Proc. Nadl. Acad. Sci. USA Vol. 87, pp. 523-527, January 1990 Biochemistry Myristoylation-dependent replication and assembly of human immunodeficiency 1 (AIDS/virus assembly/gag precursor) MARTIN BRYANT*t AND LEE RATNERtf§ Departments of *Pediatrics, tMedicine, and §Molecular Microbiology, Washington University School of Medicine, Saint Louis, MO 63110 Communicated by Stuart Kornfeld, September 25, 1989 (receivedfor review May 23, 1989)

ABSTRACT Covalent linkage of to the N- virus types I and II, human immunodeficiency virus 1 (HIV-1), terminal residue ofPr55gag, the precursor ofthe major and simian immunodeficiency virus type I] has recently been structural proteins of human immunodeficiency virus 1 (HIV- described (1, 18-23). In cells infected by MoMLV (22) or 1), facilitates an essential step in virus assembly and propaga- MPMV (23), myristoylation of the gag polypeptide is required tion. Substitution of the myristoyl-acceptor glycine with ala- for normal virion assembly and infectious particle production. nine, in a functional clone of HIV-1, eliminates virus replica- HIV, like other mammalian , encodes a gag tion. Complementation of this defect, in trans, restores precursor, Pr55gag, which is normally processed by the viral infectious particle production. The nonmyristoylated (myr-) protease to the major structural proteins found in the mature gag precursor accumulates in infected cells and is not processed extracellular virus particle. Both Pr55gag and the matrix into the mature capsid components of the intact virion. How- protein (p17), which is proteolytically cleaved from the ever, myr- Pr55gag can be processed by purified HIV protease N-terminal end ofPr55gag, are myristoylated (see Fig. 1) (24, in vitro, demonstrating that the myristoyl moiety is not required 25). Analysis of the sequence of these proteins for cleavage by the protease. Myristoylation of Pr55gag is not from different HIV-1 isolates shows that the myristoylation necessary for localization but is required for stable membrane acceptor is always an N-terminal glycine residue (26-28). association and assembly of HIV-1. Furthermore, the Prl80gag-pol precursor, which contains gag proteins as well as the viral replicative [prote- The covalent attachment of long-chain fatty acids to select ase, reverse transcriptase (RT), and integrase], is probably subsets of eukaryotic cellular proteins and viral polypeptides also myristoylated. The current study examines the role of confers a change in hydrophobicity that may be important in myristoylation of the structural gag polyprotein precursor defining protein quaternary structure, directing intracellular Pr55gag of HIV-1 in virus replication. Our results show that trafficking and , and/or in specific pro- addition of the myristate moiety is required for stable mem- tein-protein or protein-lipid interactions at the plasma mem- brane association, proteolytic processing, and assembly of brane. The specific fatty acid and its distinct mode of Pr55gag into infectious extracellular particles of HIV-1. attachment probably influence the unique biochemical func- tion served by each acylated protein (1-5). MATERIALS AND METHODS The most thoroughly studied acyl proteins have been Cell Lines. T-lymphoid cell lines H9 and Jurkat were virus-related polypeptides because of their high level of provided by J. Hoxie and R. C. Gallo (National Cancer expression in infected cells. Schmidt and Schlesinger (6) were Institute). COS-1 and HeLa cell lines were obtained from the first to describe the attachment of fatty acid to the envelope American Type Culture Collection. glycoproteins of vesicular stomatitis and Sindbis and DNA Clones. pGG1 contains a functional clone of HIV-1 to suggest an association of this type of protein modification and was previously designated pHXB2gpt2 (29). Plasmid with virus assembly and propagation (7). Subsequently, the pAenv is a derivative of pGG1 constructed by digestion with structural proteins of nonenveloped viruses have also been the restriction endonucleases Sal I and Xho I, followed by shown to contain covalently bound fatty acid. The capsid treatment with T4 DNA ligase thus deleting the entire enve- protein VP4 and its precursors in picornaviruses are myris- lope gene of HIV-1 (nucleotides 5366-8474). toylated (8-11). The position of the myristate moiety at the Mutagenesis. The glycine to alanine substitution was made interface between protein subunits in poliovirus suggests that with the 20-mer oligonucleotide