Orientation-specific cis complementation by bulge- and loop-mutated human immunodeficiency virus type 1 TAR RNAs.

Copyright © 1994, American Society for Microbiology

Orientation-Specific cis Complementation by Bulge-

and

Loop-Mutated

Human

Immunodeficiency Virus

Type 1 TAR RNAs

MARTIN BRADDOCK, ROBERT POWELL, JULIASUTTON, ALAN J. KINGSMAN, AND SUSAN M. KINGSMAN*

RetrovirusMolecularBiology Group, Department of Biochemistry, University of Oxord,

Oxford

OX1

3QU,

United

Kingdom

Received 31 May1994/Accepted 19August1994

Tat activates human immunodeficiency type 1gene expression by bindingto TAR RNA.TARcomprises a partiallybasepairedstemandhexanucleotide loop withatripyrimidinebulge in theupperstem. Invitro, Tat binds tothebulge and upper stem, with no requirementfor the loop. However, in vivo, loop sequences are critical foractivation,implyingthat aloop bindingcellular factormaybeinvolved intheactivationpathway. Giventhatactivation appearstobe atwo-componentsystemcomprisingaTat-bulgeinteraction and a cellular factor-loop interaction,weconsidered thatitmight bepossible tospatiallyseparate the twocomponents and retainactivation. Wehaveconstructed aseries of doubleTAR elements comprisingvarious combinations of mutated TAR structures. Defective TARs with nucleotide substitutions in either the bulge or the loop complemented eachother togivewild-typeactivation.However, thecomplementationwasorientation specific, requiring the intact Tat binding sitetoresideonthe5'-proximalTAR. These data suggestthatprovidedthe wild-typeorientation ofthebulge andloopelements isretained,there is norequirementfor them tocoexiston the same TAR structure.

Humanimmunodeficiency virus type 1 (HIV-1) replication

is critically dependent upon the virally encoded Tat protein.

Tat binds to the transactivation response (TAR) element, which is localized between +14 and +44 atthe 5' end of all HIV-1RNAs(7).TARisapartially base paired RNAstructure

comprisingatripyrimidinebulge andanunpaired hexanucleotide

loop (10,23).Theposition of TAR relativetothetranscription

startsite is critical foractivation, andTARisfunctional only when placed in the correct orientation with respect to the HIV-1long terminal repeat (15, 17, 24, 27). The predominant mode of action of Tat istostimulate transcription, and Tat is envisaged asbeing introducedto the transcription machinery inafeedback process as aprotein-bound TAR RNA complex

(8).

Invitro, Tat binds specifically at the bulge, with no require-ment for the loop sequence. However, in contrast to the in vitro binding data,mutations in the loop sequence abolish both Tatactivation of transcription and translation (4, 5, 11). This

finding implies arequirement for cellular factors to facilitate

thebinding and/or the activation process. A number of TAR RNA-binding proteins have been identified (12, 14, 19, 22, 25, 29, 37), some of which bind to the loop sequence, but the function of these cellular factors that bind the loop is contro-versial. The finding that Tat activates transcription

indepen-dently of TAR when tethered to RNA via a heterologous

RNA-protein interaction (28, 31) or when located upstream of theRNAstart in a TAR-less configuration (30) suggests that

loop-bindingfactors cannot be central to the activation

path-way. This isconsistent with the fact that Tat directly contacts components of the transcription initiation complex (18, 20). Thereis, however, some genetic evidence that implies that in the normal TAR configuration, Tat can bind to TAR only when complexed with a loop-binding protein (21). It has been

*Corresponding author. Mailing address: Retrovirus Molecular

Biology Group, Department of Biochemistry, South Parks Road, Oxford OX1 3QU, United Kingdom. Phone: 865 27548. Fax: 865 275259.

suggested that the loop-binding protein also mediates the

activation. Analternative view is that theloop-bindingfactors areessentialtofacilitateaccessof Tattothebulge byexcluding

competing cellularbulge-bindingfactors (4).

