the matrix protein of hiv-1 is not sufficient for assembly ... · cv-1 cells were infected with...

The Matrix Protein of HIV-1 Is Not Sufficient for Assembly and Release of Virus-like Particles

Angela M. Giddings,† G. D. Ritter, Jr.,* and Mark J. Mulligan*,†,1

Departments of *Medicine and †Microbiology, University of Alabama at Birmingham, Birmingham, Alabama 35294-2170

Received May 6, 1997; returned to author for revision June 18, 1997; accepted June 9, 1998

The matrix (MA) proteins of human immunodeficiency virus type 1 (HIV-1) and simian immunodeficiency virus (SIV) areknown to be important for the targeting and assembly of lentiviral proteins. The objective of the present study was todetermine whether the MA protein of HIV-1 was sufficient for particle assembly and release. Eukaryotic expression ofwild-type HIV-1 Gag-Pol, HIV-1 MA alone, or SIV MA alone was analyzed with radio-immunoprecipitation, density centrifu-gation, and a protease protection assay. Cells that expressed HIV-1 Gag-Pol or SIV MA alone released virus-like particles(VLPs) with sucrose gradient densities of 1.15 or 1.12 g/ml, respectively. The MA and/or capsid proteins in these particleswere protected from protease degradation, indicating the presence of a protective outer membrane. Expression of HIV-1 MAprotein alone resulted in release of MA which pelleted through a 20% sucrose cushion but failed to enter a 20–60% sucrosegradient and was not protected from protease degradation. The MA protein of SIV was previously reported to be sufficientfor production of VLPs (S. A. Gonzalez, H, K, Affrachino, H. R. Gelderblom, and A. Burney. Virology 194, 548–556, 1993; V. Liska,D. Spehner, M. Mehtali, D. Schmitt, A. Kirn, and A. M. Aubertin. J. Gen. Virol. 75, 2955–2962, 1994). Our study confirmed thatresult, but indicated that the MA protein of HIV-1 was not sufficient to assemble and release VLPs. © 1998 Academic Press

INTRODUCTION

HIV-1 and SIV are lentiviruses that utilize a type-C-likeretroviral strategy for morphogenesis. The precise mo-lecular mechanisms that achieve virion assembly, bud-ding, and maturation are still largely unknown. As withother oncoviruses and lentiviruses, the HIV-1 and SIVGag polyproteins in the absence of other viral proteinsare sufficient for the assembly and budding of VLPs(Gheysen et al., 1989; Shioda and Shibuta, 1990; Willsand Craven, 1991) However, several other viral mole-cules, including glycoproteins, enzymes, and RNA, arerequired for the production of infectious particles. HIV-1and SIV Gag polyproteins assemble into immature viri-ons during budding and are subsequently cleaved by theviral protease to produce mature virions containing theprocessed Gag subunit proteins: matrix (MA), capsid(CA), nucleocapsid (NC), and p6 (Henderson et al., 1998;Mervis et al., 1988).

The HIV-1 MA protein is modified by the addition of amyristic acid moiety at an N-terminal glycine. It has beenshown that this modification plus other downstream se-quences are responsible for the targeting of MA to theplasma membrane (Bryant and Ratner, 1990; Freed et al.,1994; Lee and Linial, 1994; Spearman et al., 1994; Zhou etal., 1994). Mutations in the HIV-1 MA protein can redirect

particle formation to internal cellular compartments(Facke et al., 1993; Freed et al., 1994; Gallina et al., 1994;Chazal et al., 1995). MA also contains a nuclear localiza-tion signal which targets the viral preintegration complexto the nucleus after viral penetration (Bukrinsky et al.,1993). In addition, HIV-1 MA was shown to be responsi-ble for the specific incorporation of viral glycoproteinsinto viral particles (Yu et al., 1992; Facke et al., 1993;Wang et al., 1993; Lee and Linial, 1994).

Regions within the central and C-terminal portions ofHIV-1 MA were shown to be important for efficient par-ticle assembly and release (Freed et al., 1994; Wagner etal., 1994; Chazal et al., 1995). In an MuLV system, it wasshown that wild-type MuLV Gag could coassemble witha chimeric Gag containing either heterologous MA orheterologous CA, suggesting that either protein couldmediate Gag–Gag interactions necessary for particleproduction (Deminie and Emerman, 1993). It was re-ported that MA protein of SIV of sooty mangabeys(SIVsmm), in the absence of other viral molecules, wassufficient to direct the release of VLPs (Gonzalez et al.,1993; Liska et al., 1994).

