a,ll

It

+147,

l

gn

I

u [

*{i1L

fwIIi1Ii

r+2_IP

E~Z

LZIrIIIII

Lz~cI

1111

env

1

R

+289-i .

...

env

l er r+322 +386, r+410 .471, r5203

M~~~~~~~C~

a

lZ~EIIVII

W

...-

env

I

I

t

FIG. 3. Structures of deletion and fusion mutants.Wild-typeU5-5'gagismappedforU5, PBS, leader,andgag regions. Positions of restrictionenzymecleavagesites used for theconstructionof mutanttemplatesareshown. y,PBS,a, and regionsare also indicated. (A)

RNAs with various deletionsin the leaderand/or5' terminusof thegagregion.5GdP, SGda, 5GdPda, 5GdI, 5GdadI3,5GdPdadp,LG,and gag were synthesized by SP6 RNApolymerase reactions using as templates DNAs pAM5GdP, pAM5Gda, pAM5GdPda, pAM5Gdp, pAM5Gdadp3, pAM5GdPdadI3, pAMLG, and pAMgag, linearized by SalI digestion. Deleted sequences are indicated by dotted lines. Nucleotide positionsarealso indicated.(B)Recombinant RNAsmade between the 5'regionand theenvregion.E,5LE,(5L)E,LE,Pa+ ,E, Pa+E,andPEweresynthesizedfromtemplates pAME, pAM5LE,pAM(5L)E, pAMLE, pAMPa+,3E, pAMPa+E,andpAM,3Ecleavedby

AvaI. Linkages generated bygenemanipulationareshownbydotted lines. 5'-terminalsequencesarealignedwiththewild-typeU5-5'gag

shown above.

reaction (lane 10), although the dimerandmonomer bands appearedtobeslightlysmeared.

Thus,in agreementwiththeresults obtained in the cDNA probing assays, the deletion analyses (Fig. 4A) suggested that the PBS and a regions contribute primarily to dimer

formation and that a sequences contribute to MA(p15) binding.

To compare the affinities for MA(p15) amongwild-type andmutant RNAs, we next performed competition assays

with unlabeled U5-5'gagRNAasdescribedpreviously (Fig. 4B). Inthewild-type RNA, 90and 40% of the total radioac-tivities were retained in the shifted dimer band in the

presence of 10- and 100-fold excess amounts of the cold competitor, respectively (lanes 3 and 4). In contrast, the complexes of 5GdPda, 5GdP, and 5GdPdaodI3 RNAs were

effectively dissociated by a 10-fold concentration of the wild-typeRNA. Less than 5% of theinput 32pcountswere

found in the retarded bandin lanes 7, 10, and 13.

Anothermutant,LG, consistingof the whole Lregionand thegagsequences,hadaninteresting phenotype.The rate of

dimerization (Fig. 4B,lane 14)was comparable tothoseof 5Gdat, 5GdP, and their relatedmutants. Despite the lackof sufficient dimerization activity, the LG RNA sustained a

moreintensereactivitywithMA(p15)thandid5Gdot,5GdP,

5Gd,B, or the related mutants. The LG RNA made an

additional band of slowermigration representing the mono-mercomplexaswellasthe band ofshifted dimers (lane15). Inaddition,one-fourth of theinputmonomerpopulationwas

incorporatedinto the dimercomplexesafterincubation with MA(p15);the dimercomplexinlanes 15 and 16 containedan

increased 32P radioactivity. Moreover, the LG

dimer-MA(p15) complex was retained in the competition assay

with 10-fold concentration of the wild-type RNA (lane 16) butnotthemonomer-MA(pl5) complex.Theyregionlikely

participates in the RNAdimerization but is lessresponsible forMA(p15) bindingandtheMA(p15)-induced RNA dimer-ization.