GCTCTCGCAGCCATCTC- it may play a role in virus capsid assembly (12). The structural TCT using oligonucleotide-directed mutagenesis as de- proteins ofpolyomavirus (10) and simian virus 40 (11), as well scribed by Kunkel (30). The resultant plasmid was designated as the PreS1 protein ofhepadnaviruses (i.e., hepatitis B virus) pGA1 and the mutation was confirmed by the dideoxynucle- (13), are also modified by attachment of myristic acid and otide chain-termination method for sequencing DNA (31). may influence nucleocapsid assembly and/or infectivity. Transfection. Cells were transfected by the calcium phos- Myristoylation of p6Ov-src of Rous sarcoma virus is re- phate precipitation technique for gene transfer (32). Stable quired for its stable association with cellular membranes and transfectants were isolated by cotransfection of HeLa cells morphological transformation of cells (14, 15). Similarly, with either pGA1 or pGG1 and SFneo [similar to pSV2neo myristoylation has been proposed to be an important factor in with the addition of the spleen focus-forming virus long transformation by gag-onc fusion proteins with kinase activity terminal repeat (33)]. Cell clones resistant to G418 (400 (16, 17). In addition, fatty acid modification of the gag pre- ,ug/ml, GIBCO) and positive by HIV-1 p24 ELISA (DuPont) cursor to the internal structural proteins of many mammalian were selected for protein studies. retroviruses [Moloney murine leukemia virus (MoMLV), Ma- son-Pfizer monkey virus (MPMV), human T-cell leukemia Abbreviations: MoMLV, Moloney murine leukemia virus; MPMV, Mason-Pfizer monkey virus; HIV, human immunodeficiency virus; RT, reverse transcriptase. The publication costs of this article were defrayed in part by page charge tTo whom reprint requests should be addressed at: Washington payment. This article must therefore be hereby marked "advertisement" University School of Medicine, Clinical Science Building Box 8125, in accordance with 18 U.S.C. §1734 solely to indicate this fact. 660 South Euclid Avenue, Saint Louis, MO 63110. 523 Downloaded by guest on September 26, 2021 524 Biochemistry: Bryant and Ratner Proc. Natl. Acad. Sci. USA 87 (1990) Virus Assays. Cell supernatants were tested on various Biological Effect of Myristoylation Minus Mutation. The days posttransfection for virus-specific antigen using the p24 pGA1 mutant and the parent plasmid pGG1 were transfected ELISA (sensitivity of 0.03 ng/ml). Both the major structural into CD4- HeLa (Fig. 2) or COS-1 cells (not shown) using nucleocapsid protein of HIV-1, p24, and the p24-containing calcium phosphate coprecipitation (32). On day 4 posttrans- Pr55gag can be detected by this assay (unpublished obser- fection, CD4' human lymphoid cells (H9) were added to vation). Virions were concentrated 10:1 from culture super- allow propagation and amplification of infectious virus re- natants by polyethylene glycol precipitation (30% in 150 mM leased from the transfected cells. The nonadherent H9 cells NaCl/0.1 mM phenylmethylsulfonyl fluoride), solubilized, were removed from the adherent HeLa cells on day 8 and and virus-associated RT activity was measured by [32P]TTP were maintained in culture for an additional 2 weeks. Virus- incorporation using a poly(rA)-oligo(dT) template (34). specific antigen production, as measured by p24 antigen Immunoprecipitation and Immunoblotting. HIV-1-specific ELISA, was detected in supernatant solutions from both antigens in lysates from 4 x 106 cells or cell-free supernatants pGG1 and pGA1 transfected cultures prior to the addition of were characterized by immunoblot techniques (35) or were the H9 cells (Fig. 2). However, neither virus antigen nor RT immunoprecipitated after labeling with [35S]/ activity could be detected in the H9 cultures after cocultiva- (specific activity, 1031 Ci/mmol; 1 Ci = 37 GBq) or tion with pGA1 transfected HeLa (Fig. 2) or COS-1 cells (not [3H]myristic acid (39.3 Ci/mmol). Cells labeled with myris- shown). In contrast, production of infectious HIV by the tate were grown in 5% delipidated fetal calf serum. Pooled pGG1 transfected cells was readily identified by syncytia AIDS patients' sera containing a high titer of antibody to HIV formation among the H9 cells, p24 soluble antigen produc- antigens were used in both procedures. Cell pellets were tion, and RT activity (Fig. 2). The pGA1 defect could, how- lysed in RIPA buffer [50 mM Tris-HCI, pH 7.4/0.5% Triton ever, be complemented in trans, by the gag-pol expression X-100/100 