We hypothesized that HIV-1 MA, independent of allother viral proteins, would be capable of directing theprotein–protein interactions necessary for the assemblyand release of VLPs. We expressed HIV-1 Gag-Pol, HIV-1MA alone, or SIV MA alone, in a eukaryotic expressionsystem. Using metabolic labeling of recombinant pro-teins, sucrose density gradients, and a protease protec-tion assay, we found that HIV-1 MA alone was incapableof forming VLPs. Interestingly, employing these same

1 To whom correspondence and reprint requests should be ad-dressed at The University of Alabama at Birmingham, 845 19th StreetSouth; BBRB 220, Birmingham, AL 35294-2170. Fax: (205) 975-6027.E-mail: [email protected].

VIROLOGY 248, 108–116 (1998)ARTICLE NO. VY989284

0042-6822/98 $25.00Copyright © 1998 by Academic PressAll rights of reproduction in any form reserved.

108

methodologies, we confirmed that SIV MA was sufficientfor the assembly and release of VLPs.

RESULTS

Recombinant HIV-1 MA protein was released fromcells and pelleted through a 20% sucrose cushion

To confirm recombinant protein expression, the re-combinant vaccina vectors, V V:gag (HIV-1), rVVmatrix(HIV-1), or rV V/SC11, were inoculated onto H9 cells,and the cultures were metabolically labeled with 3H-leucine overnight. Cell lysate supernatants or lysedviral pellets were immunoprecipitated and separatedby SDS–PAGE. In the cell lysate supernatants from VV:gag (HIV-1) infected cells, we detected the Gag polypro-tein (p55), the p41 processing intermediate, and the ma-ture products, p24 capsid (CA) and p17 matrix (MA)(Fig. 1). In the cell lysate supernatant of cells infectedwith the rVVmatrix (HIV-1) construct, high levels of a17-kDa protein consistent with MA were observed. In thepellets derived from the culture media of VV:gag (HIV-1)infected cells, predominantly CA was detected but MAwas also observed. The pellets from the rVVmatrix(HIV-1) construct contained the 17-kDa immunoreactiveprotein corresponding to MA.

In other experiments, 3H-myristic acid was employedfor metabolic labeling. For VV:gag (HIV-1), the Gagpolyprotein, the p41 intermediate, and the 17-kDa MAprotein were myristylated and, as expected, CA was not(Fig. 2). For rVVmatrix (HIV-1), the 17-kDa MA protein wasmyristylated. These results indicated that VV:gag (HIV-1)and rVVmatrix (HIV-1) expressed posttranslationallymodified HIV-1 Gag or MA proteins with the expectedelectrophoretic mobilities. Furthermore, these recombi-nant HIV-1 Gag or MA proteins were released into mediaand pelleted through a sucrose cushion.

We also wished to analyze and confirm SIV MA aloneassembly using the same methodologies we were ap-plying to HIV-1 MA alone. CV-1 cells were infected withrVVp17 (SIV) and labeled with 3H-leucine to demonstrateSIV MA expression. A 16- to 17-kDa protein was ob-served in both the cell lysate supernatant and in theculture medium (data not shown).

The released HIV-1 MA protein failed to enter asucrose density gradient

To determine whether the MA proteins that were ex-pressed by rVVmatrix (HIV-1) or rVVp17 (SIV) assembledinto VLPs, density gradient centrifugation was employed.H9 cells were infected with rVVmatrix (HIV-1), VV:gag

FIG. 2. Autoradiograph demonstrating myristic acid labeled HIV-1Gag and MA proteins. H9 cells infected with the rV Vs were labeledovernight with 3H-myristic acid. The cells were lysed and immunopre-cipitated followed by SDS–PAGE and autoradiography. Control vac-cinia, C; rV Vmatrix (HIV-1), M; or V V:gag (HIV-1), G.

FIG. 1. Autoradiograph demonstrating leucine-labeled recombinantHIV-1 MA proteins both in cells and pellet fractions. H9 cells wereinfected with the control vaccinia virus (C), rV Vmatrix (HIV-1) (M), orV V:gag (HIV-1) (G). The infected cells were leucine starved and thenlabeled overnight with 3H-leucine. The media were layered over asucrose cushion and pelleted by ultracentrifugation. The pellets or cellswere lysed and immunoprecipitated followed by SDS–PAGE and auto-radiography.

109HIV-1 MATRIX ALONE UNABLE TO FORM PARTICLES

(HIV-1), or control. CV-1 cells were infected with rVVp17(SIV) or VV:gag (HIV-1). After an overnight labeling pe-riod, the media were filtered, layered over a 20–60%sucrose gradient, and centrifuged. For VV:gag (HIV-1)infection of H9 cells, MA and CA migrated into the gra-dient and were localized to fractions 8–12 (Fig. 3a). The

average of the measured densities of these fractionswas 1.15 g/ml, similar to that reported for HIV-1 VLPs(Shioda and Shibuta, 1990). The p41 intermediate wasobserved only in fractions at the top of the gradient. Therelease of this p41 protein in a soluble form into culturemedia may help explain why relatively little particulate