To investigate further the relationship between the Tat-bulge interaction and the cellular factor-loop interaction, we

separated the twobinding sitesby constructing tandem TAR

elements downstream of the HIV-1 promoter. The TAR elementswere mutated such that they lacked either the Tat

bindingsiteatthebulgeorthecellular factorloopbindingsite.

The bulge mutation was a substitution of residue U-23 for

C-23, and the loop mutationwas athree-base substitutionof the sequence30/CUGGG/34for30/AGGGU/34. Both of these classesof mutationseriously impairthetat activation process

(4, 26). Thegeneralconfigurations of single and double TAR

mutants areshown inFig. 1. Single TAR mutants were derived from the wild-type (WT) long terminal repeat-CAT plasmid

pOGS210 (1) and retained TAR in the precise WT position

but with the base changes indicated (Fig. 1A). Plasmids

containing tandem TARelements were constructed by

insert-ingasynthetic oligonucleotide, comprising alinker containing anSstII site andresidues +1 to + 62 of TAR as aWT,bulge mutant (BM), orloop mutant(LM) sequence into the unique

NheIsite of plasmid pOGS210, pPE511, or pPE682. The TAR elements studied were either WT, BM, or LM, and the double TARelements were combinations of these as indicated in Fig.

iC.

To ensure that theanalysis was quantitative, we established

limiting conditions for Tat activation. The activation of BM

and LM TARs in pPE682 and pPE511, respectively, was compared with that of WT (pOGS210) at a range of Tat concentrationsprovided by expression from plasmid pOGS213

(1). Plasmids were transfected into subconfluent HeLa cells,

andproteinextractswereassayed for chloramphenicol

acetyl-transferase (CAT) activity after 48 h. At a low input of Tat (1

ng of pOGS213), neither the BM nor LM TAR supported

activation, whereasWTgave about a 60-fold activation, (Fig.

2). Consistent with other studies (3, 26), the mutations were

8396

on November 9, 2019 by guest

http://jvi.asm.org/

Nh

HIV-1 LT AT I

+20 TAR

pOGS210

pPE682

pPE511

+20 bulge loop +38

AGAUCUGAGCCUGGGAGCU wildtype(WT)

AGACCUGAGCCUGGGAGCU bulgemutant(T1) AGAUCUGAGCAGGGJAGCU bopmutant(T2)

WT BM

LM c

CD

B

+1 +80

Nti SS Nh pA

TAR1 TAR2

Key: x

-X- wild type (WT)

_ -0- bulgemutant(T1)

-0_- loopmutant(T2)

0 1 1 0 100 1000

TAR1 TAR2

WT WT DT

WT LM T3

BM LM T4

BM WT T5

LM LM T6

LM WT T7

EM BM T8

LM BM T9

LM-10o-EM T10

BM-101-LM Ti1

FIG. 1. Structures of the expression plasmids used for WT (pOGS210), BM, and LM sequences. LTR, long terminal repeat; CAT, bacterialCATcodingregion; I, simianvirus 40 small t intron; Nh,NheI restriction site;pA,polyadenylation site. The TAR stem loop isrepresented byarrows.TheLMis plasmidpPE511(5), and the BM, which containsapointmutation of U23 to C23, was generated by PCR mutagenesis using standard techniques. The primerused for primer extension analysis of RNA was as described previously (3) and corresponded to nucleotides +33 to +13 in the CAT gene. (B) Structures of the HIV-1 LTR constructs harboring double TAR elements. Ss, SstII restrictionsite. PlasmidspOGS210, pPE682(T1) andplasmid pPE511(T2)werelinearizedattheuniqueNhel site, and

a synthetic oligonucleotide which comprised a linker containing an SstII restrictionsite and either theWT, LM,orBM TAR sequence (panelA)wasinserted. TAR1 is eitherWT,BM,orLM,and TAR2is

asyntheticoligonucleotide consisting ofalinkerandresidues +1 to +

62ofeither theWT,BM, or LM TAR sequence.(C)Configurations anddesignations of themutants.MutantsT10 andTllareT9andT4 in which the TARelementsareseparated bya101-nucleotidespacer.

lessdeleteriousat ahigh(500-ng) input ofTat.Allsubsequent

experimentswereconducted with 500 ng of TAR DNA and 1

ngofTat-expressing plasmid.