mM NaCl/aprotinin (100 units/ml)/0.1 mM phe- plasmid pAenv (Fig. 2). Similar results were obtained after nylmethylsulfonyl fluoride] and used directly for immunoblot direct transfection of the CD4+ Jurkat cell line (data not analysis. The HIV-1 and cellular proteins were separated by shown). These results show that the glycine to alanine sub- SDS/PAGE (36) and transferred by electroblotting to a stitution at the N terminus of the gag polyprotein Pr55gag nitrocellulose membrane. HIV-specific proteins were iden- eliminates propagation of infectious HIV-1. To determine tified by a series of reactions with goat anti-human IgG whether noninfectious particles were assembled and released conjugated with biotin and avidin conjugated with horserad- from pGA1 transfected cells, culture supernatant was concen- ish peroxidase (HRP) and the HRP substrate 4-chloro- trated and then fractionated by buoyant density centrifugation 1-naphthol. For immunoprecipitations, postnuclear superna- through a linear sucrose gradient (Fig. 2 Insets). Neither virus tant containing -5 x 106 cpm or cell-free supernatant was capsid antigen, as measured by p24 ELISA, nor [3H]uridine- precleared with 5 ,ul of normal human sera and protein labeled viral RNA could be detected at the density of HIV-1 A-Sepharose (Pharmacia) and then incubated with 5 ,u of particles from pGA1 transfected cells. Soluble gag antigen, HIV-1 antibody-positive serum for 8 hr at 4°C, and an measured as p24-equivalent material by ELISA, was present additional 2 hr with 40 ,ul of protein A-Sepharose (50% in in the concentrated supernatant solution at the top of the RIPA buffer). The beads were washed four times with RIPA sucrose gradient. In contrast, intact virions with a density of buffer and once with 10 mM Tris HCI, pH 7.4/0.1 M NaCI. 1.14-1.16 g/ml were present in the culture supernatant solu- The pellet was resuspended in 50 ,lI of 2x SDS/PAGE tion of pGG1 transfected cells. Immunoblot analysis of the sample buffer (10 mM Tris-HCI, pH 8.0/2% SDS/2% 2- gradient fractions yielded identical results and demonstrated mercaptoethanol) and boiled 5 min, and the eluted proteins nonvirion-associated, p24-containing material in the medium were separated by SDS/PAGE. The gel was fixed in isopro- from pGA1 transfected cells (data not shown). panol (25%), acetic acid (10%), treated with En3Hance (New Synthesis of Pr55gag in Mutant Transfected Cells. To con- England Nuclear/DuPont), dried, and exposed to preflashed firm that Pr55gag is produced but not myristoylated in the x-ray film (X-Omat AR, Kodak). rev Subcellular Fractionation. Cells were rinsed twice in phos- phate-buffered saline (PBS), scraped, pelleted by centrifu- gag j ElifE HIV-1 UE pol I Pe ED gation at 250 x g for 10 min, and resuspended in a solution LTR LTR containing 1 mM MgCI2, 10 mM Tris HCI (pH 7.4), 1 mM 5 H HH H H H 3' EDTA. Cells were then broken with 45 strokes of a Dounce pGG1 b homogenizer with a Teflon pestle and the nuclei and unbro- ken cells were removed by centrifugation for 10 min at 1000 x g after adjusting the final salt concentration to 0.15 M NaCI. Bs/H 11 Cla The soluble cytosol fraction (S-100) was separated from the 256 ATGGGTGCG 374 membrane fraction (P-100) by centrifugation for 30 min at 4°C Gly p24 p15 in a type 50 Ti rotor at 45,000 rpm. The pellet (P-100) was BssH 11 4 Cla Um resuspended in STE (0.15 M NaCI/10 mM Tris HCI, pH 7.4/1 pGA1 ----- J-ATGGCTGCGi .11 mM EDTA) and afraction was saved for immunoblot analysis. The remaining sample was adjusted to 1 M NaCl and centri- Ala fuged again at 45,000 rpm. The two new fractions were FIG. 1. Mutagenesis of the N-terminal glycine of Pr55gag, the designated salt wash (high ionic strength) P-100(+) and S- polyprotein precursor of the structural proteins of HIV-1. The 100(+). The original fractions were designated P-100(-) and positions of HindIII restriction sites (H) are shown for the S-100(-) to indicate no change in ionic strength of the buffer. functional clone, pGG1. The 551-kilobase HindIII fragment (nucle- otides 80-631) derived from plasmid pGG1 was subcloned into M13, RESULTS and site-directed mutagenesis (30) was used to change the codon for Mutagenesis. To investigate the role of myristoylation of the N-terminal glycine to a codon for alanine. The 116-kilobase BssHII/CIa I fragment was isolated from the replicative form of the Pr55gag in HIV assembly and replication, we used oligonu- mutated M13 clone and ligated in place of the corresponding frag- cleotide-directed