Experiments with recombinant RNAs. We tested whether the y, PBS, a, and ,B sequences can exert their functions when linked 5' to the env sequences that are not directly A

5GdP 5Gda

5GdPda

5Gdf

5Gdadf

5GdPdad,O

LG

gag

B E

5LE (5L)E

LE

5E

PcxOE

Poi+E

OE

u

WV

n

11

on November 9, 2019 by guest

http://jvi.asm.org/

DIMERIZATION AND gag PROTEIN BINDING IN BLV 1835

AU5-5gag 5GdP 5Gdit 5GdPd/ gag 5GdrJ 5Gd1dr: 5GdPd ,d,,

MA(p15) - + - + - + - + - + - + - + - +

--wo 4

64.

t

0 _ * _a

ea so

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

U5-S gag 5Gdpdci 5Gd3 5GdPdad3

MA(p15} + i + - + + + +

-wilabeledU5-5gag - - lOx lOOx - - lOx - - lOx - - lox

LG

+ l 1 - - 10K lOOx

Fi> > > >

Z~~~~~~~~-

0 ; Z

seem

00 **

sa-.-4

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

FIG. 4. Analysis of deletionmutantsfor dimer formationand complex formation with MA(p15). (A) MA(plS)bindingassaywithmutants

havingdeletions in the PBS-leader region. The wild-typeandmutantRNAswereincubated with (lanes +)orwithout (lanes -) MA(plS) (25

,uM)andanalyzedon anativegel. Positions of dimersandmonomers areindicated byopenandfilled circles,respectively, for theU5-5'gag

andgagRNAs(lanes1and9).The complex of the U5-5'gagRNAdimerand MA(p15) is indicated byatriangle in lane 2. (B) Band shift and

competitionassays.U5-5'gagandmutantRNAs with deletionsin the PBS-leaderorR-U5 regionwereexamined for MA(p15)-binding ability.

In the competition assays, concentrations of the wild-type competitor, U5-5'gag, are indicated as 1Ox (10-fold) or 10Ox (100-fold) in

comparisonwith that of the labeled RNA. -,reactions without thecompetitor.

relatedtothe packaging mechanism.Aseriesof recombinant RNAssynthesized for thispurposeis shown in Fig.3B, and the results oftheMA(pl5) bindingassay areshown in Fig.5.

Thecontrol envRNA, E,waspoorly dimerized, and it did not make a discrete band shift after the MA(plS) binding

reaction(Fig. 5A,lanes5 and6).

The 5LE RNA with a sense insertion of the 5'-terminal portion ofU5-5'gag formed a solid dimerband containing

40% of the total3ZPradioactivity(Fig. 5A,lane7) and made

aclear dimer shift(lane 8)similartothat ofU5-5'gag(lanes

2to4).Themonomershiftwasnotdetectableas afirm band.

Therecombinant RNA retained theprotein-binding activity in thecompetitionassaywitha10-foldexcessamountof the cold U5-5' gag RNA (lane 9) but not with a 100-fold

concentration of thecompetitor (lane 10).Incontrast,(5L)E with the opposite strand of the U5-leader region was

defi-cient in both self-dimerization and complex formation with MA(pl5) (lanes 11 to 14). These results indicate that the positive strand of the U5-leader segment from positions +147to+491 bears minimum structuralrequirementsfor the

twoRNA functions.

Neither the LE RNA containing the complete leader

sequencesnor5Ewith theU5 regiondimerized sufficiently (Fig. 5B, lanes 1 and9) in comparison with 5LE (Fig. SA, lane 7), ensuring the cooperation of U5 with PBS-leader

region in viral RNA dimerization. LE could make the

monomer and dimer band shifts firmly; shifts were,

how-ever, prevented by the 10-fold concentration ofU5-5'gag

(Fig. SB,lanes 1 to4). ThePa+1E RNA with PBS-a and 1B elementsinatruncated formmadeaband shift profile similar tothat of LE(lanes5to8) but less efficiently.

Dimers of5E behaved like those of5GdPdad3 (Fig. 4B, lanes 11 to 13) after the protein binding reaction (Fig. 5B, lanes 9 to 11), while monomers smeared through lane 10.