Although RNA polymerase is poorly processive in the

absence ofcorrect Tatactivation (28), itseems unlikely that there is an accumulation oftranscripts truncated afterTAR,

as,forexample,aribozymeplaceddownstreamofTARis still

functional(18).Wethereforeanticipatedthat the second TAR element should besynthesized. However, to confirm this,we determined the level of activation fromasingleTARelement

compared with tandem TAR elements in which the first

[image:2.612.320.561.73.248.2] [image:2.612.65.303.74.386.2]

elementwaseither aBM

(T5)

or anLM

(T7).

In

addition,

to control for the possibility that a

super-secondary

structure, deleterioustoWTactivation,mightarise from the

duplication

[pOGS213]ng

FIG. 2. Effectsof titrating Tat-expressing (pOGS213)onthe

acti-vation of WT andmutantTARsequences.Transient transfectionwere

carried outby calcium phosphate coprecipitation onto subconfluent HeLa cells. Extractswereassayed after 48h for CAT activity by using

Quan-T-CAT, ascintillation counting-based assay(Amersham

Inter-national plc). RNAwas prepared as described previously (6) and

analyzed by primer extension. Transfection efficiencywasdetermined

by cotransfection of human cytomegalovirus promoter-driven lucif-erase plasmid (pMW41 [35]), and luciferase assays carried out as

describedpreviously (36). Plasmids encoding WT bulge BM Ti,and LM T2TAR RNAs (500ngof each)weretransfected into HeLa cells

with 0to500ngof pOGS213 and 250ngofpMW41. Proteinextracts

containing thesamenumberof luciferase unitswereassayed for CAT

activity.

of TAR elements,we analyzed the activation of tandem WT

TARs(DT)andatandem element in which the first TARwas

WT and thesecond TARwasmutated (T3). The double TAR

plasmids were cotransfected with pOGS213 under limiting

conditions, and after 48 h, extracts were analyzed for the

presence of CAT transcripts and CAT enzyme activity (Fig.

3A).

Interestingly, the tandem WT TAR gave twice the level of activation ofthe single TAR (Fig. 3A, lane 4compared with lane 2). AdefectiveTARplaced downstreamofWT had no

deleterious effect, as activation levels were the same as

ob-tained witha singleTAR (compare lanes 2 and6). Placinga

defectiveTAR upstream ofa WTTAR element also had no

effecton activation,whether itwas a BM (lane 8) or an LM (lane 10).Inthese latterconfigurations, activation levelswere

thesame asobtained with thesingleTAR(lane 2).These data indicated thataWT TARelementwasfullyfunctional whether placed upstream or downstream of a second defective TAR element, indicatingthatTARfoldingandproteininteractions

wereunperturbed bythe tandemstructure.Thisfindingis also consistent with reports that multiple tandem TAR elements functioneffectivelyasTARdecoys(32, 33).Thereasonfor the increased activation by the tandem WT TARs is not clear, although similar observations were made when HIV-2 TAR elementswereduplicated (2).ThefindingthataWTTARcan

function when placed downstream of a defective element about 90 nucleotidesawayfrom the mRNA startappears to be

inconsistent with thepreviouslydescribedposition dependence of TAR(15, 17, 24, 27).The result couldbe explained if the

compactsecondarystructureof thefirst TARservesto shorten

the distance between the initiation site and the active TAR.

Alternatively,an asyet uncharacterizedTAR-specific protein

A

on November 9, 2019 by guest

http://jvi.asm.org/

A.