mutagenesis to change the codon for the ment in pGG1. The resultant mutant plasmid was designated pGA1. myristoyl-acceptor glycine of Pr55gag to a codon for the The HIV-1 precursor, Pr55gag, and processed gag gene products, nonacceptor alanine in pGG1, a functional clone of HIV-1 matrix antigen (p17), capsid antigen (p24) are shown. M, myristoyl (Fig. 1). The mutant clone was designated pGA1. group at the N terminus of Pr55gag and p17. Downloaded by guest on September 26, 2021 Biochemistry: Bryant and Ratner Proc. Natl. Acad. Sci. USA 87 (1990) 525 those expressed in the pGG1 transfected cells (Fig. 3). The z >0o3 wild-type Pr55gag of HIV-1 was labeled with [3H]myristate (Fig. 3B, lanes 1 and 3) as has been shown (24, 25). In contrast, the Pr55gag produced in pGA1 transfected cells could not be labeled with [3H]myristate (Fig. 3B, lane 2) but 0.1 could be detected by immunoblot using either antibody- positive AIDS patients' sera (Fig. 3A, lane 2) or a mouse anti-p24 monoclonal antibody (not shown). Processing of the gag precursor Pr55gag to the virus capsid antigen p24 was detected in the pGG1 transfected cells (Fig. 3A, lanes 1 and > 50 3) but not in the pGA1 transfected cells (Fig. 3A, lane 2). < ~0 1. 2 8474),aec on pGA1p Nonmyristoylated Pr55gag Is Not Processed in Vivo. We further investigated the fate of the nonmyristoylated HIV-1 gag precursor (myr-) in the pGA1 transfected cells in a CL pGGI I pulse-chase experiment. Confluent monolayers of the trans- TIME3 (day pottasfcn fected HeLa cells were metabolically labeled with [35S]_ methionine/cysteine for 8 hr. A fraction of the labeled cells cotolpasi p 0n(n minus mutantlcigncloie 36 was refed with complete medium and harvested for immu- 844,n th combntonIofpGA&nvwretasfceditlu noprecipitation 0 and 24 hr later (Fig. 4A). Cell-free super- natant solution was also harvested at the end of the chase period. Virus-specific proteins were immunoprecipitated, separated by SDS/PAGE, and detected by autoradiography. In the pGG1 transfected cells, processed p24 can be seen immediately after the 8-hr pulse (no chase) (Fig. 4A, pGG1, lane a). After the 24-hr chase period (lane b), nearly complete CD4- HeLa cells on day 0. The CD4+ human lymphoid cells, H9, capable of exogenous by HIV-1 were added on day 4, disappearance of cell-associated precursor coincided with removed from the adherent cells on day 8, and maintained in culture the appearance of the virion-associated major structural for an additional 2 weeks. Similar results were observed when protein p24 (lane c). In contrast, only minimal turnover of transfected COS-1 cells were used (not shown). (Upper) Soluble p24 Pr55gag, and no processed p24 in the pGA1 transfected cells antigen production in the transfected cultures as determined by was seen at the 24-hr time point (Fig. 4A, pGA1, lane b) or HIV-1 p24 ELISA (DuPont). (Lower) RT activity present in the same at 48 hr (not shown). These data indicate that myristoylation samples. (Inset) Virus particle production as determined by sucrose is required for normal processing of the 55-kDa structural density gradient fractionation and soluble p24 antigen detection (0-3 protein precursor of HIV-1 in HeLa cells. In COS-1 cells, ng/ml) using supernatant solutions from pGG1 and pGA1 transfected HeL~a cells (p, density; linear scale, 1.1-1.2 g/ml). A C pGA1 pGG1 kD a bc a bca bc pUA1 transfected cells, virus-specific antigens were identi- fied both by immunoprecipitation of metabolically labeled proteins and by immunoblot detection and were compared to 971 iI

68 A B -Pr55 1 2 3 4 M kD M 1 2 3 4 43- _- jre 29on% -200 ___-p24 B C pGAl pGG1 97 s .* ps p Pspspsps N\ Pr55--__ -PrI80 \68 97- _ !:

-Pr55 43~ - 43rk 43- -Pr55 p24- -_ 29 29- _ -p24

FIG. 4. Processing and ofPr55gag in vivo. (A) Pulse-chase analysis of transfected cells. Cells transfected with FIG. 3. Comparison of Pr55gag produced by pGA1 and pGG1 SFneo (control, C), pGA1, or pGG1 were metabolically labeled with transfected cells. HeLa cells were cotransfected with either pGA1 or [35S]methionine/cysteine for 8 hr. Fresh medium containing unla- pGG1 and SFneo. Cell clones resistant to G418 (400 ,ug/ml; GIBCO) beled amino acids was added for 0 (lane a) or 24 (lane b) hr. Cell-free and positive by HIV-1 p24 ELISA were used to further analyze the supernatants (lane c) were harvested after the 24-hr chase. Proteins effect of myristoylation on HIV-1 assembly and propagation. (A) were separated on a 7.5-20% linear SDS/polyacrylamide gel