AIDS RESEARCH AND HUMAN RETROVIRUSESVolume 11, Number 8, 1995Mary Ann Liebert, Inc.

Expression and Purification of the HIV Type 1 Rev ProteinProduced in Escherichia coli and Its Use in the Generation of

Monoclonal Antibodies

MICHAEL J. ORSINI,1'2 ARVIND N. THAKUR,3 WILLIAM W. ANDREWS,13-4MARIE-LOUISE HAMMARSKJÖLD,35 and DAVID REKOSHI-3-5

ABSTRACT

We have developed a simple and rapid procedure for the purification of large amounts of Rev protein over-

expressed in E. coli. The purification method, which does not require denaturation of the protein, takes ad-vantage of the positively charged nature of Rev and the ability of Rev to interact with nucleic acids. The pu-rified protein was used to develop three novel murine monoclonal antibodies against Rev. Using fusion proteinsbetween glutathione S-transferase (GST) and various fragments of the Rev protein, we mapped the specificityof these antibodies to different regions of the Rev protein. One antibody, 3H6, is directed against the maleo-lar localization/RRE-binding domain of Rev between amino acids 38 and 44. Another antibody, 3G4, recog-nizes an epitope between amino acids 90 and 116 of Rev. A third antibody, 2G2, does not recognize any ofthe fusion proteins, and may be directed against a conformational epitope. All three antibodies are able to de-tect Rev on Western blots and to immunoprecipitate Rev under native conditions. However, only 3H6 and3G4 immunoprecipitate Rev under denaturing conditions and are able to detect Rev expressed in transfectedcells by indirect immunofluorescence. These antibodies should prove useful in further studies of Rev function.

INTRODUCTION as a nuclear/nucleolar localization signal20-22 and appears tomediate the binding of Rev to the RRE-binding re-

THE HIV-1 Rev protein has been shown to promote the g¡on 11,18,19,22-25 ¡t nas aiso ^sen rep0rted that Rev protein can

transport of unspliced and incompletely spliced viral form multimers both in vitro19-23-25'21 and when expressed inmRNAs from the nucleus to the cytoplasm.1"7 The exact mech- cells.25 The ability of Rev to form multimers may either facil-anism by which this occurs remains unclear. However, Rev itate or enhance its ability to interact with the RRE.15-23-25-28function is dependent on the presence of a ci's-acting element Two other regions that flank the basic region appear to be as-

present in Rev-regulated mRNAs known as the Rev-responsive sociated with this property.19'25 Mutations made in either theelement (RRE).3,4'8"1 ' The Rev protein has been shown to bind basic domain or the multimerization domains of Rev result into RRE-containing RNA in vitro11'15 and to be associated with proteins that are nonfunctional.17_19'25 These mutant proteinsRRE-containing RNA in vivo.16 Using mutants of Rev, the abil- have no effect on wild-type Rev function when they are coex-

ity of Rev to function has been shown to correlate with the abil- pressed.18,19'25ity of Rev to bind to the RRE.1117"19 In contrast, mutations made within a leucine-rich region that

The 116-amino acid Rev protein contains several discrete re- lies between amino acids 75 to 83 of Rev result in nonfunc-gions defined by extensive mutagenesis that appear to be re- tional proteins that are capable of interfering with the activitysponsible for different properties of Rev. Rev contains an argi- 0f wild-type Rev.18-29 It has been suggested that these trans-nine-rich basic region from amino acids 38 to 50 that functions dominant mutant Rev proteins may inhibit Rev function by in-

'Department of Biochemistry, University at Buffalo, Buffalo, New York 14214.2Present address: Department of Molecular Genetics, SmithKline Beecham, King of Prussia, Pennsylvania 19406.3Department of Microbiology, University at Buffalo, Buffalo, New York 14214.4Present address: Chrion Corporation, 4560 Horton Street, Emeryville, California 94608.'Present address: Myles H. Thaler Center for AIDS and Human Retrovirus Research, and the Department of Microbiology, University of

Virginia, Charlottesville, Virginia 22908.

945

ORSINI ET AL.

terfering with the ability of Rev to interact with a cellular fac-tor required for Rev to function.18'30,31 The amino-terminal 10amino acids and the carboxy-terminal 33 amino acids appearto be dispensable for Rev activity.24,32

To generate tools for further studies of Rev and the proper-ties of the various domains of this protein, we overexpressedthe wild-type Rev protein in Escherichia coli and purified theprotein to apparent homogeneity. The purification method thatwe describe in this article yields large amounts of native, func-tional Rev protein without the use of dénaturants. We then usedthis protein to develop novel monoclonal antibodies againstRev. These antibodies were mapped to determine the region ofRev against which they were directed and were then charac-terized for their ability to recognize Rev by Western blot analy-sis, immunoprecipitation, and immunofluoresence.

MATERIALS AND METHODS

Prokaryotic expression plasmidsThe prokaryotic expression vector pOTSNcol was a kind gift

from M. Rosenberg at SmithKline Beecham (King of Prussia,PA).33-36 The rev gene was introduced into this vector as an

Ncol-Xhol fragment to create the plasmid pOTSrev. To do this,an Ncol site was introduced at the start of the coding region forrev using a polymerase chain reaction (PCR) method.37-38 Therev gene was amplified from pCVl39 using an upstream oligo(5' CATCTCCCATGGCAGGAAGAAGCG 3') that changedpCVl nucleotide 969 from T to C (indicated in bold), creatingan Ncol site. The downstream oligo (5' CTAGGTCTCGA-GATGCTGCTCCCA 3') 3' to the rev stop codon, which con-

tained an Xhol site (CTCGAG), was entirely complementary to

pCVl. The resulting PCR product was cleaved with Ncol andXhol, gel purified, and ligated to a 5.8-kb Ncol-Xhol fragmentfrom pOTSNcol.

pOTSrev(A38-44) was constructed by SOE-PCR40 in a man-

ner similar to that described for pRev(MlO) below. In this case

the internal oligonucleotides deleted the coding region foramino acids 38^14 of Rev (RRNRRRR) and the Ncol-Xholfragment containing the deletion was then cloned into pOTSrev.

Bacterial strains

N99cl+ is an E. coli A lysogen that was used for all cloninginto pOTS.33,36 This strain contains an integrated copy of theA genome, which synthesizes sufficient amounts of A repressorto inhibit transcription from the PL promoter. This represses thesynthesis of heterologous protein. It was maintained on Luria-Bertani broth (LB) and grown at 37°C.