32-28,119

24-0

0

B. 32

24

0

C;

-r

, 16 CO

10,916

C.

0

CO

x

0

am

I

mayrecognizethedefective TAR andfacilitate the activity of thesecondelement despite its distancefromthe initiationsite. Given that tandem HIV-1 TAR elements were functional, we then examined whether different mutated TAR elements could be combinedin tandem to allowactivation by comple-mentationbetween the Tatbinding site and the loop factor bindingsite (Fig. 3B). The combination T4 (BM-LM), which effectively presentsaWTloop in the firstTARfollowed bya

WT bulge in the second TAR, showed no activation by Tat

(lane 6). Similarly,doublebulgemutants (T8)ordouble loop

mutants(T6)weredefective (lanes12and10), paralleling the

LANE: 1

12

3 4 5 6 7

1

8 9 10 11 12 13

CONSTRUCT: WT DT |T4 TT9 T6 T8 IN

TAT:

l

+

|

+

T

+

r

+

-

+

ACTIVATION: 19.4 50.8 1.8 20.9 1.9 1.7

212W "a

RNA:

150_ _

FIG. 3. Activation ofdouble TAR elements. Transient transfec-tions were carried out by calcium phosphate coprecipitation onto subconfluent HeLacells(13). Extractswereassayed after 48 h forCAT activityasdescribedabove.RNA waspreparedasdescribedpreviously (6) and analyzed by primer extension. Transfection efficiency was determinedasdescribedabove.

results with single TAR elements (Fig. 2). However, the combination T9 (LM-BM), which presents an intact bulge

followed by an intactloop, gave activationtothesamelevel as asingle WT TAR element (lane 8).These data show that the Tatbinding sitecan function witha loopon aseparate TAR element but only inoneorientation.

These observations then led us to consider whether this orientation-specific complementationwas dependentonclose proximity of the twoTARsalongthelengthof thetranscript.

Toaddressthis, plasmidconstructions T4 and T9were modi-fied to produce plasmids Ti1 and T10, respectively, by the insertion of a 101-nucleotide oligonucleotide (5'-GGTAA GACACGACTTATCGCCACTGGCAGCAGCCACTGGTA

ACAGGATTAGCAGAGCGAGGT'lTGTAGGCGGTGCTA

CAGAGTTCTAGAAGTGGTGGCCGC-3')

at the unique SstII site, thus spacing the two TAR elements. Plasmid con-structions Tll and T10were tested in a Tat activation assay

(Fig. 3C). As formutant T4, Tat was unable to activate the

TARconfiguration inmutantTll(Fig.3C, lane8).Activation ofmutantT10,however,wasmuch reducedcomparedwiththe parental construct T9 (compare lanes 2 and 6). Taken

to-gether,these datashow that while there issurprisingflexibility

intherelationshipof the Tatbinding site to the TAR loop, that

flexibilitydoesnotextendto aseparation of 101 nucleotides.

The resultsaredepictedschematically in Fig.4.Wepropose

thatTatandtheloop-bindingproteinareasymmetric and that

theypresent interaction surfaces in the WTconfiguration. In

T4,althoughTatand theloopfunctionaredisplayingthesame

facetothe promoter, theyare notdisplayingthesamefaceto each other. Thisconfiguration isinactive. IfTatand theloop factor bind in thisconfiguration,thenclearlytheirindependent

presentation to the initiation complex is not sufficient for

36

on November 9, 2019 by guest

http://jvi.asm.org/

[image:3.612.66.551.80.507.2]

CONSTRJCT

IER I

rpp LJ0iR

WT

DT

T4

T9

T10

FRMTP-- - L- ~o-

~-%ACTIVATION

100

203

0

100

12.0

FIG. 4. Activationconfigurationsof doubleTAR elements. Shown

isasimplecartoonrepresentingTARsequence,indicatingthebinding

oftwoasymmetric factors,Tat andaloop-bindingfactor. Thepercent

activation of thesequence configurations by Tatas determined from

data inFig. 3 isindicated.

activation. Alternatively,neither of themmaybind. InT9,Tat

and the loop-binding protein display the same face to each

otherand to the initiationcomplexas in theWTsingleTAR, and thisisnowactive. InT10,althoughTat andaloop-binding

cellular factor are in the correct orientation with respect to

each other, we would suggest that the spacing of the two

protein binding sites does not permit efficient interaction to

allowactivation to occur.