Neither PotaE nor ,BE formed a visible dimer band. The

Pa+Emonomersfailedtogiveaclear shift in lane13,while

the ,BE monomers could form a broad and diffuse band

barely separatedfrom theunboundmonomerband(lane 16)

in the absence of thewild-type competitor. The results in Fig. 5B indicate that each of the small elements in the U5-PBS-a-,B regionispoorlyreactive toMA(pl5).

It should be noted that thegag-tailed RNAs, 5GandLG,

were significantly more active in dimerization aswellas in complexformation thanwere the correspondingenv-tailed RNAs,5LEandLE, respectively.Since thegagregionfrom position +535to +940 itself hadadetectabledimer-forming naturebut didnotcause theband shiftwith MA(pl5) (Fig. 4A,lanes 9 and 10), itislikelythat thegagRNA structure influences the overallRNAfoldingandthereby supportsthe upstreamRNA functions.

B VOL. 67,1993

on November 9, 2019 by guest

http://jvi.asm.org/

[image:6.612.148.473.77.387.2]

A

U5-5'gag E 5LE (5L)E

MA(p15) - + + + - + - + + + - + + +

unlabeled U5-5 gag - - lOx100x - - - - lOx100x - - lOx100x

... A

_0

0 as*, d

1 2 3 4 5 6 7 8 9 10 11 12 13 14

B

LE Pt3dE 5E PL+E 3E

MA(p15) - + + + - + + + - + + - + + - + +

unlabeledU5-5'gag - - lox10Ox - - lOx 1OOx - - lox _ - lOx - - lox

0~~~~~~~

[image:7.612.157.440.349.754.2]

1 23 4 56 78 910 11121314 1516 17

FIG. 5. Analysisofrecombinant RNAs for dimerization andMA(p15)-bindingactivities. (A)Results for theparentalRNAs,U5-5 gag, and Eand for recombinantRNAs,5LE,and(5L)E; (B)results for other recombinantscontainingshortsegmentsfrom the5'-terminalregion

andthebodyoftheenvregion.Themonomerbandof eachRNAis indicatedbyafilled circle. Detectable dimer bandsaremarkedwith open circles. Shifted dimers andmonomers areshownby trianglesandfilledtriangles,respectively. SampleswereincubatedwithMA(p15)(+)or

in buffer alone (-). Concentrations of thecompetitorU5-5' gagareindicated as lOx (10-fold)or 1OOx (100-fold) relativetothe testRNA concentration.

Computer-assisted analysis of BLV 5'-terminal RNA fold-ing. The dimer-forming and MA(p15)-binding sequences

determined above were examined for their formation of

secondarystructures by free-energyminimizationanalysis. As shown in Fig. 6A,abranchedstem-loopstructurewas

predictedintheregion encompassingU5tothe axregion. In

theprediction,theU5sequencesformlongstemsII andIV,

a shortstemIII,and associated bulges andloops. ThePBS andasequencesformstemV. Three basepairingsbetween the5' terminus ofU5and the3' terminus ofamakestemI.

In the region,twoseparatestem-loopstructuresVI and VII, consistingof residues +417to +437and +447to+471, respectively, were predicted (Fig. 6B). Stem-loop VI is

complementary to cDNA probe 3.1 and stem-loop VII is complementary to probe 3.2, which may account for their inhibitory effectonthe MA(pl5) bindingreaction(Fig. 2A).

DISCUSSION

Dimerization of BLVandother retrovirus RNA.We have examined thedimer-forming feature of the 5'-terminal BLV RNAbyacDNAprobing method and mutation experiments

with theU5-5'gagRNAfragments.Allof the experimental

results maybe interpreted asimplying that RNA

dimeriza-tion ismediatedby theregion encompassing U5, PBS, and the 3'flanking30 bases of PBS(a).