N5151 is an E. coli strain that contains a temperature-sensi-tive mutation in the A ci repressor gene (cI857). It is propa-gated at 32°C except when performing inductions.33,36 Whenshifted to 42°C, the mutant form of the repressor fails to findthe Pl promoter, and synthesis of the heterologous protein be-gins.Mammalian expression constructs

pRev was constructed as previously described.41 It containsthe Rev protein coding sequence from pCVl39 under the con-

trol of the simian cytomegalovirus IE94 promoter enhancer re-

gion from -650 to +30.42To facilitate the exchange of mutated Rev fragments between

pOTSrev and pRev, pRev was modified by SOE-PCR to cre-

ate the plasmid pRev(N-X). pRev(N-X) contains an Ncol sitejust upstream from the ATG, which makes it identical to

pOTSRev in this region.pRev 11-116 was constructed using pCVl as a template for

a PCR reaction. For this reaction, the upstream oligonucleotidecontained an Ncol site near its 5 ' end followed by sequences thatencoded Rev amino acids 11-15. The downstream oligo was

complementary to pCVl nucleotides 1546-1569. The PCR prod-uct was cut with Ncol and Xhol and cloned into pRev(N-X).

pRev(MlO) and pRev(A78-79) were constructed using SOEPCR to introduce the desired changes into pRev.44 pRev(MlO)was constructed using pCVl as a template for two initial PCRreactions. One fragment was generated using a top strandoligomer identical to pCVl nucleotides 962-985, together witha bottom strand oligomer containing a mutation from nucleotide1200 to 1205 of pCVl (changing GCTTGA to AGATCT). Thisresulted in the substitution of amino acids 78 and 79 from L-Eto D-L. pRev(M10) is identical to the original M10 mutant.18pRev(A78-79) was constructed using an identical strategy ex-

cept that in this case the top strand oligomer and complemen-tary bottom strand oligomer deleted six nucleotides, resultingin the deletion of amino acids 78 and 79 of Rev.

Purification of HIV-1 Rev proteinpOTSrev in E. coli strain N5151 was inoculated in a 500-ml

culture in LB medium containing ampicillin (100 pg/mi). Whenthe OD660 reached 0.6-0.7, 500 ml of LB warmed to 65°C was

added to the initial culture. The temperature of the incubatorwas raised to 42CC, and the culture was grown for another 5hr. The induced cultures were pelleted, washed once with 50mM Tris-HCl, pH 8.0, and frozen at -70°C.

The pellets from 3 liters of induced culture were resuspendedin 25 ml of ice-cold 50 mM Tris-HCl (pH 8.0), 150 mM NaCl(buffer A). This was passed once through a French press(Aminco) set to high ratio at 1000 psi. All successive steps were

performed at 4°C or on ice. The lysate was centrifuged at 15krpm for 20 min and the supernatant was made 0.5% in poly-ethyleneimine (PEI) by adding a solution of 5% PEI in 50 mMTris-HCl (pH 7.4). Polyethyleneimine has been shown to pre-cipitate nucleic acids and nucleic acid-binding proteins.45 Theresulting suspension was stirred for 20 min and centrifuged as

described above. The pellet was then resuspended in 25 ml ofa solution containing 50 mM Tris-HCl (pH 8.0) and 0.7 M NaCl.The brownish suspension was successively passed through a

10-ml pipette, a 30-ml syringe with no needle, an 18-gauge nee-

dle, a 20-gauge needle, and a 21-gauge needle. This processserved to homogenize the sticky pellet to a uniform consistency.The suspension was then centrifuged as described above. Theresulting pellet was now off-white and the supernatant was

brown. A saturated solution of (NH4)2S04 was slowly added tothe supernatant until the mixture reached 40% saturation. Thissuspension was centrifuged as described above, and the pelletwas resuspended in 10 ml of buffer A, and filtered through a

0.45-/u,m pore size filter. The filtrate was applied to a Mono-S(Pharmacia, Piscataway, NJ) fast protein liquid chromatogra-

MONOCLONAL ANTIBODIES AGAINST HIV-1 REV 947

phy (FPLC) column that had been equilibrated with 60% bufferA and 40% buffer B (buffer B is 50 mM Tris-HCl [pH 8.0],1.5 M NaCl), at a flow rate of 1 ml/rnin. The column was

washed until baseline was reestablished, usually 30 min. A lin-ear gradient of 40-100% buffer B in buffer A was begun over

30 min and 2-ml fractions were collected. A peak containingRev protein eluted between 55 and 60% buffer B. The peakspanned four or five fractions, which were pooled and storedin 1-ml aliquots at —70°C. Protein concentration was deter-mined by the bicinchoninic acid (BCA) assay (Pierce, Rockford,IL).

Generation of monoclonal antibodies against Rev

Rev protein purified from bacteria as described above was

used to immunize BALB/c mice using a modification of theprocedure of Claflin and Williams.46 The mice were given an

initial immunization intraperitoneally with 50 pg of Rev pro-tein in Freund's complete adjuvant. Four weeks later, theywere boosted with 50 pg of Rev protein in Freund's incom-plete adjuvant. A final boost was given 2 weeks later that con-

sisted of 50 pg of Rev protein in phosphate-buffered saline(PBS). The immune response of the mice against Rev was

monitored by enzyme-linked immunosorbent assay (ELISA),using plates coated with Rev protein. To do this, a 96-wellmicrotiter plate was coated with 100 pi of purified Revprotein dissolved in carbonate-bicarbonate buffer (15 mMNa2C03, 35 mM NaHC03 [pH 9.6]) at a concentration of 5pg/mi. The plate was incubated overnight at 37°C, and thefollowing day the plate was washed three times with PBS andblocked with a solution of 2% bovine serum albumin (BSA)in PBS for 30 min at room temperature. The plate was againwashed as described above and 100 pi of serum dilution was

added for 2 hr at 37°C. The plate was then washed as de-scribed above and incubated at room temperature for 1 hr withalkaline phosphatase-conjugated anti-mouse polyvalent anti-serum (Sigma, St. Louis, MO), diluted according to manu-

facturer directions in PBS containing 0.05% Tween 20. Theplate was then washed sequentially three times with PBScontaining 0.05% Tween 20 and once with PBS. The assaywas developed by adding to each well 100 pi of a phos-phatase substrate solution made by adding one tablet ofphosphatase substrate (Kirkegaard & Perry, Gaithersburg,MD) and 1 ml of substrate solution (Kirkegaard & Perry) to4 ml of water. The plate was read at 15 and 30 min after sub-strate addition at 405/540 nm in an automated ELISA platereader.