gradi- Immunoblot analysis of transfected cells. Lysates were prepared ent. (B) Immunoblot analysis of subcellular fractions prepared from from cells transfected with pGG1 (lanes 1 and 3), pGA1 (lane 2), or transfected cells. Unlabeled cells were broken by Dounce homoge- SFneo alone (lane 4). The proteins were separated by SDS/PAGE nization and separated by centrifugation at 45,000 rpm into cytosol and detected as described. Molecular size markers are shown in the or soluble (s), and membrane or pellet (p), fractions: SFneo alone (C), center in kDa. The relative positions of p24 and/or Pr55gag are pGAl, or pGG1. The position of the 180-kDa gag-pol precursor, shown on the left in A and on the right in B. (B) [3H]Myristate Pr55gag, and p24 are indicated on the right. These samples were incorporation into Pr55gag. Cell lines were metabolically labeled prepared in a 150 mM NaCl buffer and are designated (-) for low salt. with [3H]myristic acid and the HIV proteins were immunoprecipi- The membrane fractions were resuspended in the same buffer, tated with pooled AIDS patients' sera. Autoradiography was con- adjusted to 1 M NaCl, and separated into soluble (s) and pelleted (p) tinued for an extended period in an attempt to detect labeled Prl80. fractions by centrifugation. +, High salt wash fractions. Downloaded by guest on September 26, 2021 526 Biochemistry: Bryant and Ratner Proc. Natl. Acad. Sci. USA 87 (1990) although some processing of nonmyristoylated Pr55gag was specific proteins in each fraction were identified by immu- detected, it was markedly delayed compared to processing of noblot detection (Fig. 4B). the myristoylated form of Pr55gag (not shown). In the first set of experiments, HIV-specific proteins Nonmyristoylated Pr55gag Can Be Processed in Vitro. At- present in the cytosol or membrane fractions, prepared from tachment of myristate could potentiate a conformation fa- pGG1, pGA1, and control (C) HeLa cell lines were examined vorable for recognition and cleavage of Pr55gag by the HIV {Fig. 4B, lanes s(-) and p(-) [the (-) indicates that the protease. In the absence of myristoylation, specific cleavage fractions were separated in low ionic strength buffer]}. No sites may not be accessible to the protease. Alternatively, the HIV-specific proteins were detected in the subcellular frac- myristoyl moiety could serve to localize or stabilize the pre- tions prepared from the control HeLa cells (an irregular, nonspecific band appears in each lane in the middle portion cursor molecule to a site in the cell necessary for association of the gel). The membrane fraction of the pGG1 cell line with the protease as a prerequisite to processing. To distinguish contained the majority of the myr' Pr55gag and all of the between these possibilities, processing of myr' and myr- processed p24, as expected. The distribution ofthe nonmyris- Pr55gag was compared in vitro using purified HIV protease. toylated Pr55gag in the pGA1 cell line was similar to that seen Extracts of selected HeLa cell clones expressing myr' in the pGG1 cell line. However, in repeated experiments, the pGG1 or myr- pGA1 Pr55gag were incubated with purified total amount of myr- Pr55gag in the cell was greater than that HIV protease (50 ng/ml) and the time-dependent of the myr' Pr55gag. As suggested above, this probably of Pr55gag was followed by immunoblot analysis of the reflects the slower turnover (lack of processing) of the non- cleavage products using a mouse monoclonal antibody to p24 myristoylated precursor. (Fig. 5). The disappearance of myr' Pr55gag in the pGG1 Washing the membrane fractions with low ionic strength extract could be detected at 0.5 min (lane 2), was essentially buffer did not alter the relative distribution of Pr55gag (or complete by 12 min (lane 4), and coincided with the appear- p24) in either the pGA1 or pGG1 preparations (data not ance ofthe virus capsid antigen p24. The degree ofprocessing shown). However, when the ionic strength ofthe wash buffer of myr+ Pr55gag in the steady state is reflected by the was increased (1 M NaCl) the majority of the myr- Pr55gag presence of p24 at time 0 (lane 1). polyprotein redistributed to the soluble fraction [s(+)], while The myr- Pr55gag in the pGA1 extract was also specifically the localization of myr+ Pr55gag and p24 in the membrane cleaved by the HIV protease in vitro (Fig. 5 Center). The fraction did not change. These data indicate that while reaction showed kinetics similar to that seen when the myr+ targeting