FIG. 3. Autoradiographs indicating that released HIV-1 Gag and SIV MA proteins, but not released HIV-1 MA alone, penetrated into sucrose densitygradients. (A) H9 cells were infected with V V:gag (HIV-1), starved, and labeled with 3H-leucine overnight. The medium was filtered, layered over asucrose density gradient, and ultracentrifuged for 3 h at 125,000 g. (B) H9 cells were infected with rV Vmatrix (HIV-1), starved, and labeled with3H-leucine overnight. The medium was filtered, layered over a sucrose density gradient, and ultracentrifuged for 3 h at 125,000 rpm. (C) CV-1 cellswere infected with rV Vp17 (SIV), starved and labeled with 3H-leucine overnight. The medium was filtered, layered over a sucrose density gradient,and ultracentrifuged for 3 h at 125,000 g.

110 GIDDINGS, RITTER, AND MULLIGAN

Gag protein was observed in fractions 8–12, despite highlevels of Gag expression in the cells (e.g., Fig. 1). Inter-estingly, HIV-1 MA expressed by rVVmatrix (HIV-1), un-like the MA and CA proteins from VV:gag (HIV-1) in H9cells, remained mostly among the soluble proteins at thetop of the gradient in fractions 1–3 (Fig. 3b). A smallamount of HIV-1 MA protein penetrated all the waythrough the gradient into fraction 17, most likely the resultof nonspecific aggregation or MA associated with filter-able cell fragments. Similar results were obtained withVV:gag (HIV-1) or rVVmatrix (HIV-1) infections of CV-1cells (not shown).

MA expressed by rVVp17 (SIV) in CV-1 cells, as ex-pected based on the previous work in CV-1 cells ofGonzalez et al. (1993) migrated into the sucrose gradient(Fig. 3c). The 17-kDa bands localized to fractions 9–11,the average density of which was 1.12 g/ml. Low-densityVLPs have also been reported following expression ofmutant HIV-1 gag genes that lacked the I domain (San-defur et al., 1998). These sucrose gradient results indi-cated that while recombinant HIV-1 Gag or SIV MAformed VLPs, recombinant HIV-1 MA alone did not formVLPs in either H9 or CV-1 cells. However, it remainedpossible that aberrant membrane-bound particlesformed by HIV-1 MA might be of insufficient density topenetrate the gradient or that the protein in fraction 17was due to aggregated particles. It was also possiblethat the apparent low-density VLPs of SIV MA wereprotein aggregates or incompletely enveloped particles.

Expression of HIV-1 Gag, but not HIV-1 MA alone,produced membrane-bound particles which weretrypsin resistant

To further investigate whether the proteins which wereexpressed by rVVmatrix (HIV-1), VV:gag (HIV-1), orrVVp17 (SIV) assembled as membrane-bound particles,

we performed trypsin protection assays. After metaboliclabeling, the culture medium was clarified, filtered, anddivided into six aliquots. The first aliquot remained un-treated, the second received trypsin inhibitor, the thirdreceived triton X-100, the fourth received trypsin, the fifthreceived trypsin and triton X-100, and the sixth receivedtrypsin and trypsin inhibitor. After an incubation period,trypsin inhibitor was added, and the samples were lysed,immunoprecipitated, and analyzed by SDS–PAGE. Tryp-sin inhibitor alone, triton X-100 alone, or untreated sam-ples served as controls. MA produced by VV:gag wasreleased into the media and had an approximate molec-ular mass of 17 kDa: see lanes 1–3 under Gag-Pol (Fig.4). After treatment of the VV:gag media with trypsin (lane4), the MA protein still migrated at 17 kDa, indicating thatit had been protected from trypsin digestion by a lipidbilayer. However, when trypsin and triton x-100 wereadded together, so that lipid bilayers were disrupted bythe detergent, the 17-kDa protein was not observed anda new cleavage product of lower molecular mass wasseen (lane 5). These results indicated that VV:gag re-leased MA proteins as components of membrane boundVLPs. It was also of interest that the Gag processingintermediate, p41, was degraded by trypsin in the ab-sence (lane 4) or presence (lane 5) of triton X-100, con-firming the sucrose gradient data (Fig. 3a), which indi-cated this protein was released in soluble form. This p41proteolysis was specific to the activity of trypsin becausealiquots treated with trypsin inhibitor as well as trypsinmaintained the 41-kDa protein (lane 6). Interestingly, thep24 CA protein, unlike MA, did not undergo proteolysis inthe presence of trypsin and triton X-100 (lane 5). This wasprobably due to poor exposure of the single potentialdibasic trypsin cleavage site present in the CA of HIV-1HXB2.