The spleen of the mouse that showed the best responseagainst Rev protein was fused with SP2 myeloma cells.46 Initialscreening of hybridoma clones was done by ELISA as describedabove. Three clones (designated 2G2, 3G4, and 3H6) were iso-lated, subcloned by limiting dilution, and screened in the same

manner. Ascites was generated by injection of Pristane-primedBALB/c mice with 5 X 106 hybridoma cells in 1 ml of culturemedium. The ascites was harvested at intervals until the micewere sacrificed.

The isotype of each monoclonal antibody was determined byusing a mouse monoclonal isotyping kit (Sigma) according tomanufacturer instructions. The three monoclonals characterizedin this study were all determined to be IgGi(/c).

Construction and expression of glutathioneS-transferase Rev fusion proteins

To map the regions of Rev recognized by the monoclonalantibodies, fragments of Rev were amplified using PCR andsubcloned into the pGEX-2T vector (Pharmacia) for expressionin bacteria. Upstream oligonucleotides that encoded the amino-terminal end of the fragment contained a Bamrll site followedby the Rev sequence. Downstream oligonucleotides that were

complementary to coding strand contained an EcoRl site. ThePCR products were cleaved with BamRl and EcoRl, gel puri-fied, and ligated with pGEX-2T that had been cleaved with thesame enzymes. To express the various fusion proteins, E colicontaining the different pGEX-rev plasmids were grown untilthe OD660 reached 0.4. At this time, isopropyl-/3-D-thiogalac-topyranuside (IPTG) was added to the culture to a final con-

centration of 0.1 mM. The bacteria were grown for an addi-tional 2-3 hr, whereafter 1 ml of culture was pelleted. The pelletwas resuspended in 200 pi of 2X sample buffer and analyzedby immunoblot analysis.

Immunoprecipitations for protein analysisImmunoprecipitation (IP) analysis using total cell protein ex-

tracts was performed as previously described.43,47'48 Typically,one 150-mm plate of transfected cells was used. Cells were har-vested by scraping, washed twice in PBS, and resuspended in2 ml of NET-2 (50 mM Tris-HCl [pH 7.4], 150 mM NaCl,0.05% Nonidet P-40) or RIPA (radioimmunoprecipitation as-

say) (50 mM Tris-HCl [pH 7.2], 150 mM NaCl, 0.1% sodiumdodecyl sulfate [SDS], 0.1% deoxycholate, 0.1% Triton X-100)buffer. Cells were subjected to three 30-sec cycles of sonica-tion with chilling on ice between cycles, using a Branson(Fischer Scientific, Rochester, NY) sonicator with microtip at

setting 3. The lysates were cleared of debris by centrifugationat 10 krpm for 20 min at 4°C.

Antiserum used for IP was first bound to protein A- or pro-tein G-Sepharose by incubation of 10 pi of polyclonal rabbitor monoclonal ascites with 10% (w/v) protein A- or G-Sepharose that had been prepared by swelling overnight in ei-ther NET-2 or RIPA buffer. The Sepharose was washed fourtimes in 1 ml of the respective buffer, and then resuspended toa final volume of 1 ml. The antiserum or ascites was added andthe mixture was rocked vigorously at room temperature for 2hr. The Sepharose bound with antibody was then pelleted andwashed as above. One-third to one-half of the 2-ml extract was

then added to the pellet. The mixture was incubated on ice for1 hr with vigorous shaking. The final antigen-antibody-Sepharose mixture was then pelleted, and washed four or fivetimes with 1 ml of buffer. The final pellet was resuspended in20 pi of water and 20 pi of 2X SDS-PAGE sample buffer andprocessed for immunoblot analysis as described below.

Western (immuno-) blot analysis and anti-Revpolyclonal antiserum

Immunoblot analysis was performed as previously de-scribed.3-49 Proteins were transferred onto Immobilon poly-vinylidene difluoride (PVDF) membranes (Millipore, Bedford,MA). Rev polyclonal antiserum was produced in rabbits using

948 ORSINI ET AL.

an affinity-purified glutathione S-transferase-Rev (GST-Rev) fusion protein containing amino acids 61-116 of Rev asan immunogen.

Cells and transfectionsCMT3 COS cells50-51 were transfected by the DEAE-

dextran method as previously described.49 The cells were har-vested at 72 hr posttransfection.

RESULTS

Overexpression of Rev in Escherichia coliOur initial goal was to overexpress and purify large amounts

of biologically active Rev protein. Other groups have previ-ously described the purification of Rev to near homogene-jjy 12,27,52,53 However, most of these methods involve denatu-ration of the Rev protein during purification, or use modifiedRev proteins. We decided to attempt to express authentic Rev

12 3 4 5 6 78 9101112131415161718

PASAEH801 pOTSrev01 234701234

hours7 post-induction

Ml 2 3 4 5 6 7 8 9 10 11 12

PASAEH801 pOTSrev hm]rs01 2 34701 2347 post-induction

12 3 4 5 6 7 9 1011 12

FIG. 1. SDS-PAGE and Western blot analysis of proteinexpression in E. coli N5151 transformed with pOTSrevand pASAEH801. Strain N5151 bacteria containing eitherpASAEH801 or pOTSrev were grown and induced as describedin Materials and Methods. Cell lysates were analyzed at thetimes indicated on an 18% SDS polyacrylamide gel. Proteinswere detected by staining with Coomassie blue (A) or byWestern blot (B) using a polyclonal antiserum directed againstthe C terminus of Rev. In both panels, lanes 1-6 are lysates ofN5151 containing pASAEH801 and lanes 7-12 are lysates ofN5151 containing pOTSrev. The positions of proteins used asmolecular weight markers are indicated and their sizes are in-dicated in thousands of daltons (kDa). The arrows indicate themajor protein species observed after induction.