Computer-assisted RNA sequenceanalysis suggested for-mation ofasecondarystructureofastem-loopwithmultiple branches throughout the entire region determined. In that model,the ysequencescontributetostem-loops II, III, and IV. Elevenresidues of PBS and one base of its 5' adjacent region arefoundtopairwith thealsequencesto makestem V. Because retrovirus RNA dimerization is assumed to result fromintermolecular basepairingsduetothe presence ofpalindromic sequences (8), the model seems to support ourdata.

Stem-loop V is analogous to the secondary structure proposed for avian retrovirus RNAs, in which the 3' U5-PBS region hybridizesthe 5'-terminal leader sequences to build up a more extended stem without branching (2, 13). The involvement of thePBS-aregion in BLV RNA dimerization agreeswith the dimer modelpredicted for avian and murine viruses in earlierstudies (8).

We have alsoanalyzedHIV RNAfor dimer formationby cDNAprobing. The preliminary result shows that the 5'-terminal extended stem-loop II of the reported secondary

on November 9, 2019 by guest

http://jvi.asm.org/

DIMERIZATION AND gag PROTEIN BINDING IN BLV 1837

loop IV

stemIv

loopIII

(7)

(515)

U5

47

+245 (456)

PBS

(a) stemV kader

A

A0A +478 (689)

+410 +406

B (S121) (617)

FIG. 6. RNA secondary structures predicted for the dimer-forming andMA(p15)-binding regions by computer-assisted energy

minimization analysis. (A) Multibranched stem-loop structure

formed by thedimerization sequences +228 to +366. Nucleotide numbers originate from the RNAcapsite(+ 1). Position numbers in

thesequenced provirusgenome(38)arealso shown in parentheses.

Stems ItoIVandassociating loopsareindicated. Thestart(+245)

and end (+304) positions of the -y region and ofPBS (+322 and

+339) aremarked. Nucleotide numberings in the proviralgenome arealsoindicated in parentheses. (B) Twostem-loops found in the ,B

region. Sequences from +406to+478wereanalyzed for secondary

structure. The AUG translation start codon of thegag gene is marked by dots. The ,B region encompassed nucleotides +410 to

+469. Stem-loopsVIand VIIareindicated.

structure in the HIV leader (12) is probably involved in dimerization and thatU5 andPBS maynotberequired.

The MuLV dimerization site has been reported to be presentwithin the

*

sequences,where a double stem-loop was predicted (37). However, the folding of the recent prediction deviates from the long-standing model (8) and

seemstobecomparabletothe BLV structurewithregard

tobothlocation and overall structure.

BLV RNA dimerization and primer

tRNAPr'

binding. In retrovirusparticles,tRNAprimerbindingiscompatiblewith

genome RNA dimerization (8). A primer-associated RNA dimer model, as well as a primer-free dimer model, was

proposedfor avian andmouseretroviruses(8,13). However, it is inferred fromourcDNAprobingand deletion analyses

that BLV RNA dimerization is competitive with tRNAPr, binding; bindingof cDNA 1.1 and deletion of PBSinhibited the dimerization. Also, it should be noted that the 3'-terminal primer tRNAsequences are locked inthe charac-teristic cloverleafstructurebyformation of the acceptorand TgCstems. Theprocessof tRNAannealingmaybea more

sophisticated dynamicreaction than hasbeeninterpreted by

earlier experiments (35). Unfolding of the genome RNA dimer structure, inadditiontounwinding of the tRNAstems (11), might beinvolved.

RNA structures required for MA(pl5) binding. Deletion and chimericmutationanalyses implicatethe region encom-passing PBStothe5' end ofthegag sequences, PBS-a-,B,as sequences essential for thespecificinteraction withMA(pl5) tomake acomplex that isdetected as adiscrete band in the gel retardation assay. The 1 region was predicted to form twoisolatedstem-loops VI and VII, while the PBS-a region likely forms stem-loop V, involved in the dimer conforma-tion asdescribed above.