In contrast, the 17-kDa protein corresponding to HIV-1

FIG. 3—Continued


MA in the culture medium of cells infected with rVVma-trix (HIV-1) was cleaved following treatment with eithertrypsin alone (lane 4, under MA-only) or trypsin plustriton X-100 (lane 5). This HIV-1 MA protein remaineduncleaved in the control aliquots treated with trypsininhibitor alone (lane 3), triton X-100 alone (lane 2), oruntreated (lane 1). The protein remained uncleavedwhen treated with trypsin inhibitor and trypsin (lane 6),demonstrating that the digestion of HIV-1 MA was trypsinspecific.

When this assay was performed on the SIV MA ex-pressed by rVVp17 (SIV), it was found that trypsin did notcleave the MA protein even in the presence of detergent(data not shown). This was probably due to SIV MAprotein folding, causing inaccessibility of the cleavagesites to trypsin or to absence of one dibasic site whichwas present in HIV-1. Therefore, trypsin was not usefulfor assessing the membrane envelopment of the re-leased SIV MA. When the experiment was repeated witha different protease, useful data were obtained (seebelow). The overall results from this set of experimentsprovided strong evidence that the HIV-1 MA protein,expressed in the absence of other HIV-1 proteins, wasunable to assemble and release VLPs.

Expression of SIV MA alone produced membrane-bound particles which were resistant to proteinase K

To obtain definitive data on whether or not SIV MAalone formed VLPs, we employed a different enzyme in

the protease protection assay. Since trypsin was notuseful, proteinase K and the protease inhibitor, phe-nylmethylsulfonylfluoride (PMSF), were utilized. CV-1cells were infected, labeled, and clarified. The filteredmedium was treated as follows: the first aliquot wasuntreated, the second received PMSF, the third re-ceived triton X-100, the fourth received proteinase K,the fifth received proteinase K and triton X-100, and thesixth received proteinase K and PMSF. After the incu-bation period, PMSF was added and the samples werelysed, immunoprecipitated, and analyzed by SDS–PAGE. Medium from V V:gag (HIV-1) infected cells wasused as a positive control and results similar to thosefor trypsin were obtained, except that CA in Aliquot 5was digested by proteinase K, which cleaves peptidebonds with much less specificity than does trypsin(data not shown). SIV MA produced by rV Vp17 (SIV)was released into the media and had an approximatemolecular mass of 16–17 kDa (lanes 1–4 and lane 6,Fig. 5). After treatment with proteinase K (lane 4), theSIV MA protein of 16–17 kDa was still present, indi-cating that it was protected from proteinase K diges-tion by a lipid bilayer. However, when proteinase K andtriton X-100 were added together, disrupting the lipidlayer bilayer, the 16–17 kDa protein disappeared (lane5). These results indicated that rV Vp17 (SIV)-infectedcells released SIV MA proteins as components ofmembrane bound particles, thereby confirming thatSIV MA can form VLPs.

FIG. 4. Trypsin protection assays indicating HIV-1 Gag, but not HIV-1 MA alone, assembled into lipid bound particles. H9 cells infected with rV Vswere labeled with 3H-leucine overnight. The medium were filtered and split into six aliquots. Reagents were added to each aliquot as follows: Aliquot1, untreated; Aliquot 2, trypsin inhibitor at a 500 mg/ml final concentration; Aliquot 3, triton x-100 to 1%; Aliquot 4, trypsin at a 200 mg/ml finalconcentration; Aliquot 5, triton x-100 to 1% and 200 mg/ml trypsin; and Aliquot 6, 200 mg/ml trypsin and 500 mg/ml trypsin inhibitor. After 30 min, trypsininhibitor was added to the tubes that had not received it to stop reactions. The media fractions were immunoprecipitated followed by SDS–PAGE andautoradiography.


DISCUSSION

As demonstrated above by metabolic labeling, densitygradient centrifugation, and protease protection, recom-binant expression of HIV-1 gag-pol resulted in Gag syn-thesis with posttranslational modification, polyproteinprocessing, and production of membrane bound VLPs ofappropriate density. Similarly, recombinant expression ofHIV-1 MA alone in the absence of other HIV-1 proteinsresulted in a 17-kDa immunoreactive protein that wasreleased into the culture medium, pelleted through a 20%sucrose cushion, and was posttranslationally modifiedwith myristic acid. These results initially suggested thatthe HIV-1 MA alone might be sufficient for the assemblyand release of VLPs, as had previously been describedfor SIVsmm (Gonzalez et al., 1993; Liska et al., 1994).However, when the culture medium from cells express-ing HIV-1 MA alone was layered over a sucrose densitygradient, most of the MA protein failed to enter into thegradient. These data were consistent with two interpre-tations: (1) the released HIV-1 MA protein was in solubleform and so failed to penetrate the gradient or (2) theHIV-1 MA protein was released within aberrant mem-brane bound VLPs possessing insufficient density toenter the gradient. The small amount of HIV-1 MA whichpelleted completely through the 20–60% density gradient

was likely an artifact; i.e., an aggregation of soluble MAprotein or VLPs or MA protein associated with lysedcellular fragments.