^18

FIG. 2. SDS-PAGE analysis of steps in the purification ofRev expressed in E. coli. Purification of Rev was performed asdescribed in Materials and Methods. Samples from the variouspurification steps were separated on an 18% SDS polyacry-lamide gel, and proteins were visualized by staining withCoomassie blue. Lane 1, whole bacterial lysate; lane 2, Frenchpress supernatant; lane 3, French press pellet; lane 4, super-natant after polyethyleneimine precipitation; lane 5, PEI pelletafter elution; lane 6, 0.7 M salt eluate from PEI precipitate; lane7, 40% ammonium sulfate pellet; lanes 8-18, fractions fromMono-S column. The position of the Rev protein is indicatedby an arrow. The positions of proteins used as molecular weightmarkers are indicated.

protein in large amounts and to devise a simple and rapid pu-rification method for this protein that did not involve denatu-ration.

To express the Rev protein, the region of the pCVl clone39containing the rev gene was subcloned between the Ncol andXhol sites of the expression vector pOTSNcol3336 using thePCR strategy described in Materials and Methods. This resultedin the creation of the plasmid pOTSrev. The bacterial strainN5151 was transformed with pOTSrev. This strain contains a

temperature-sensitive mutation in the A repressor. Thus ex-

pression of Rev was not expected when the bacteria were prop-agated at 32°C, but shifting to 42°C was expected to result inthe inactivation of the repressor and subsequent expression ofRev. A typical induction is shown in Fig. 1. As a control, a cul-ture of N5151 containing the plasmid pASAEH801 was also in-duced. pASAEH801 contains the influenza NS1 gene under thecontrol of the Pl promoter and expresses a 25-kDa protein on

induction.36Aliquots of both cultures were harvested at various times af-

ter thermal induction and lysates were analyzed by SDS gelelectrophoresis. The proteins were detected by staining withCoomassie blue (Fig. 1A) or by Western blotting (Fig. IB) us-

ing a rabbit polyclonal antiserum directed against Rev. Asshown in Fig. 1A, on induction of the N5151 strain containingpOTSrev, a band migrating at an apparent molecular size of ap-proximately 18 kDa was seen by Coomassie staining (Fig. 1A,lanes 7-12). This species was not present in the induced strain

MONOCLONAL ANTIBODIES AGAINST HIV-1 REV 949

N5151 containing pASAEH801 ; however, the predicted 25-kDaspecies was observed (Fig. 1A, lanes 1-6). Figure IB showsthat the rabbit polyclonal antiserum directed against the C ter-minus of Rev recognized an 18-kDa protein from pOTSrev, as

well as a slightly larger species, but that the antiserum, as ex-

pected, did not detect the 25-kDa species produced by induc-tion of bacteria containing pASAEH801.

Purification of Rev

To purify the Rev protein, we developed a protocol in whichthe bacteria were first lysed using a French press. This was fol-lowed by precipitation with PEI and ammonium sulfate frac-tionation. The Rev protein was then further purified usingFPLC. The procedure is described in detail in Material andMethods. Representative samples from the various steps in theprotocol were analyzed on an 18% SDS polyacrylamide gelstained with Coomassie blue and are shown in Fig. 2.

After lysis of bacteria followed by passage through a Frenchpress, most of the induced protein appeared in the soluble frac-tion (Fig. 2, lanes 1-3). Most of the induced protein was pre-cipitated by the addition of PEI (Fig. 2, lanes 4—6). The proteinwas dissociated from the nucleic acid pellet by homogenizationof the pellet in high-salt buffer (Fig. 2, lane 6), leaving most ofthe nucleic acids in the precipitate. This supernatant was con-

centrated and further purified by making the supernatant 40%in ammonium sulfate (Fig. 2, lane 7). The 40% ammonium sul-fate pellet was resuspended in 50 mM Tris, pH 8.0, containing0.15 M NaCl, filtered to remove insoluble material, and appliedto a Mono-S cation-exchange FPLC column in the same buffercontaining 0.6 M NaCl. The majority of contaminating proteins,but no Rev protein, were eluted in the column flow-through.To elute the Rev protein, a gradient of 0.6 to 1.5 M NaCl was

applied to the column. Rev protein was eluted when the saltconcentration reached 0.8-0.9 M (Fig. 2, lanes 8-18). TheRev protein was apparently homogeneous as determined byCoomassie staining.

Production and characterization ofmonoclonal antibodies

The purified Rev protein was used to immunize mice andspleens were used to generate hybridomas as described inMaterials and Methods. Clones of hybridomas were thenscreened by ELISA, using plates that were coated with purifiedRev protein. Antibodies obtained from three different positivehybridoma clones were further examined for their ability to de-tect Rev protein by immunoblot using lysates of cells trans-fected with vectors expressing Rev (Fig. 3). In this analysis, we

tested whether the different antibodies could detect wild-typeRev protein, as well as two mutant Rev proteins, RevMlO andRevA78-79. In RevMlO amino acid residues 78 and 79 are

changed from L-E to D-L,18 whereas RevA78-79 is deleted inthese amino acids. Both mutant proteins were shown to be/rans-dominant inhibitors of Rev function. All three mono-

clonal antibodies recognized both the wild-type and the mutantRev proteins (Fig. 3A-C, lanes 2-4), which were not detectedin untransfected cells (Fig. 3A-C, lane 1). Rev was usually de-tected as a doublet band. The upper band may result from post-translational modification of Rev, such as phosphorylation.54Monoclonal antibody 3G4 also detected a species with an ap-

2G2 3G4 3H6

FIG. 3. Western blot analysis of the ability of anti-Rev mon-oclonal antibodies to recognize wild-type and mutant Rev pro-teins. CMT3 cells were transfected with 5 pg of pRev (lane 2),pRevMlO (lane 3), or pRevA78-79 (lane 4). Lane 1 of (A-C)contains untransfected cells. Proteins were separated on an 18%SDS polyacrylamide gel and electrophoretically transferred toan Immobilon membrane. The blot was developed with ascitescontaining the monoclonal antibody 2G2 (A), 3G4 (B), or 3H6(C). All of the different ascites fluids were diluted 1:1000 priorto use.

parent molecular size of 28 kDa that was present in both un-transfected and transfected cells. Additional bands that migratedbelow Rev were detected only in transfected cells. These maybe the result of proteolytic degradation of the Rev and mutantRev proteins.