It is possible that PBS-a and ,B cooperatively make a

three-dimensional conformationactive in protein binding. It is alsopossible that the PBS-a region in concert with thegag sequences exhibits the dimer-forming activity and thereby increases the affinity of the a structure to the protein.

BLV U5-5' gag-MA(p15) complex formation and genome packaging. Although there has been no direct information about BLV genome RNApackaging, our in vitro character-ization of U5-5'gag-MA(p15) complex formation is related to the accumulated experimental results concerning the genome encapsidation in MuLV and other retrovirus sys-temsfromtheviewpoints described below.

PBS-ox-1 covers the BLV RNA region analogous to the packaging sites determined for MuLV (28, 41), Rous sar-coma virus (33), avian sarcoma-leukosis virus (18, 43), spleen necrosis virus(45), and HIV (3, 23). Recent observa-tion that gag sequences are also included in the signal sufficient forheterologous RNA packaging into MuLV and Rous sarcoma virusvirions (1, 5-7) supports our results; the BLV 13 region contained the gag start codon and its 3' sequences. In addition, the downstream gag sequenceswere found to performfacilitating roles in the MA(pl5) binding reaction.

Further support is provided by the fact that deletions in the U5region of the MuLV genome resulted inblockingof the RNApackaging into the virus particles(32).Inthisstudy with BLV RNA, U5 was shown to contain a part of the region essential for RNA dimerization. Deletion of U5 resulted in a more than a 10-fold decrease in complex-forming ability (compare LG with U5-5' gag in Fig. 4B), although the RNA feature to be annealed upon MA(pl5) bindingwasretained. These notionsagreewith the explana-tion that the U5 region of MuLV may not be directly involved in the selective packaging of viral RNA but is requiredtomaintainthecorrectoverall structureof the viral genome (24).

BLV RNA dimerization appeared to lead to enhanced reactivitywithMA(pl5)protein.Conversely,MA(pl5) bind-ing often resulted inannealingofthe RNAmonomers.These observations could offer anexplanation forwhy retrovirus particles contain a70S dimer asthe genome.

With all of these in vitro results, it is necessary to determine at the cell culture level the RNA and protein sequencesrequired for BLV

packaging,

using

aDNA clone capable of infectious particle production or an equivalent assay system.

There is disagreement concerning the function of avian retrovirus MA(pl9); the specific

RNA-binding

activity

de-scribed in earlier studies(20, 21)wasundetectable inrecent

experiments (42).

Furthermore,

detection of the specific interaction between MuLVppl2and the

homologous

RNA (39, 40) has never been repeated with other methods or further

investigated

to determine the

responsive

RNA se-quences. Since retrovirus gag precursor

processing

occurs

VOL.67, 1993

on November 9, 2019 by guest

http://jvi.asm.org/

[image:8.612.66.299.63.363.2]

after formation of the immature core (19, 47), it is thegag polyprotein that first associates with the viral RNA. Unlike avian virus MA(p19) or MuLV ppl2, BLV MA(pl5) is the NH2-terminal cleavage intermediate generated by the Pr44gag processing (46) or the precursor of mature proteins MA(plO) and p4 (17).

The5'-terminal RNAstructureand virus replication. In the predicted folding of the 1 region (Fig. 6B), the gag start codon isfound in stemVI. Although the model needstobe validated experimentally, it is conceivable that the 5'-termi-nal conformation of the viral RNA and its interaction with cellular and viral proteins may determine the fate of viral RNAmolecules eithertoactasmRNA or to beencapsidated intoparticles asthe genome RNA. In HIV, the large folded structure lies immediately 5' to the AUG codon (12).

Manyfunctional elements for reverse transcription prim-ing, RNAdimerization, packaging, translation initiation and activation, and splicing nest in the 5'-terminal region of retroviral RNA. Some of these elements overlap each other orarein closeproximity, suggesting the possibility that one RNAfunction can regulate other activities. The 5'-terminal region may form an intricate three-dimensional conforma-tion. In this regard, retroviral RNA may be reminiscent of poliovirus RNA,whose 5'-terminal noncoding region forms a defined cloverleaf structure that controls translation (34, 44) and genome replication (4) by interacting with various cellular and viral factors.