To permit direct comparison of our HIV-1 MA resultswith SIV MA, recombinant expression of SIV MA alonewas also performed utilizing these systems, and resultedin an immunoreactive protein of 16–17 kDa that wasreleased into the culture medium and pelleted through asucrose cushion. The density gradient analysis of culturemedium obtained from cells expressing SIV MA alonerevealed that this 16–17 kDa immunoreactive proteinmigrated into the gradient to a density of 1.12 g/ml.

To determine which of the possible interpretationscited above for the HIV-1 MA data were correct, a trypsinprotection assay was utilized. In this experiment, MA andCA released into the medium of cells expressing thecomplete HIV-1 gag and pol genes were protectedagainst trypsin digestion, indicating Gag has assembledinto particles surrounded by a lipid bilayer. In contrast,the HIV-1 MA protein expressed alone and released intocell culture medium was susceptible to trypsin digestioneven in the absence of detergent and therefore was notpresent in membrane bound particles. Taken togetherwith the density gradient results, these trypsin protectiondata clearly indicated that HIV-1 MA protein alone wasincapable of assembling into VLPs.

This conclusion for HIV-1 MA differed from two previ-ous reports which also employed vaccinia expressionbut indicated that SIVsmm MA was sufficient for the as-sembly and release of VLPs (Gonzalez et al., 1993; Liskaet al., 1994). The SIV MA density gradient data that weobtained appeared to confirm those results. To be cer-tain, the trypsin protection assay was employed, but itwas discovered that trypsin did not cleave the SIV MAparticles even in the presence of detergent. Therefore,proteinase K was utilized in the protection assay, and itwas observed that the SIV MA VLPs were protected fromproteolysis unless detergent was added. Thus SIV MAalone was capable of forming low density VLPs in theabsence of other viral proteins.

Why was HIV-1 MA incapable of assembly into VLPswhile SIV MA was capable of forming particles? Onepossibility is that the assembly requirements for HIV-1MA differ from those for SIV MA. Several residues thatwere shown to be important for the proper folding ofSIVmac MA are altered in HIV-1 (Rao et al., 1995). Theseinclude the glycine-45 and alanine-47 of SIV MA whichmay be involved in the hydrogen bonding responsiblefor trimerization of MA. In HIV-1 MA, these residuesare alanine-45 and asparagine-47, respectively. HIV-1MA also has one less cysteine and methionine thanSIV MA, which might also affect its conformation andmake it less likely to form tight self-associations. TheHIV-1 and SIVsmm MA proteins also have very littlesequence similarity in their C-terminal regions, and

FIG. 5. Proteinase K protection assay indicating SIV MA assembledinto lipid bound particles. CV-1 cells infected with rV Vp17 (SIV) werelabeled with 3H-leucine for 4 h, fresh medium was added, and labellingwas continued overnight. The medium was filtered and split into sixaliquots. Reagents were added to each aliquot as follows: Aliquot 1,untreated; Aliquot 2, PMSF at a 2.5 mg/ml final concentration; Aliquot 3,triton X-100 to 1%; Aliquot 4, proteinase K at 10 mg/ml final concentra-tion; Aliquot 5, triton X-100 to 1% and 10 mg/ml proteinase K; Aliquot 6,10 mg/ml proteinase K and 2.5 mg/ml PMSF. After 1 h at 37°C, PMSFwas added to the tubes that had not received it to stop reactions. Themedia fractions were immunoprecipitated followed by SDS–PAGE andautoradiography.


this region could play a role in the self assembly ofthese proteins.

Interestingly, it was shown by deleting the entire MAcoding sequence that MA may be dispensable forparticle formation by HIV-1 Gag (Franke et al., 1994;Lee and Linial, 1994). However, smaller mutationswithin HIV-1 MA disrupted particle formation (Freed etal., 1994; Wagner et al., 1994; Chazal et al., 1995). Smallmutations in SIV MA which could possibly change itsconformation were also shown to abolish particle for-mation in the context of Gag (Gonzalez et al., 1996). Itappears that the contributions of MA to particle as-sembly depend on MA conformation. SIVsmm MA mayassemble without the aid of the other Gag componentsbecause its conformation permits more efficient self-association than occurs with HIV-1 MA. Despite thefact that in some systems the Gag proteins of HIV-1and SIV may bind together (Franke et al., 1994), therequirements for their MA association and assemblyappear to be different. Our results with HIV-1 and SIVclearly indicate that self-assembly of MA into VLPs isnot a general feature of the primate lentiviruses.