The antibodies were then tested for their ability to immuno-precipitate Rev from transfected cells. For this experiment, ex-

tracts of transfected cells were prepared in either NET-2 or

RIPA buffer. The three Rev monoclonal antibodies, a mouse

polyclonal antiserum, and a control monoclonal antibody (IT,directed against purified tropomyosin from Schistosoma man-

soni) were used for immunoprecipitations as described inMaterials and Methods. The RIPA buffer contains sodium do-decyl sulfate and would be expected to disrupt conformationalepitopes to a greater extent than the NET-2 buffer. The im-munoprecipitates were subjected to immunoblot analysis usinga Rev polyclonal antiserum developed in rabbits, as describedin Materials and Methods (Fig. 4). As seen in Fig. 4, the poly-clonal mouse antiserum and the three monoclonal antibodies(2G2, 3G4, and 3H6) all precipitated Rev from the lysates madein NET-2 buffer (lanes 5, 7, 9, and 11) while the control anti-body IT did not (lane 3). In contrast, 2G2 failed to immuno-precipitate Rev in RIPA buffer (cf. lane 8 to lanes 6, 10, and12).

The ability of the monoclonal antibodies to detect Rev byindirect immunofluorescence was also examined. To do this,CMT3 cells were transfected with the Rev-expressing plasmidpRev or with its parental vector pCMV, which lacks the rev

cDNA insert. Strong specific immunofluoresence was observedin pRev-transfected cells using antibodies 3G4 and 3H6, re-

spectively. In contrast, only weak, nonspecific immunofluores-ence was observed with antibody 2G2. No specific fluorescence

ORSINI ET AL.

ext IT

N R N R

FIG. 4. Western blot analysis of the ability of anti-Rev mon-oclonal antibodies to immunoprecipitate Rev expressed in trans-fected cells. CMT3 cells in 150-mm dishes were transfectedwith 15 ixg of pRev. Cell extracts were made as described inMaterials and Methods in either NET-2 (N) or RIPA (R) buffer.In lanes 1 and 2, 20 pi of the 2-ml extract was added directlyto the gel without immunoprecipitation. Immunoprecipitationswere performed with one-fifth of each extract, using the fol-lowing antibodies: Lanes 3 and 4, IT, a monoclonal ascitesdirected against schistosome tropomyosin; lanes 5 and 6,antiserum directed against Rev from the mouse used for pro-duction of the monoclonal antibodies; lanes 7 and 8, 2G2 as-

cites; lanes 9 and 10, 3G4 ascites; lanes 11 and 12, 3H6 ascites.Immunoprecipitated protein was detected by Western blotanalysis using a polyclonal Rev antiserum directed against thecarboxy terminus of Rev. The arrow indicates the position ofRev protein.

was obtained in the control transfected cells with any of the an-

tibodies (data not shown).To determine the specific regions of Rev recognized by the

different monoclonal antibodies, we made fusion constructs be-tween various regions of Rev and schistosome glutathione 5transferase (GST). This was done by subcloning different fragments of the rev gene into the pGEX-2T vector as described inMaterials and Methods. The resultant constructs contained GSTfused to Rev amino acids 1-30, 1-61, 30-61, 61-90, 90-116,and 61-116. We also utilized a plasmid that expressed GSTfused to the full-length Rev protein. This plasmid was a kindgift from M. Zapp (University of Massachusetts, Amherst, MA).The expression of these fusion proteins was induced in bacte-ria using IPTG as described in Materials and Methods, and an-

alyzed by SDS-PAGE. After induction, all of the plasmids ex-

pressed GST-Rev fusion proteins that were detected both bystaining with Coomassie blue and by immunoblot using a Revpolyclonal antibody (data not shown).

The same extracts used to analyze expression of the fusionproteins were then subjected to immunoblot analysis using as-

cites fluid containing the different monoclonal antibodies(Fig. 5B-D) or the control antibody IT (Fig. 5A). Monoclonalantibody 3H6 detected the full-length Rev fusion protein as wellas fusion proteins containing amino acids 1-61 and 30-61 ofRev (Fig. 5C, lanes 1, 3, and 6). Antibody 3G4 detected thefull-length Rev fusion protein as well as fusion proteins con-

taining amino acids 61-116 and 90-116 of Rev (Fig. 5D, lanes

1, 2, and 4). In most cases, lower molecular weight species thatprobably represented breakdown products of the fusion proteinswere also detected. Antibody 2G2 did not specifically recog-nize any of the fusion proteins at dilutions of antibody that rec-

ognized Rev expressed in transfected cells (Fig. 5B). This maybe caused by the inability of the fusion proteins to fold into a

proper conformation for recognition by antibody 2G2. Fromthis analysis we could conclude that monoclonal antibody 3H6recognizes an epitope contained within amino acids 30-61, andthat 3G4 recognizes an epitope contained within amino acids90-116 of Rev. In contrast, 2G2 apparently does not recognizean epitope of Rev that is present within the context of these fu-sion proteins.

Amino acids 30-61 of Rev contain a stretch of arginineresidues that has been shown to be involved in the nuclear im-port of the protein. Studies in our laboratory (data not shown)and elsewhere55'56 have indicated that a Rev protein deleted inthese amino acids (RevA38^44; RRNRRRR) remains localizedin the cytoplasm of transfected cells and behaves functionallyas a dominant negative mutant. It was thus of interest to deter-

1 2 3 4 5 6 7 1 2 3 4 5 6 7

1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8

FIG. 5. Epitope mapping of anti-Rev monoclonal antibodiesusing GST-Rev fusion proteins. Bacteria containing plasmidsthat expressed different GST-Rev fusion proteins were inducedas described in Materials and Methods. Proteins from bacteriallysates were separated on a 15% SDS polyacrylamide gel andelectrophoretically transferred to an Immobilon membrane. Thedifferent lanes contain extracts from bacteria that express thefollowing: lane 1, GST-Rev(l-116); 2, GST-Rev(61-116); 3,GST-Rev(l-61); 4, GST-Rev(90-116); 5, GST-Rev(61-90);6, GST-Rev(30-61); 7, GST-Rev(l-30). The blots were de-veloped with a monoclonal ascites directed against schistosometropomyosin (IT, diluted 1:10,000) (A) or the different mono-

clonal ascites directed against Rev: 2G2, diluted 1:250 (B), 3H6,diluted 1:10,000 (C), or 3G4, diluted 1:500 (D).