ACKNOWLEDGMENT Wethank YokoSakamoto forBLVpreparation.

REFERENCES

1. Adam, M.A., and A. D. Miller. 1988.Identification ofasignal in amurine retrovirus that is sufficient forpackaging of nonretro-viral RNA into virions. J. Virol. 62:3802-3806.

2. Aiyar, A., D. Cobrinik, S. Ge, H.-J. Kung, and J. Leis. 1992. InteractionbetweenretroviralU5 RNA and the

T*C

loopofthe tRNATrP primer is required for efficient initiation ofreverse transcription. J.Virol. 66:2464-2472.

3. Aldovini, A., and R. A. Young. 1990. Mutations of RNA and protein sequences involved in human immunodeficiencyvirus type 1packaging resultinproduction of noninfectious virus. J. Virol. 64:1920-1926.

4. Andino, R., E.Rieckhof, and D.Baltimore. 1990. A functional ribonucleoprotein complex forms around the 5' end of poliovi-rus RNA. Cell 63:369-380.

5. Armentano, D., S.-F. Yu, P. W. Kantoff, T. von Ruden, W. F. Anderson, and E. Gioboa. 1987. Effect of internal viral se-quences in the utility ofretroviral vectors. J. Virol. 61:1639-1646.

6. Aronoff, R., and M.Linial.1991. Specificity of retroviral RNA packaging.J.Virol. 65:71-80.

7. Bender, M. A., T. D. Palmer, R. E. Glinas, and A. D. Miller. 1987. Evidence that the packaging signal of Moloney murine leukemia virus extends into the gag region. J. Virol. 61:1639-1646.

8. Coffin, J. 1982. Structure of the retroviral genome, p. 261-368. In R.Weiss, N. Teich, H. Varmus, and J. Coffin (ed.), RNA tumor viruses. Cold Spring Harbor Laboratory, Cold Spring Harbor,N.Y.

9. Gorelick, R. J., L. E. Henderson, J. P. Hanser, and A. Rein. 1988. Pointmutants of Moloney murineleukemia virus that fail

topackageviral RNA: evidence for specific RNA recognition by

a "zinc finger-like" protein sequence. Proc. Natl. Acad. Sci. USA 85:8420-8424.

10. Gorelick, R. J., S. M. Nigida, Jr., J. W. Bess, Jr., L. 0. Arthur, L. E. Henderson, and A. Rein. 1990. Noninfectious human immunodeficiency virus type 1 mutants deficient in genomic RNA.J.Virol. 64:3207-3211.

11. Harada, F., G. G. Peters, and J. E. Dahlberg. 1979. The primer tRNA for Moloney murine leukemia virus DNA synthesis. J. Biol. Chem. 254:10979-10985.

12. Harrison, G. P., and A. M. L.Lever. 1992. The human immu-nodeficiency virus type 1 packaging signal and major splice donor region have a conserved stable secondary structure. J. Virol. 66:4144-4153.

13. Haseltine, W. A., A. M. Maxam, and W. Gilbert. 1977. Rous sarcoma virus genome is terminally redundant: the 5' sequence. Proc. Natl. Acad. Sci. USA 74:989-993.

14. Heaphy, S., C. Dingwall,I. Ernberg, M. J. Gait, S. M. Green, J. Karn, A. D. Lowe, M. Singh, and M. A. Skinner. 1990. HIV-1 regulator of virion expression (Rev) protein binds to an RNA stem-loop structure located within the Rev response element region. Cell 60:685-693.