and transport of Pr55gag to the plasma membrane Pr55gag was used as the substrate. Specificity of proteolysis may be influenced by myristoylation, additional properties of was confirmed by selective inhibition of cleavage of Pr55gag the polyprotein must control these events. In addition, the (Fig. 5 Right) by pepstatin A (aspartyl protease inhibitor) affinity of Pr55gag to the membrane (stabilization) is clearly (lane c) but not leupeptin ( protease inhibitor) (lane d), enhanced by the myristate moiety. The larger gag-pol precur- or antipain (cysteine protease inhibitor) (not shown). sor polypeptide Prl80gag-pol, which is generated by a trans- Subcellular Localization of Nonmyristoylated Pr55gag. It lational frameshift between the gag and pol reading frames, is has been proposed that the hydrophobic myristate moiety also detected in pGA1 cells (Fig. 4B) (37). Its identity was plays a role in the association of the modified protein with confirmed by immunologic detection using a monoclonal an- cellular membranes (1). The presence of the fatty acyl group tibody to p24 (not shown). In the pGA1 cell line, this protein could influence protein-protein or protein-lipid interactions is not myristoylated (Fig. 3, lanes 2) and displays a cellular directly, or indirectly, by a change in protein conformation. distribution similar to that of the myr- Pr55gag (Fig. 4B). To begin to examine the various possibilities, the distribution of Pr55gag and its cleavage product p24 in pGG1 and pGA1 DISCUSSION transfected cells was determined by subcellular fraction- were swollen in The major structural nucleocapsid proteins of most mamma- ation. Unlabeled cells hypotonic buffer, are cleaved broken in a Dounce homogenizer, and the nuclei and unbro- lian retroviruses including HIV-1 specifically from a myristoylated polyprotein precursor. Comparison of ken cells were removed by centrifugation. The remaining amino acid of this from different cellular components were separated into cytosol [soluble (s)]- the sequences precursor The virus- HIV-1 isolates, as well as other mammalian retroviruses, and membrane [pellet (p)]-containing fractions. shows that the myristoyl acceptor is always an N-terminal pGG1 pGA1 pGA1 glycine residue (Gly-1). In this study, we used site-directed kD 1 2 3 4 5 1 2 3 4 5 a b c d mutagenesis to replace the N-terminal myristoyl-acceptor glycine of Pr55gag, the structural polyprotein precursor of 97- HIV-1, with the nonacceptor amino acid alanine. Our results _._-~- __ -Pr55 demonstrate that myristoylation is required for virus assem- 43 _ - _- _ bly and infectious particle production. Without the myristoyl 29- accumulates in infected HeLa cells and mm --Am An-4 ___924 moiety, Pr55gag jj,-- processing into the mature structural virion components is prevented. In COS-1 cells, although proteolysis of nonmyris- to can be detected data not FIG. 5. Processing of nonmyristoylated Pr55gag by the HIV toylated Pr55gag p24 (ref. 21; protease in vitro. Extracts (50 ,ug) of HeLa cell clones transfected shown), efficient processing is dependent on myristoylation. with pGG1 or pGA1 were incubated with purified HIV-1 protease (50 A level of expression of the gag-pol precursor in COS-1 cells ng/ml) at 250C for 0, 0.5, 2, 12, or 30 min (lanes 1-5, respectively). higher than that in HeLa cells could account for this quan- The samples were precipitated with 10%o trichloroacetic acid, rinsed titative difference in myristoylation-dependent proteolysis. with acetone, solubilized in sample buffer, and separated on a Under these conditions, protease activation may be more 7.5-20%o gradient SDS/polyacrylamide gel. HIV-specific proteins efficient and less dependent on association with a specific were detected by immunoblot techniques using a mouse monoclonal subcellular compartment. Alternatively, it may reflect a antibody to p24 (Pr55gag, which contains p24, is also recognized). or cell Extracts from transfected cells were incubated for 0 min (lane species line-specific phenomenon. pGA1 or of the to alanine in a), or 12 min (lane b) in the absence of added inhibitor or in the Deletion of Gly-1 mutagenesis Gly-1 presence of 1 mM pepstatin A (lane c) or 1 mM leupeptin (lane d). Pr65gag of MoMLV also blocked myristoylation (22). The Molecular mass markers are shown on the left in kDa and the nonmyristoylated polyprotein failed to associate with the positions of Pr55gag and p24 are indicated on the right. membrane fraction and virus particle formation was inhibited. Downloaded by guest on September 26, 2021 Biochemistry: Bryant and Ratner Proc. Natl. Acad. Sci. USA 87 (1990) 527 Similarly, a Gly-1 to valine substitution in the MPMV Pr78gag 1. Towler, D. T., Gordon, J. I., Adams, S. P. & Glaser, L. (1988) Annu. Rev. Biochem. 