MATERIALS AND METHODS

Cells and viruses

The human T-cell lymphoma cell line, H9, was main-tained in RPMI supplemented with 12% fetal bovineserum (FBS). The simian cell line, CV-1, was main-tained in DMEM supplemented with 12% FBS. Therecombinant vaccinia virus (rV V) constructs utilizedwere V V:gag (provided by the NIH AIDS repository,catalog No. 405, contributed by Edgar Engleman), rV-Vmatrix (HIV-1) (provided by Connie Bergman, USC),rV Vp17 (SIVsmm) (provided by Silvia Gonzalez; desig-nated v-p17gag in Gonzalez et al., 1993), and rVV/SC11,a control recombinant which expresses the lacZ gene(Owens and Compans,1989). The cloning strategy forV V:gag was described (Gowda et al., 1989). This con-struct expresses the Gag and Pol proteins of HIV-1HXB2. rV Vmatrix (HIV-1) was created using PCR am-plification from the cDNA for the HXB2 strain of HIV-1using the following primers: 59-59-CTCTGCGCTAGC-CGCCATGGGTGCGAGAGCGTC; and 39-59-CGGGGCC-TCGAGACTAGTTAATTAGTAATTTTGGCTGAC. This strat-egy results in a stop codon being placed directly afterthe codon for the last amino acid of MA, the p17/p24cleavage site. The PCR products were cut with NcoI/SpeI and inserted into the pK vector, a previouslydescribed derivative of the pSC11 plasmid (Bergmanet al., 1993). The recombinant virus was constructedthrough homologous recombination using the WRstrain of vaccinia (Connie Bergman, personal commu-nication). This construct expresses HIV-1 MA in theabsence of other HIV-1 proteins. The cloning strategyfor the rV Vp17(SIV) has been described (Gonzalez et

al., 1993). This construct expresses SIVsmm MA in theabsence of other SIV viral proteins. Stocks of thevaccinia viruses were grown in either TK- or BHK-21cell lines maintained in DMEM supplemented with12% FBS. These stocks were titered in CV-1 cells.

Protein expression, labeling, and immunoprecipitation

H9 cells were infected with the HIV-1 or control rV Vsin serum-free RPMI at a multiplicity of infection (m.o.i.)of 8 plaque-forming units (pfu)/cell. After 2 h the mediawas removed after a low-speed centrifugation and thecells were incubated for 30 min in leucine-free Eagle’smedium. Cultures were then labeled for 4 h with 200mCi/ml 3H-leucine in Eagle’s medium. Fresh leucine-free media and RPMI supplemented with 12% FBS(complete media) were added at a ratio of 3:1, andlabeling was continued overnight. The final concentra-tion of complete media was 20%. CV-1 cells wereinfected with HIV-1 Gag, HIV-1 MA, SIV MA, or controlrV Vs in serum free DMEM at a m.o.i. of 8 and treatedas above. All incubations were done at 37°C. Alterna-tively, H9 cells in RPMI were infected with the HIV-1 orcontrol rV Vs at an m.o.i. of 5 for 2 h, then washed andlabeled overnight with 3H-myristic acid. The mediafrom these cultures was then layered over a 20%sucrose cushion and centrifuged in an ultracentrifuge(Beckman Instruments, Inc., Palo Alto, CA) using anSW41 ti rotor (Beckman Instruments) at 125,000 g for3 h to pellet virus-like particles. The pellets wereresuspended in lysis buffer (1% triton X-100, 1% so-dium deoxycholate, 150 mM NaCl, 50 mM Tris-HCl pH7.5, 10 mM EDTA). The cells were subjected to low-speed centrifugation and lysed with lysis buffer. HIV-1or SIV proteins in media, pellets, or cell lysate super-natants were immunoprecipitated overnight at 4°Cwith rotation using protein A-agarose beads andpooled sera from HIV-1 infected patients or SIV in-fected monkeys. The beads were collected by low-speed centrifugation, washed twice with lysis buffer,and boiled for 10 min in reducing buffer (2% SDS, 10%glycerol, 5% 2-mercaptoethanol, 50 mM Tris-HCl pH6.8, 0.1% bromphenol blue). The proteins were sepa-rated on 11.5 or 15% SDS–PAGE gels and visualized byautoradiography.

Density gradient centrifugation

Culture media from infected H9 or CV-1 cells wereclarified by low-speed centrifugation for 5 min and fil-tered through 0.45-mm filters. The clarified media werelayered over 20–60% linear sucrose gradients and cen-trifuged at 125,000 g for 3 h using an SW41 ti rotor.Gradient fractions were collected from the top and di-luted with lysis buffer. Density measurements with arefractometer (Bausch & Lomb) were taken before dilu-


tion. The samples were then immunoprecipitated andanalyzed as above.