15. Jentoft, J. E., L. M. Smith, X. Fu, M. Johnson, and J. Leis. 1988. Conserved cysteine and histidine residues of the avian myelo-blastosis virus nucleocapsid protein are essential for viral rep-lication but are not "zinc-binding fingers." Proc. Natl. Acad. Sci. USA 85:7094-7098.

16. Karpel, R. L., L. E. Henderson, and S. Oroszlan. 1987. Interac-tions of retroviral structural proteins with single-stranded nu-cleicacids. J.Biol. Chem. 262:4961-4967.

17. Katoh, I., H. Kyushiki, Y. Sakamoto, Y. Ikawa, and Y. Yoshi-naka. 1991. Bovine leukemia virus matrix-associated protein MA(pl5): further processing and formation of a specific com-plexwith thedimer of the 5' terminal genomic RNA fragment. J. Virol. 65:6845-6855.

18. Katz, R. A., R. W.Terry,and A. M. Skalka. 1986. A conserved cis-actingsequencein the5' leader of avian sarcoma virus RNA isrequired forpackaging.J. Virol. 59:163-167.

19. Kohl, N. E., E. A. Emini, W. A. Schleif, L. J. Davis, J. C. Heimbach, R. A. F. Dixon, E. M.Scolnick, andI. S. Sigal. 1988. Active human immunodeficiencyvirus protease is required for viralinfectivity. Proc.Natl. Acad. Sci. USA 85:4686-4690. 20. Leis, J., S. Johnson, L. S. Collins, and J. A. Traugh. 1984.

Effects of phosphorylation of avian retrovirus nucleocapsid proteinppi2onbinding ofviral RNA. J. Biol. Chem. 59:7726-7732.

21. Leis, J., J. McGinnis, and R. W. Green. 1978. Rous sarcoma virus p19 binds to specific double-stranded regions of viral RNA: effect of p19 oncleavage ofviral RNA by RNase III. Virology 84:87-98.

22. Lenardo, M.J.,C.-M. Fan, T. Maniatis, and D. Baltimore. 1989. The involvement of NF-KB in 3-interferon gene regulation reveals itsrole as widely induciblemediatorofsignal transduc-tion. Cell57:287-294.

23. Lever, A., H. Gottlinger, W. Haseltine, and J. Sodroski. 1989. Identification of a sequence required forefficientpackaging of human immunodeficiency virus type 1 RNA into virions. J. Virol.63:4085-4087.

24. Linial, M. L., and A. D. Miller. 1990. Retroviral RNA packag-ing: sequencerequirements and implications.Curr.Top. Micro-biol.Immunol. 157:125-152.

25. Lowry,0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem.193:265-275.

26. Luban, J.,and S. P. Goff. 1991. Binding of human immunode-ficiency virus type 1(HIV-1) RNA to recombinant HIV-1 gag protein. J. Virol. 65:3203-3212.

27. Malim, M. H., L. S. Tiley, D. F. McCarn, J. R. Rusche, J. Hauber, and B. R. Cullen. 1990. HIV-1 structural gene expres-sion requires binding of the Rev trans-activator to its RNA targetsequence. Cell60:675-683.

28. Mann, R., R. C.Mulligan, and D. Baltimore.1983.Construction of aretrovirus packaging mutant and its use toproduce helper-freedefective retrovirus. Cell33:153-159.

29. Meric, C.,and S. P. Goff. 1989. Characterization of Moloney murineleukemia virus mutants withsingle-amino-acid substitu-tions in the Cys-His box of thenucleocapsid protein. J.Virol. 63:1558-1568.

30. Meric, C., E. Gouilloud, and P.-F. Spahr. 1988. Mutations in Roussarcoma virus nucleocapsid proteinp12(NC): deletions of

on November 9, 2019 by guest

http://jvi.asm.org/

DIMERIZATION AND gagPROTEIN BINDING IN BLV 1839 Cys-His boxes.J.Virol. 62:3328-3333.

31. Meric, C., and P.-F. Spahr. 1986. Rous sarcomavirusnucleic acid-bindingprotein p12 isnecessaryfor viral70SRNAdimer formationandpackaging. J.Virol.60:450-459.