57, 69-99. polyprotein resulted in an accumulation of immature type A 2. Schultz, A. M., Henderson, L. E., Oroszlan, S., Garber, E. A. & particles containing unprocessed precursor and prevented pro- Hanafusa, H. (1985) Science 227, 427-429. duction of infectious virions (23). A single point mutation 3. Olson, E. N., Towler, D. A. & Glaser, L. (1985) J. Biol. Chem. 260, present in the endogenous murine ecotropic provirus EMV-3 3784-3790. also accounts for the nonmyristolated gag polyprotein 4. Schmidt, M. F. G. (1983) Curr. Top. Microbiol. Immunol. 102,101-129. probably 5. Buss, J. E., Solski, P. A., Schaeffer, J. P., MacDonald, M. J. & Der, and the lack of infectious virus particles observed in vivo (38). C. J. (1989) Science 243, 1600-1603. Taken together, these studies suggest that protein N- 6. Schmidt, M. F. G. & Schlesinger, M. J. (1979) Cell 17, 813-819. myristoylation plays a central role in gag polyprotein asso- 7. Schmidt, M. F. G. & Schlesinger, M. J. (1980) J. Biol. Chem. 255, membrane component 3334-3339. ciation with a specific cellular where 8. Paul, A. V., Schultz, A., Pincus, S. E., Oroszlan, S. & Wimmer, E. maturation and budding occurs. In this regard, we found that (1987) Proc. Nadl. Acad. Sci. USA 84, 7827-7831. in the absence of the myristoyl moiety, Pr55gag was only 9. Chow,M., Newman,J. F. E., Filman, D.,Hogle,J. M.,Rowlands,D. J. weakly associated with the membrane fraction. In contrast, & Brown, F. (1987) Nature (London) 327, 482-486. 10. Streuli, C. H. & Griffin, B. E. (1987) Nature (London) 326, 619-622. myristoylated Pr55gag was tightly bound. This suggests that 11. Schmidt, M., Muller, H., Schmidt, M. F. G. & Rott, R. (1989) J. Virol. assembly and budding of virus requires membrane targeting 63, 429-431. ofPr55gag and stable complex formation mediated, at least in 12. Marc, D., Drugeon, G., Haenni, A., Girard, M. & van der Werf, S. (1989) part, by the myristoyl moiety. The activity of the virus- EMBO J. 8, 2661-2668. 13. Persing, D. H., Varmus, H. E. & Ganem, D. (1987) J. Virol. 61, encoded protease may also be dependent on the correct 1672-1677. topographical association of its precursor molecule, Prl8M. 14. Buss, J. E., Kamps, M. P., Gould, K. & Sefton, B. M. (1986) J. Virol. Alternatively, failure to myristoylate this precursor might 58, 468-472. prevent dimerization-dependent autocatalytic release of the 15. Kaplan, J. M., Mardon, G., Bishop, J. M. & Varmus, H. E. (1988) Mol. Cell. Biol. 8, 2435-2441. protease, the proposed mechanism for activation of the 16. Schultz, A. M. & Oroszlan, S. (1983) J. Virol. 46, 355-361. aspartyl protease (39). Without the active protease and/or 17. Schultz, A. M. & Oroszlan, S. (1984) Virology 133, 431-437. association with its specific substrate myr' Pr55gag, matu- 18. Henderson, L. E., Krutsch, H. C. & Oroszlan, S. (1983) Proc. Nall. ration and budding of HIV infectious particles cannot occur. Acad. Sci. USA 80, 339-343. 19. Henderson, L. E., Benveniste, R. E., Sowder, R., Copeland, T. D., It is interesting to note that avian and certain mammalian Schultz, A. M. & Oroszlan, S. (1988) J. Virol. 62, 2587-2595. retroviruses including two members of the lentivirus subfam- 20. Delchambre, M., Gheysen, D., Thines, D., Thiriart, C., Jacobs, E., ily (visna and equine infectious anemia virus) and the human Verdin, E., Horth, M., Burny, A. & Bex, F. (1989) EMBO J. 8, produce gag precursors that are not myris- 2653-2660. spumaretrovirus 21. Gottlinger, H. G., Sodroski, J. G. & Haseltine, W. A. (1989) Proc. Natl. toylated (40, 41). An alternative mechanism of gag precursor Acad. Sci. USA 86, 5781-5785. localization and/or association with membranes in cells 22. Rein, A., McClure, M. R., Rice, N. R., Luftig, R. B. & Schultz, A. M. infected by these viruses is likely to be responsible but is (1986) Proc. Natl. Acad. Sci. USA 83, 7246-7250. currently uncharacterized. 23. Rhee, S. S. & Hunter, E. (1987) J. Virol. 61, 1045-1053. 24. Veronese di Marzo, F., Copeland, T. D., Oroszlan, S., Gallo, R. C. & Although it is possible that substitution or elimination of the Sarngadharan, M. G. (1988) J. Virol. 62, 795-801. N-terminal glycine residue could itself affect the processing of 25. Mervis, R. J., Ahmad, N., Lillehoj, E. P., Raum, M. G., Salazar, the gag polyprotein, our data demonstrating cleavage of the F. H. R., Chan, H. W. & Venekatesan, S. (1988)J. Virol. 