If, as proposed previously, Tat must form a stable

het-erodimer beforebindingto TAR(21),then thiscomplex must

beabletobridgetwo TARelements. Given the in vitro binding

propertiesof Tat (9, 10, 34),itisperhaps more likely that Tat

and the loop-binding factor bind independently but upon contact,facilitatedbytheflexibilityof the RNA, theremightbe aconformationalchangeinTat tocreate anactivation surface. Thenotion of conformationalchangein Tat is consistent with

thefindingthat isolated activation domainscomprisingamino

acids 1 to 48 bind different proteins from full-length Tat in

vitro

(16).

Itis interestingthat separating thebulge and loop

ondifferent TAR elements isfunctional,whereasspacingthem

by increasingthe upper stemresults in anonfunctional TAR

(2, 29).

Either there is insufficient flexibility in this latter structuretoallowproteininteractionorthestretchedstructure binds a competing factor such as TRBP1 (12), which may inhibit Tat binding. Irrespective of whether Tat binds TAR

independentlyor as aheterodimer withaloop factor,it is clear

that the bulge region in the first TAR is available for Tat

binding.

Thisis notconsistent with the notionproposedfora

single

TAR

(4)

that in the absence of

loop-binding factors,

competing

cellular bulge-binding factors block Tat

binding.

However, it is possible that the loop-binding factors on the

second TARalso

effectively

exclude other

bulge-binding

pro-teins.

Taken

together,

these resultsmeanthatanydetailed model of the Tat-TAR

phenomenon

would needtotakeaccountofa greater

flexibility

inthe

relationship

of the

bulge

and the

loop

than has been

appreciated

todate.

MartinBraddock isaRoyal SocietyResearchFellow;Robert Powell isaSERCpostgraduate.This researchwasfundedbyGLAXO andthe SERC.

We thank our colleagues in the Retrovirus Molecular Biology Groupforstimulating discussions.

REFERENCES

1. Adams,S.E.,

L.

D.Johnson,M.Braddock,A.J.Kingsman,and R. Edwards. 1988. Synthesis of a gene for the HIV transactivator proteinTatbyanovelsinglestrandedapproachinvolvingin vivo gaprepair.Nucleic AcidsRes.16:4287-4298.

2. Berkhout, B.,A.Gatignol, J.Silver,andD.Jeang.1990.Efficient transactivation bythe HIV-2 Tat protein requires a

duplicated

TAR RNAstructure.Nucleic AcidsRes. 18:1839-1846. 3. Blanchard,A.D., R.

Pow4l1,

M.Braddock,A.J.Kingsman,and

S. M.Kingsman.1992. An adenosineatposition27in the human immunodeficiencyvirustype 1trans-activationresponse element is

notcriticalfortranscriptionalortranslationalactivationbyTat.J. Virol.66:6769-6772.

4. Braddock, M.,R Powell,A. D.Blanchard,A.J. Kingsman,and S. M.Kingsman.1993. HIV-1 TARRNA-binding

proteins

control Tat activation of translation in Xenopus oocytes. FASEB J. 7:214-222.

5. Braddock, M., A. Thorburn, A. Chambers, G. D. Elliott, G.J. Anderson,A.J.Kingsman,and S. M.Kingsman. 1990. Anuclear translational blockimposed bytheHIV-1 3

region

is relieved

by

theTat/TARinteraction.Cell 62:1123-1133.