Protease protection assay

Cultures of 5 3 106 H9 or CV1 cells were infectedwith V V:gag (HIV-1), rV Vmatrix (HIV-1), rV Vp17 (SIV), orthe control rV V at an m.o.i. of 8 and labeled as above.After the overnight labeling period, the cells werecollected and treated as above with lysis buffer. Theculture media were filtered, and each was divided intosix equal aliquots and treated following a methodpreviously described (Wills et al., 1989). The first ali-quot received no further treatment. The second aliquotreceived soybean trypsin inhibitor (Sigma, Inc., St.Louis, MO) to a final concentration of 500 mg/ml or 2.5mg/ml PMSF (Sigma). The third aliquot received tritonx-100 to 1%. The fourth aliquot received trypsin to afinal concentration of 200 mg/ml or proteinase K to afinal concentration of 10 mg/ml. The fifth aliquot re-ceived both triton X-100 to 1% and trypsin (200 mg/ml)or proteinase K (10 mg/ml). The sixth aliquot receivedtrypsin (200 mg/ml) and an excess of trypsin inhibitor(500 mg/ml) or proteinase K (10 mg/ml) and an excessof PMSF (2.5 mg/ml). The trypsin samples were thenincubated at room temperature for 30 min, after whichtrypsin inhibitor was added to all samples that did notpreviously receive it. The proteinase K samples wereincubated at 37°C for 1 h, after which PMSF wasadded to all samples that did not previously receive it.Lysis buffer or 5x was added to all samples, followedby immunoprecipitation, SDS–PAGE, and autoradiog-raphy as above.

ACKNOWLEDGMENTS

This work was supported by U.S. Public Health Service GrantsAI33784, AI28147, and AI45209 (G.D.R. and M.J.M.), and AI07150(A.M.G.). We thank Cathy Hardin and Alesia Hatten for typing themanuscript, and Paul Johnson, Connie Bergman, Silvia Gonzalez, Pa-tricia Fultz,and the NIH AIDS Repository for providing reagents.

REFERENCES

Bergman, C., Stohlmann, S. A., and McMillan, M. (1993). An endog-enously synthesized decamer peptide efficiently primes cytotoxic Tcells specific for the HIV-1 envelope glycoprotein. Eur. J. Immunol. 23,2777–2781.

Bryant, M., and Ratner, L. (1990). Myristoylation-dependent replicationand assembly of Human Immunodeficiency Virus 1. Proc. Natl. Acad.Sci. USA 87, 523–527.

Bukrinsky, M. I., Haggerty, S., Dempsey, M. P., Sharova, N., Adzhubel, A.,Spitz, L., Lewis, P., Goldfarb, D., Emerman, M., and Stevenson, M.(1993). A nuclear localization signal within HIV-1 matrix protein thatgoverns infection of non-dividing cells. Nature 365, 666–669.

Chazal, N., Gay, B., Carriere, C., Tournier, J., and Boulanger, P. (1995).Human Immunodeficiency Virus Type 1 MA deletion mutants ex-pressed in baculovirus-infected cells: cis and trans effects on theGag precursor assembly pathway. J. Virol. 69, 365–375.

Delchambre, M., Gheysen, D., Thines, D., Thiriart, C., Jacobs, E., Verdin,E., Horth, M., Burny, A., and Bex, F. (1989). The Gag precursor of

simian immunodeficiency virus assembles into virus-like particles.EMBO J. 8, 2653–2660.

Deminie, C. A., and Emerman, M. (1993). Incorporation of HumanImmunodeficiency Virus Type 1 gag proteins into murine leukemiavirus virions. J. Virol. 67, 6499–6506.

Facke, M., Janetzko, A., Shoeman, R. L., and Krausslich, H.-G. (1993).A large deletion in the matrix domain of the Human Immunodefi-ciency Virus gag gene redirects virus particle assembly from theplasma membrane to the endoplasmic reticulum. J. Virol. 67,4972–4980.

Franke, E. T., Yuan, H. E. H., Bossolt, K. L., Goff, S. P., and Luban, J.(1994). Specificity and sequence requirements for interactions be-tween various retroviral gag proteins. J. Virol. 68, 5300–5305.

Freed, E. O., Orenstein, J. M., Buckler-White, A. J., and Martin, M. A.(1994). Single amino acid changes in the Human ImmunodeficiencyVirus Type 1 Matrix protein block virus particle production. J. Virol. 68,5311–5320.

Gallina, A., Mantoan, G., Rindi, G., and Milanesi, G. (1994). Influence ofMA internal sequences, but not of the myristylated N-terminus se-quence, on the budding site of HIV-1 gag protein. Biochem. Biophys.Res. Commun. 204, 1031–1038.

Gheysen, D., Jacobs, E., de Foresta, F., Thiriart, C., Francotte, M.,Thines, D., and De Wilde, M. (1989). Assembly and release of HIV-1precursor Pr55gag virus-like particles from recombinant baculovirus-infected insect cells. Cell 59, 103–112.

Gonzalez, S. A., Affranchino, J. L., Gelderblom, H. R., and Burny, A.(1993). Assembly of the matrix protein of Simian ImmunodeficiencyVirus into virus-like particles. Virology 194, 548–556.