62,3993-4002. myr Pr55gag in vitro, using purified HIV protease, suggests 26. Ratner, L., Haseltine, W., Patrarca, R., Livak, K. J., Starcich, B., Josephs, S. F., Doran, E. R., Rafalski, J. A., Whitehorn, E. A., Bau- that the defect in processing in transfected cell lines is a specific meister, K., Ivanoff, L., Petteway, S. R., Jr., Pearson, M. L., Lauten- effect ofthe fatty acid moiety. In support ofthis conclusion, the berger, S. A., Papas, T. S., Ghrayeb, J., Chang, N. T., Gallo, R. C. & inhibitor of de novo fatty acid synthesis, cerulenin, prevents Wong-Staal, F. (1985) Nature (London) 313, 277-284. myristoylation of Pr55gag of HIV-1 and its proteolytic cleavage 27. Sanchez-Pescador, R., Power, M. D., Barr, P. J., Steimer, K. S., Stem- pien, M. M., Brown-Shimer, S. L., Gee, W. W., Renard, A., Randolph, to p24 (42). In addition, the potential inhibitor of protein A., Levy, J. A. & Luciw, P. A. (1985) Science 227, 484-492. N-myristoylation, N-myristoyl glycinal diethylacetal, prevents 28. Meyers, G. (1989) Human Retroviruses and AIDS (Los Alamos NatI. myristoylation of pl7gag, the N-terminal component of Lab., Los Alamos, NM). Pr55gag, and may disrupt viral assembly (43). 29. Fisher, A. G., Ratner, L., Mitsuya, H., Marselle, L. M., Harper, M. E., Broder, S., Gallo, R. C. & Wong-Staal, F. (1986) Science 233, 655-659. It may be possible to further examine the contribution of 30. Kunkel, T. A. (1985) Proc. Nail. Acad. Sci. USA 82, 488-492. acyl chain hydrophobic character to the biological function(s) 31. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. Acad. Sci. served by protein by generating a family of fatty USA 74, 5463-5467. acid analogues that act as alternative acyltransferase sub- 32. Graham, F. L. & Van der Eb, A. J. (1973) Virology 52, 456-467. 33. Southern, P. J. & Berg, P. (1982) J. Mol. Appl. Genet. 1, 327-340. strates (44). Also, radiolabeled analogues of myristic acid 34. Poiesz, B. J., Ruscetti, F. W., Gazdar, A. F., Bunn, P. A., Minna, J. D. could potentially be used to analyze the intracellular steps & Gallo, R. C. (1980) Proc. NatI. Acad. Sci. USA 77, 7415-7419. involved in protein processing and targeting. The develop- 35. Towbin, H., Staehelin, T. & Gordon, J. (1979) Proc. Nail. Acad. Sci. ment of specific inhibitors ofprotein acylation that are active USA 76, 4350-4354. 36. Laemmli, U. K. (1970) Nature (London) 227, 680-685. in vivo without substantial cellular toxicity would also be 37. Jacks, T., Power, M. D., Masiarz, F. R., Luciw, P. A., Barr, P. J. & desirable given the importance of gag polyprotein myristoy- Varmus, H. E. (1988) Nature (London) 331, 280-283. lation in human replication (45). 38. Jorgensen, E. C., Kjeldgaard, N. O., Pedersen, F. S. & Jorgensen, P. (1988) J. Virol. 62, 3217-3223. We thank 39. Navia, M. A., Fitzgerald, P. M. D., McKeever, B. M., Leu, C., Heim- J. I. Gordon and R. 0. Heuckeroth for helpful criticism; bach, J. C., Herber, W. K., Sigal, I. S., Darke, P. L. & Springer, J. P. E. Sadler and L. Westfield (Washington University Howard Hughes (1989) Nature (London) 337, 615-620. Institute) for providing the oligonucleotides; M. Arens for the pooled 40. Stephens, R. M., Casey, J. W. & Rice, N. R. (1986) Science 231, AIDS patients' sera; S. Fine and D. Loh for the gift of SFneo; R. Finn 589-594. and B. Salsgiver (Monsanto Corporation, Saint Louis) for the 41. Mauer, B., Banner, T. H., Darai, G. & Flugel, R. M. (1988) J. Virol. 62, purified HIV-1 protease; and the AIDS Research and Reference 1590-1597. Reagent Program, National Institute of Allergy and Infectious Dis- 42. Pal, R., Gallo, R. C. & Sarngadharan, M. G. (1988) Proc. NatI. Acad. eases, National Institutes of Health, for the monoclonal antibody to Sci. USA 85, 9283-9286. 43. Shoji, S., Tashiro, A. & Kubota, Y. (1988) J. Biochem. (Tokyo) 103, p24 (no. 249). M.B. is a scholar ofthe American Foundation for AIDS 747-749. Research. L.R. is a Hartford Foundation Fellow. Research was also 44. Heuckeroth, R. O., Glaser, L. & Gordon, J. 1. (1988) Proc. Nail. Acad. supported in part by Contract DAMD 17-87-C-7102 and Grants Sci. USA 85, 8795-8799. AI-25903 (manuscript 8 of the Washington University AIDS Clinical 45. Bryant, M. L., Heuckeroth, R. O., Kimata, J., Ratner, L. & Gordon, Trial Group), AI-27179, and AI27302. J. I. (1989) Proc. Natl. Acad. Sci. USA 86, 8655-8659. Downloaded by guest on September 26, 2021