MONOCLONAL ANTIBODIES AGAINST HIV-1 REV 951

A B12 3 12 3

wpolyclonal 3H6

FIG. 6. Western blot analysis of the ability of monoclonalantibody 3H6 and polyclonal serum to recognize Rev deletionmutants. Lanes 1 and 2 contain extracts from CMT3 cells trans-fected with 5 pg of pRev(ll-116) (lane 1) or pRev (lane 2).Lane 3 contains an extract of strain N5151 bacteria containingpOTSRevA38^14. Proteins were separated on an 18% SDSpolyacrylamide gel and electrophoretically transferred to anImmobilon membrane. The blot was developed with a Revpolyclonal antiserum directed against the C terminus of Revthat was used at a 1:250 dilution (A) or Rev monoclonal anti-body 3H6 that was used at a 1:1000 dilution (B).

mine if monoclonal antibody 3H6, which recognized an epitopebetween Rev amino acids 30 and 61, was capable of binding toRevA38^14. To do this, E. coli containing a plasmid expressinga Rev protein deleted in amino acids 38^14, pOTSrev(A38—44),were grown and induced as described for pOTSrev. Bacteriallysates were subjected to immunoblot analysis using either a

polyclonal serum directed against the carboxy terminus of Revor monoclonal antibody 3H6. As controls, lysates from CMT3cells transfected with pRev or pRev(ll-116) were also ana-

lyzed. The results of this experiment are shown in Fig. 6A andB. Figure 6A demonstrates that the polyclonal anti-Rev serum

recognized the protein produced from all of the constructs (lanes1-3). In contrast, Fig. 6B shows that monoclonal antibody 3H6recognized only wild-type Rev (lane 2) and Rev(ll-116) (lane1), whereas the RevA38-44 protein was not recognized by thisantibody (lane 3).

DISCUSSION

In this article we describe a procedure for the expression andpurification of the HIV-1 Rev protein in E. coli. The method we

developed to purify Rev takes advantage of the basic, positivelycharged nature of the Rev protein, as well as its ability to inter-act with nucleic acids. Most of the Rev protein was found to besoluble after lysis of the bacteria and coprecipitated with bacte-rial nucleic acids when these were precipitated with the polyca-tion polyethyleneimine. Rev, as well as other proteins, could bereleased from the nucleic acid pellet by using a high salt con-centration. This eluate was further purified and concentrated byammonium sulfate precipitation. The final step consisted ofchromatography over a cation-exchange column under condi-tions in which Rev, but not contaminating proteins, bound to thecolumn. Rev was then specifically eluted using a salt gradient.The protein purified in this manner was apparently homogeneouswhen analyzed by staining with Coomassie blue.

We have also used this method to purify the M10 mutantRev protein that contains substitutions of Rev amino acids 78and 79, as well as a mutant that deletes these amino acids.53a

As mentioned previously, Rev proteins containing mutations inthis region of the protein have been shown to inhibit the func-tion of wild-type Rev in trans.18-29-51 We believe that the ex-

pression and purification method described in this article willbe applicable to the purification of several other mutant Revproteins, provided that they retain their nucleic acid-bindingproperties. We have also demonstrated that the Rev protein pu-rified using this method is functional in vivo, using a chloro-quine-mediated uptake assay similar to that previously de-scribed for the HIV-1 Tat protein.533

Three monoclonal antibodies directed against Rev were pro-duced using the purified Rev protein. The three antibodies, 2G2,3G4, and 3H6, each recognized Rev protein on Western blots.To map the regions in Rev to which the antibodies were di-rected, we analyzed the ability of the antibodies to recognizefusion proteins that consisted of various fragments of Rev fusedto glutathione S-transferase (GST). This analysis revealed that3G4 was directed against an epitope contained within aminoacids 90-116 of Rev and that 3H6 was directed against an epi-tope between amino acids 30 and 61 of Rev. We also showedthat 3H6 failed to recognize a mutant Rev protein that containeda deletion of amino acids 38 to 44, suggesting that 3H6 is di-rected against an epitope that lies between or overlaps aminoacids 38 to 44. This conclusion must be regarded as tentative,however, because it has not yet been confirmed by the demon-stration of binding between 3H6 and a peptide correspondingto the sequence of the deleted region. This arginine-rich, posi-tively charged region of Rev has been implicated as being in-volved in localizing Rev to the nucleus and in binding to theRRE.18"23,25

Interestingly, 2G2 failed to recognize any of the GST-Revfusion proteins on Western blots, including full-length Revfused to GST. Because 2G2 did recognize authentic Rev pro-tein expressed in CMT-3 cells on Western blots, it is possiblethat this antibody only recognizes a conformational epitope thatis not present when Rev is expressed as a fusion protein. Thisconformation may be restored when the Rev protein is trans-ferred from a denaturing SDS gel to a membrane. 2G2 was ableto immunoprecipitate Rev from a transfected cell extract whenthe extract was prepared in NET-2 buffer, which contains thedetergent Nonidet P-40 (NP-40). However, when the extractwas prepared in RIPA buffer, which contains SDS, deoxy-cholate, and Triton X-100, 2G2 was unable to immunoprecip-itate Rev. In contrast, both 3H6 and 3G4 were able to im-munoprecipitate Rev in either buffer. This further supports thenotion that 2G2 is directed against a sensitive conformationalepitope with Rev that is destroyed on harsh treatment with de-tergents.

The novel monoclonal antibodies we have described here are

likely to be useful reagents for investigating the ability of Revto interact with cellular proteins and nucleic acids. For exam-

ple, the ability of 3G4 to recognize a region of Rev that hasbeen previously shown to be dispensable for Rev function24'32may be of advantage in examining the association of Rev withRRE RNA in vivo as well as with cellular factors that may berequired for Rev function. On the other hand, because 3H6 rec-

ognizes the RRE-binding domain of Rev, it may be able to dis-tinguish between the pool of Rev that is bound to RNA in vivoand the free pool. These types of studies are currently under-way in our laboratory.

952 ORSINI ET AL.

REFERENCES

1. Emerman M, Vazeux R, and Peden K: The rev gene product of thehuman immunodeficiency virus affects envelope-specific RNA lo-calization. Cell 1989;57:1155-1165.

2. Felber BK, Hadzopoulou-Cladaras M, Cladaras C, Copeland T, andPavlakis GN: Rev protein of human immunodeficiency virus type1 affects the stability and transport of the viral mRNA. Proc NatiAcad Sei USA 1989;86:1496-1499.