6. Braddock, M., A. M. Thorburn, A. J.

Kingsman,

and S. M. Kingsman. 1991. BlockingofTat-dependentHIV-1 RNA modi-ficationbyaninhibitorof RNA

polymerase

II

processivity.

Nature

(London)350:439-441.

7. Cullen,B. R. 1992. Mechanism of action of

complex regulatory

proteinsencodedbycomplexretroviuses. Microbiol.Rev. 56:375-394.

8. Cullen, B. R. 1993. Does HIV-1 tat induce a

change

in viral initiationrights?Cell73:417-420.

9. Dingwall,C.,I.Ernberg,M.J. Gait,S.M.Green, S.Heaphy,J. Karn,A. D.Lowe,M. Singh,and M. A.Skinner. 1990. HIV-1tat

,-1% I I

1% ol

-1I

on November 9, 2019 by guest

http://jvi.asm.org/

[image:4.612.69.304.81.588.2]

protein stimulatestranscriptionbybindingto aU-richbulgeinthe stemof theTARRNA structure. EMBOJ.9:4145-4153. 10. Dingwall, C., I. Ernberg, M. J.Gait, S. M.Green, S.Heaphy,J.

Karn, A. D. Lowe, N. Singh, and M. A. Skinner. 1989. Human immunodeficiency tat protein binds transactivation responsive region(TAR)RNA in vitro. Proc.Natl.Acad.Sci.USA 86:6925-6929.

11. Feng, S., and E. C. Holland. 1988. HIV-1 tat transactivation requires theloop sequences withinTAR. Nature (London) 334: 165-167.

12. Gatignol, A., A. Buckler-White, B. Berkhout and K.-T. Jeang. 1991. Characterisation of a human TAR-RNA binding protein that activates the HIV-1 LTR.Science 251:1597-1600.

13. Gorman, C. M., L. F.Moffat,and B. H.Howard.1982. Recombi-nant genomeswhichexpresschloramphenicol acetyltransferasein mammalian cells.Mol.Cell. Biol.2:1044-1051.

14. Gunnery, S., S. R. Greeen, and M. B. Mathews. 1992. Tat responsive region of humanimmunodeficiency virustype 1 stim-ulatesproteinsynthesis in vivoand in vitro: relationship between structure and function. Proc. Natl. Acad. Sci. USA 89:11556-11561.

15. Hauber, J., and B. R. Cullen. 1988. Mutational analysis ofthe transactivation responsiveregion of thehumanimmunodeficiency virusTatprotein. J.Virol.62:673-679.

16. Hermann, C. H., and A. Rice. 1993. Specific interaction ofthe humanimmunodeficiency virustatproteins withacellular protein kinase.Virology 197:601-608.

17. Jakobovits,A., D. H. Smith, E. B. Jakobovits, and D. J. Capon. 1988. A discrete element 3' of human immunodeficiency virus1 (HIV-1) and HIV-2 mRNA initiation sites mediates transcrip-tional activation by an HIV trans activator. Mol. Cell. Biol. 8:2555-2561.

18. Jeang, K.-T.,R.Chun, N. H. Lin, A.Gatignol, C. G. Glabe, and H. Fan. 1993. In vitro and invivo binding of human immunodefi-ciency virus type 1 Tat protein and Spl transcription factor. J. Virol.67:6224-6233.

19. Kaczmarski, W., and S. A. Khan. 1993. Lupus autoantigen ku proteinbindsHIV-1 TAR RNAinvitro. Biochem. Biophys. Res. Commun. 196:935-942.

20. Kashanchi, F., G. Piras, M. F. Radonovich, J. F. Duvall, A. Fattaey, C.-M.Chiang, R. G. Roeder, and J. N. Brady. 1994. Direct interaction of human TFIIDwith the HIV-1 transactivator Tat. Nature(London)367:295-296.

21. Madore, S. J., and B. R. Cullen. 1993. Genetic analysis of the cofactor requirement for human immunodeficiency virus type 1 Tatfunction.J. Virol. 67:3703-3711.