Gonzalez, S. A., Burny, A., and Affranchino, J. L. (1996). Identification ofdomains in the Simian Immunodeficiency Virus matrix protein essen-tial for assembly and envelope glycoprotein incorporation. J. Virol. 70,6384–6389.

Gowda, S. D., Stein, B. S., Steimer, K. S., and Engleman, E. G. (1989).Expression and processing of Human Immunodeficiency Virus Type1 gag and pol genes by cells infected with a recombinant vacciniavirus. J.Virol. 63, 1451–1454.

Henderson, L. E., Benveniste, R. E., Sowder, R., Copeland, T. D., Schultz,A. M., and Oroszlan, S. (1988). Molecular characterization of gagproteins from Simian Immunodeficiency Virus (SIVmne). J.Virol. 62,2587–2595.

Lee, P.P, and Linial, M. L. (1994). Efficient particle formation can occurif the matrix domain of Human Immunodeficiency Virus Type 1 gag issubstituted by a myristylation signal. J. Virol. 68, 6644–6654.

Liska, V., Spehner, D., Mehtali, M., Schmitt, D., Kirn, A., and Aubertin,A. M. (1994). Localization of viral protein X in Simian Immunodefi-ciency Virus macaque strain and analysis of its packaging require-ments. J. Gen. Virol. 75, 2955–2962.

Mervis, R. J., Ahmad, N., Lillehog, E. P., Raum, M. G., Salazar, R. H. R.,Chan, H. G., and Venkatesan, S. (1988). The gag gene products ofHuman Immunodeficiency Virus Type 1. J. Virol. 62, 3993–4002.

Neidrig, M., Gelderblom, H. R., Pauli, G., Marz, J., Bickhard, H., Wolf, H.,and Modrow, S. (1994). Inhibition of infectious Human Immunodefi-ciency Virus Type 1 particle formation by gag protein-derived pep-tides. J. Gen. Virol. 75, 1469–1474.

Owens, R. J., and Compans, R. (1989). Expression of the Human Immu-nodeficiency Virus envelope glycoprotein is restricted to basolateralsurfaces of polarized epithelial cells. J. Virol. 63, 978–982.

Rao, Z., Belyaev, A. S., Fry, E., Roy, P., Jones, I. M., and Stuart, D. I. (1995).Crystal structure of SIV matrix antigen and implications for virusassembly. Nature 378, 743–747.

Sandefur, S., Varthakavi, V., and Spearman, P. (1998). The I domain isrequired for efficient plasma membrane binding of Human Immuno-deficiency Virus Type 1 Pr55Gag. J. Virol 72, 2723–2732.

Shioda, T., and Shibuta, H. (1990). Production of Human Immunode-ficiency Virus (HIV)-like particles from cells infected with recom-binant vaccinia viruses carrying the gag gene of HIV. Virology 175,139–148.


Spearman, P., Wang, J.-J., Heyden, N. V., and Ratner, L. (1994). Identifi-cation of Human Immunodeficiency Virus Type 1 gag protein do-mains essential to membrane binding and particle assembly. J. Virol.68, 3232–3242.

Trono, D., Feinberg, M. B., and Baltimore, D. (1989). HIV-1 gag mutantscan dominantly interfere with the replication of the wild-type virus.Cell 59, 113–120.

Wagner, R., Deml, L., Fliebbach, H., Wanner, G., and Wolf, H. (1994).Assembly and extracellular release of chimeric HIV-1 Pr55gag retro-virus-like particles. Virology 200, 162–175.

Wang, C.-T., Zhang, Y., McDermott, J., and Barklis, E. (1993). Conditionalinfectivity of a Human Immunodeficiency Virus matrix domain dele-tion mutant. J. Virol. 67, 7067–7076.

Wills, J. W., and Craven, R. C. (1991). Form, function, and use of retroviralgag proteins. AIDS 5, 639–654.

Wills, J. W., Craven, R. C., and Achacoso, J. A. (1989). Creation andexpression of myristylated forms of Rous Sarcoma Virus gag proteinin mammalian cells. J. Virol. 63, 4331–4343.

Yu, X., Yuan, X., Matsuda, Z., Lee, T.-H., and Essex, M. (1992). The matrixprotein of Human Immunodeficiency Virus Type 1 is required forincorporation of viral envelope protein into mature virions. J. Virol. 66,4966–4971.

Zhou, W., Parent, L. J., Wills, J. W., and Resh, M. D. (1994). Identificationof a membrane-binding domain within the amino-terminal region ofHuman Immunodeficiency Virus Type 1 gag protein which interactswith acidic phospholipids. J. Virol. 68, 2556–2569.


the matrix protein of hiv-1 is not sufficient for assembly ... · cv-1 cells were infected with...

Documents