3. Hammarskjöld M-L, Heimer J, Hammarskjöld B, Sangwan I,Albert L, and Rekosh D: Regulation of human immunodeficiencyvirus env expression by the Rev gene product. J Virol 1989;63:1959-1966.

4. Malim MH, Hauber J, Le SV, Maizel JV, and Cullen BR: The HTV-1 rev trans-activator acts through a structured target sequence toactivate nuclear export of unspliced viral mRNA. Nature (London)1989;338:254-257.

5. Schwartz S, Felber BK, Fenyö EM, and Pavlakis GN: Env and Vpuproteins of human immunodeficiency virus type 1 are producedfrom multiple bicistronic mRNAs. J Virol 1990;64:5448-5456.

6. Garrett ED, Tiley LS, and Cullen BR: Rev activates expression ofthe human immunodeficiency virus type 1 vif and vpr gene prod-ucts. J Virol 1991;65:1653-1657.

7. Schwartz S, Felber BK, and Pavlakis GN: Expression of humanimmunodeficiency virus type 1 vi/and vpr mRNAs is Rev-depen-dent and regulated by splicing. Virology 1991;183-677-686.

8. Dayton AI, Terwilliger EF, Potz J, Kowalski M, Sodroski JG, andHaseltine WA: Cis-acting sequences responsive to the rev geneproduct of the human immunodeficiency virus. J AIDS 1988;1:441^152.

9. Rosen CA, Terwilliger E, Dayton A, Sodroski JG, and HaseltineWA: Intragenic cis-acting art gene-responsive sequences of the hu-man immunodeficiency virus. Proc Nati Acad Sei USA 1988;85:2071-2075.

10. Hadzopoulou-Cladaras M, Felber BK, Cladaras C, Athanas-sopoulos A, Tse A, and Pavlakis GN: The rev (trs/art) protein ofhuman immunodeficiency virus type 1 affects viral mRNA and pro-tein expression via a cis-acting sequence in the env region. J Virol1989;63:1265-1274.

11. Malim MH, Tiley LS, McCarn DF, Rusche JR, Hauber J, andCullen BR: HTV-1 structural gene expression requires binding ofthe rev trans-activator to its RNA target sequence. Cell 1990;60:675-683.

12. Daly TJ, Cook KS, Gray GS, Maione TE, and Rusche JR: Specificbinding of HTV-1 recombinant Rev protein to the Rev-responsiveelement in vitro. Nature (London) 1989;342:816-819.

13. Zapp ML and Green MR: Sequence-specific RNA binding by theHTV-1 Rev protein. Nature (London) 1989;342:714-716.

14. Cook KS, Fisk GJ, Hauber J, Usman N, Daly TJ, and Rusche JR:Characterization of HIV-1 Rev protein: Binding stoichiometry andminimal RNA substrate. Nucleic Acids Res 1991;19:1577-1583.

15. Tiley LS, Malim MH, Tewary HK, Stockley PG, and Cullen BR:Identification of a high-affinity RNA-binding site for the humanimmunodeficiency virus type 1 Rev protein. Proc Nati Acad SeiUSA 1992;89:758-762.

16. Arrigo SJ, Heaphy S, and Haines JK: In vivo binding of wild-typeand mutant human immunodeficiency virus type 1 Rev proteins:Implications for function. J Virol 1992;66:5569-5575.

17. Dillon PJ, Nelbock P, Perkins A, and Rosen CA: Function of thehuman immunodeficiency virus types 1 and 2 Rev proteins is de-pendent on their ability to interact with a structured region presentin env gene mRNA. J Virol 1990;64:4428-4437.

18. Malim MH, Bohnlein S, Hauber J, and Cullen BR: Functional dis-section of the HIV-1 Rev trans-activator—derivation of a trans-dominant repressor of Rev function. Cell 1989;58:205-214.

19. Olsen HS, Cochrane AW, Dillon PJ, Nalin CM, and Rosen C:

Interaction of the human immunodeficiency virus type 1 Rev pro-tein with a structured region in env mRNA is dependent on multi-mer formation mediated through a basic stretch of amino acids.Genes Dev 1990;4:1357-1364.

20. Cochrane A, Kramer R, Ruben S, Levine J, and Rosen CA: Thehuman immunodeficiency virus Rev protein is a nuclear phospho-protein. Virology 1989;171:264-266.

21. Cochrane AW, Perkins A, and Rosen CA: Identification of se-

quences important in the nucleolar localization of human immun-odeficiency virus Rev: Relevance of nucleolar localization to func-tion. J Virol 1990;64:881-885.

22. Berger J, Aepinus C, Dobrovnik M, Fleckenstein B, Hauber J, andBohnlein E: Mutational analysis of functional domains in the HIV-1 Rev trans-regulatory protein. Virology 1991;183:630-635.

23. Malim MH and Cullen BR: HTV-1 structural gene expression re-

quires the binding of multiple Rev monomers to the viral RRE:Implications for HIV-1 latency. Cell 1991;65:241-248.

24. Olsen HS, Beidas S, Dillon P, Rosen CA, and Cochrane AW:Mutational analysis of the HTV-1 Rev protein and its target se-

quence, the Rev responsive element. J AIDS 1991;4:558-567.25. Zapp ML, Hope TJ, Parslow TG, and Green MR: Oligomerization

and RNA binding domains of the type 1 human immunodeficiencyvirus Rev protein: A dual function for an arginine-rich binding mo-

tif. Proc Nati Acad Sei USA 1991;88:7734-7738.26. Daly TJ, Rusche JR, Maione TE, and Frankel AD: Circular dichro-

ism studies of the HTV-1 Rev protein and its specific RNA bind-ing site. Biochemistry 1990;29:9791-9795.

27. Wingfield PT, Stahl SJ, Payton MA, Venkatesan S, Misra M, andSteven AC: TDV-l Rev expressed in recombinant Escherichia coli:Purification, polymerization, and conformational properties.Biochemistry 1991;30:7527-7534.

28. Kjems J, Brown M, Chang DD, and Sharp PA: Structural analysisof the interaction between the human immunodeficiency virus Revprotein and the Rev response element. Proc Nati Acad Sei USA1991;88:683-687.

29. Mermer B, Felber BK, Campbell M, and Pavlakis GN:Identification of trans-dominant HIV-1 Rev protein mutants by di-rect transfer of bacterially produced proteins into human cells.Nucleic Acids Res 1990;18:2037-2044.