22. Marciniak, R. A., M. A. Garcia-Blanco, and P. A. Sharp. 1990. Identification and characterisationof aHeLa nuclear protein that specificallybinds to the transactivation response element of

hu-manimmunodeficiencyvirus. Proc. Natl. Acad. Sci. USA 87:3624-3628.

23. Muesing,M.A.,D. Smith,and D.J. Capon. 1987.Regulationof mRNAaccumulationbya humanimmunodeficiencyvirus trans-activatorprotein.Cell48:691-701.

24. Peterlin, B.N.,P.J. Luciw,P.J. Barr, and M. D.Walker. 1986. Elevated levels ofmRNA can account for thetransactivation of human immunodeficiency virus. Proc.Natl. Acad. Sci. USA 83: 9734-9738.

25. Rounsville, M. P., and A. Kumar. 1992. Binding of a host cell nuclear proteinto the stemregion ofhuman immunodeficiency virus type 1 transactivation-responsive RNA. J. Virol. 66:1688-1694.

26. Roy, S., N. T.Parkin, C.Rosen, J.Itovitch, and N. Sonenberg. 1990.Structuralrequirementsof transactivation of human immu-nodeficiency virus type 1 long terminal repeat-directed gene expressionbytat:importanceofbasepairing, loop sequences,and bulges in thetat-responsive sequence region. J.Virol. 64:1402-1406.

27. Selby, M. J., E. S. Bain, P. A. Luciw, and B. M.Peterlin. 1989. Structure, sequence, and position of the stem-loop in TAR determine transcriptional elongation by Tat through the HIV-1 longterminal repeat.Genes Dev.3:547-558.

28. Selby,M.J., and M. B. Peterlin. 1990.TransactivationbyHIV-1 tatviaaheterologousRNA-binding protein. Cell62:769-776. 29. Sheline, T., L. H.Milocco, and K. A.Jones. 1991. Two distinct

nucleartranscriptionfactorsrecognize loopandbulge residuesof theHIV-1 TAR RNAhairpin.Genes Dev. 5:2508-2520. 30. Southgate, C., L. L. Zapp, and M. R. Green. 1990. Activation of

transcription by HIV-1 Tat protein tethered to nascent RNA through another protein. Nature(London)345:640-642. 31. Southgate, C. M., and M. R. Green. 1991.Tat proteinactivates

transcriptionfrom an upstreamDNAbinding site: implicationsfor tatfunction. Genes Dev. 5:2496-2507.

32. Sullenger,B.A., H. F.Gallardo,G. E. Ungers, and E. Gilboa. 1990. Overexpression of TAR sequences renders cells resistant to humanimmunodeficiency virusreplication. Cell63:601-608. 33. Sullenger,B.A., H. F. Gallardo, G. E. Ungers, and E. Gilboa.1991.

Analysisoftransactingresponse decoyRNA-mediatedinhibition of humanimmunodeficiencyvirus type 1transactivation.J.Virol. 65:6811-6816.

34. Weeks, K. M., and D. M.Crothers. 1991. RNA recognitionby Tat-derivedpeptides: interaction in the majorgroove.Cell 66:577-588.

35. White,M., andS. M.Kingman. Unpublisheddata.

36. White, M. R. H., J. Morse, Z. A. M. Boniszewski, C. R. Mundy, M. A. W. Brady, and D. J. Chiswell. 1990. Imaging of firefly luciferase expressioninsinglemammalian cells using high sensi-tivitycharge-coupled devicecameras.Technique 2:194-201. 37. Wu, F. J., D. Garcia, D. Sigman, and R Gaynor. 1991. Tat

regulatesbinding ofthehumanimmunodeficiencyvirus transacti-vating region RNA loop-binding protein TRP-185. Genes -Dev. 5:2128-2140.

on November 9, 2019 by guest

http://jvi.asm.org/