32. Murphy, J. E., and S. P. Goff. 1989. Construction and analysis of deletion mutations in the U5 region of Moloney murine leukemia virus: effects on RNA packaging and reverse tran-scription.J. Virol. 63:319-327.

33. Nishizawa, M., T. Koyama, and S. Kawai.1985.Unusual feature oftheleadersequenceof Rous sarcomaviruspackagingmutant TK15. J.Virol. 55:881-885.

34. Pelletier, J., and N. Sonenberg. 1988. Internal initiation of translation ofeukaryoticmRNAdirected byasequencederived frompoliovirusRNA. Nature(London) 334:320-325.

35. Peters, G., and J. E. Dahlberg. 1979. RNA-directed DNA synthesis in Moloney murine leukemia virus: interaction be-tweentheprimertRNAand thegenome RNA. J.Virol. 31:398-407.

36. Phalguni,G., and J. F. Ferrer. 1980. Detection ofa precursor-likeprotein ofbovine leukemia virusstructural polypeptidesin purified virions. J.Gen. Virol. 47:311-322.

37. Prats, A. C.,C. Roy, P. Wang, M. Erard, V. Housset, C. Gabus, C.Paoletti, and J.-L.Darlix. 1990.ciselements and trans-acting factors involved in dimer formation of murine leukemia virus RNA.J. Virol. 64:774-783.

38. Sagata, N.,T.Yasunaga, J.Tsuzuku-Kawamura,K. Ohishi, Y. Ogawa, and Y. Ikawa. 1985. The nucleotide sequenceof the genomeof bovineleukemia virus: its evolutionary relationship tootherretroviruses. Proc.Natl. Acad. Sci. USA 82:677-681. 39. Sen, A.,andG.J. Todaro. 1976.Specificityofin vitrobindingof

primatetypeCviral RNA andthe homologous viralp12core

protein.Science 193:326-328.

40. Sen, A., and G. J. Todaro. 1977. The genome-associated, specificRNAbinding proteins of avian and mammaliantypeC viruses. Cell 10:91-99.

41. Sorge, J. D., D. Wright, V. D. Erdman, and A. E.Cutting. 1984. Amphotropic retrovirus vector system for human cell gene transfer. Mol. Cell. Biol.4:1730-1737.

42. Steeg, C. M., and V. M. Vogt. 1990.RNA-binding properties of the matrix protein (p199a9) of avian sarcoma and leukemia viruses. J.Virol. 64:874-855.

43. Stoker, A. W., and M. J. Bissell. 1979. Development ofavian sarcomaandleukosis virus-basedvector-packaging cell lines.J. Virol.62:1008-1015.

44. Trono, D., R. Andino, and D. Baltimore. 1988. An RNA se-quence of hundreds ofnucleotides at the 5' endofpoliovirus RNAis involved in allowingviral protein synthesis. J. Virol. 62:2291-2299.

45. Watanabe,S., and H. M. Temin. 1983.Encapsidationsequences forspleen necrosis virus,anavianretrovirus,arebetweenthe 5' longterminalrepeatand the startof the gag gene. Proc. Natl. Acad. Sci.USA79:5986-5990.

46. Yoshinaka, Y., I. Katoh, T. D. Copeland, G. W. Smythers, and S. Oroszlan. 1986. Bovine leukemia virusprotease:purification, chemical analysis, and in vitro processing of gag precursor polyproteins. J.Virol. 57:826-832.

47. Yoshinaka, Y., and R. B.Luftig. 1977. Murine leukemia virus morphogenesis: cleavage of P70invitrocanbeaccompanied by ashift fromaconcentrically coiled internal strand("immature")

to acollapsed ("mature") form of the viruscore. Proc. Natl. Acad. Sci.USA74:3446-3450.

VOL. 67,1993

on November 9, 2019 by guest

http://jvi.asm.org/