30. Ahmed YF, Hanly SM, Malim MH, Cullen BR, and Greene WC:Structure-function analyses of the HTLV-I Rex and HIV-1 RevRNA response elements: Insights into the mechanism of Rex andRev action. Genes Dev 1990;4:1014-1022.

31. Bogerd H and Greene WC: Dominant negative mutants of humanT-cell leukemia virus type 1 and human immunodeficiency virustype 1 fail to multimerize in vivo. J Virol 1993;67:2496-2502.

32. Malim MH, McCarn DF, Tiley LS, and Cullen BR: Mutational de-finition of the human immunodeficiency virus type 1 Rev activa-tion domain. J Virol 1991;65:4248^1254.

33. Rosenberg M, Ho Y-S, and Shatzman A: The use of pKC30and its derivatives for controlled expression of genes. MethodsEnzymol 1983;101:123-138.

34. Shatzman AR and Rosenberg M: Efficient expression of heterolo-gous genes in Escherichia coli. The pAS vector system and its ap-plications. Ann NY Acad Sei 1986;478:233-248.

35. Shatzman A and Rosenberg M: Expression, identification and char-acterization of recombinant gene products in Escherichia coli.Methods Enzymol 1987;152:661-673.

36. Shatzman AR and Rosenberg M: The pAS vector system and itsapplication to heterologous gene expression in Escherichia coli.Hepatology 1987;7:30S-35S.

37. Saiki RK, Scharf S, Faloona F, Mullis KB, Horn GT, Erlich HA,and Arnheim N: Enzymatic amplification of beta-globin genomicsequences and restriction site analysis for diagnosis of sickle cellanemia. Science 1985;230:1350-1354.

38. Scharf SJ, Horn GT, and Erlich HA: Direct cloning and sequence

MONOCLONAL ANTIBODIES AGAINST HTV-1 REV 953

analysis of enzymatically amplified genomic sequences. Science1986;233:1076-1078.

39. Arya SK, Guo C, Josephs SF, and Wong-Staal F: Trans-activatorgene of human T-lymphotropic virus type ITT (HTLV-III). Science1985;229:69-73.

40. Horton RM, Hunt HD, Ho SN, Pullen JK, and Pease LR:Engineering hybrid genes without the use of restriction enzymes:gene splicing by overlap extension. Gene 1989;77:61-68.

41. Smith AJ, Cho MI, Hammarskjöld ML, and Rekosh D: Human im-munodeficiency virus type 1 Pr55gag and Prl60gag-pol expressedfrom a simian virus 40 late replacement vector are efficientlyprocessed and assembled into viruslike particles. J Virol1990;64:2743-2750.

42. Jeang KT, Rawlins DR, Rosenfeld PJ, Shero JH, Kelly TJ, andHayward GS: Multiple tandemly repeated binding sites for cellularnuclear factor 1 that surround the major immediate-early promotersof simian and human cytomegalovirus. J Virol 1987;61:1559-1570.

43. Rekosh D, Nygren A, Flodby P, Hammarskjöld ML, and WigzellH: Coexpression of human immunodeficiency virus envelope pro-teins and tat from a single simian virus 40 late replacement vec-

tor. Proc Nati Acad Sei USA 1988;85:334-338.44. Ho SN, Hunt HD, Horton RM, Pullen JK, and Pease LR: Site-di-

rected mutagenesis by overlap extension using the polymerasechain reaction. Gene 1989;77:51-59.

45. Jendrisak J: In: Protein Purification: Micro to Macro. Alan R. Liss,New York, 1987, pp. 75-97.

46. Claflin L and Williams K: Mouse myeloma-spleen cell hybrids:Enhanced hybridization frequencies and rapid screening proce-dures. Curr Topics Microbial Immunol 1978;81:107-109.

47. Kessler SW: Rapid isolation of antigens from cells with a staphy-lococcal protein A-antibody adsorbent: Parameters of the interac-tion of antibody-antigen complexes with protein A. J Immunol1975;115:1617-1624.

48. Matter L, Schopfer K, Wilhelm JA, Nyffenegger T, Parisot RF,and De REM: Molecular characterization of ribonucleoprotein anti-gens bound by antinuclear antibodies. A diagnostic evaluation.Arthritis Rheum 1982;25:1278-1283.

49. Hammarskjöld M-L, Wang S-C, and Klein G: High-level expres-

sion of the Epstein-Barr virus EBNA1 protein in CV1 cells and hu-man lymphoid cells using a SV40 late replacement vector. Gene1986;43:41-50.

50. Gluzman Y: SV40-transformed simian cells support the replicationof early SV40 mutants. Cell 1981;23:175-182.

51. Gerard RD and Gluzman Y: New host cell system for regulatedsimian virus 40 DNA replication. Mol Cell Biol 1985;5:3231-3240.

52. Goh WC, Sodroski JG, Rosen CA, and Haseltine WA: Expressionof the art gene protein of human T-lymphotropic virus type III(HTLV-m/LAV) in bacteria. J Virol 1987;61:633-637.

53. Cochrane AW, Chen C-H, Kramer R, Tomchak L, and Rosen CA:Purification of biologically active human immunodeficiency virusRev protein from Escherichia coli. Virology 1989;173:335-337.

54. Cochrane AW, Golub E, Volsky D, Ruben S, and Rosen CA:Functional significance of phosphorylation to the human immun-odeficiency virus Rev protein. J Virol 1989;63:4438^1440.

55. Kubota S, Nosaka T, Cullen BR, Maki M, and Hatanaka M: Effectsof chimeric mutants of human immunodeficiency virus type 1 Revand human T-cell leukemia virus type I Rex on nucleolar target-ing signals. J Virol 1991;65:2452-2456.

56. Kubota S, Furuta R, Maki M, and Hatanaka M: Inhibition of hu-man immunodeficiency virus type I rev function by a rev mutantwhich interferes with nuclear/nucleolar localization of rev. J Virol1992;66:2510-2513.

57. Venkatesh LK and Chinnadurai G: Mutants in a conserved regionnear the carboxy-terminus of HIV-1 Rev identify functionally im-portant residues and exhibit a dominant negative phenotype.Virology 1990;178:327-330.

Address reprint requests to:David Rekosh

Myles H. Thaler Center for AIDS andHuman Retrovirus Research

University of VirginiaHSC Box 441

Charlottesville, Virginia 22908