from bovine brain (1), the T-cell-specific p56 protein from. LSTRA cells (55), and a number of gag and gag-onc fusion proteins of mammaliantype B, C, and D ...
JOURNAL OF VIROLOGY, Apr. 1987, p. 1045-1053
Vol. 61, No. 4
0022-538X/87/041045-09$02.00/0 Copyright © 1987, American Society for Microbiology
Myristylation Is Required for Intracellular Transport but Not for Assembly of D-Type Retrovirus Capsids SUNG S. RHEE AND ERIC HUNTER* Department of Microbiology, University of Alabama at Birmingham, Birmingham, Alabama 35294 Received 26 August 1986/Accepted 12 December 1986
The role of myristylation, a fatty acid modification of nascent polypeptides, in the assembly and intracellular transport of D-type retroviral capsids was investigated through the use of oligonucleotide-directed mutagenesis. Myristic acid is normally esterified through an amide linkage to a glycine residue at the amino terminus of the Mason-Pfizer monkey virus gag gene products. Mutant pA-1, which has a codon for valine substituted for that of the normally myristylated glycine, is completely noninfectious. While the mutant gag polyprotein precursors are synthesized at normal levels, they are not myristylated and are not cleaved to the mature virion proteins. No extracellular virus particles are released from mutant pA-1-infected cells, but intracytoplasmic A-type particles (capsids) accumulate in the cytoplasm. Since none of the intracellular capsids can be found associated with the plasma membrane, these results strongly suggest that myristylation is a critical signal for intracytoplasmic transport of completed viral capsids to their normal site of budding and release.
Many biologically active proteins are covalently modified during or after translation. Two distinct fatty acid acylations, which result in a hydrophobic modification of the protein, have been reported to be a feature of a large number of enveloped RNA and DNA virus proteins. The most frequent is an ester-type acylation in which palmitic acid is covalently linked to membrane-spanning cell surface glycoproteins (43, 44). Recent studies have identified a second form of fatty acid addition in which myristic acid (C14:0), a shorter-chain fatty acid than palmitic acid, is covalently linked to the amino termini of both cell and viral polypeptides. The catalytic subunit of the cyclic AMP-dependent protein kinase from bovine cardiac muscle was the first protein for which it was demonstrated that myristic acid is in an amide linkage to an amino-terminal glycine residue (12). Although myristylation is not a common form of protein modification in eucaryotic cells, there are several cellular and viral proteins with a single myristic acid addition; these include the ,-chain of calcineurin, the calmodulin-binding protein from bovine brain (1), the T-cell-specific p56 protein from LSTRA cells (55), and a number of gag and gag-onc fusion proteins of mammalian type B, C, and D retroviruses (7, 28, 47, 48). In addition, the tyrosine-specific protein kinase, p6Osrc, encoded by the Rous sarcoma virus transforming gene has been shown to be similarly modified (11, 24, 46). Although the biological functions of the amino-terminal myristyl group remain unclear, the attachment of long fatty acids to proteins has been suggested to be important for the interaction of the latter with hydrophobic plasma membranes. Indeed, p6Osrc is bound tightly to cellular membranes yet lacks a sequence of hydrophobic amino acids at the amino-terminal membrane-binding domain (52), and variants of this polypeptide which do not contain myristic acid bind poorly to membranes (10, 21, 34, 40). Nevertheless, the presence of an amino-terminal myristyl group on the NADHcytochrome b5 reductase, one of the components of the microsomal stearyl coenzyme A desaturase system, cannot be explained solely on the basis of anchoring the modified protein to a lipid bilayer, because the hydrophobicity of the enzyme itself would be predicted to be sufficient to provide *
the nonpolar interaction between the plasma membrane and protein (37). It has therefore been suggested that the aminoterminal myristyl moiety mediates selective transport of these modified proteins to specific cellular membrane receptors and orients them on the membrane to optimize their functions (37, 48). As with other mammalian retroviruses, the gag precursor, Pr785ag, of Mason-Pfizer monkey virus (M-PMV) is myristylated (47). M-PMV, first isolated from a spontaneous breast carcinoma in a female rhesus monkey (17), represents the prototype of the D-type retroviruses. This immunosuppressive virus (22) was classified into a new genus because although it shared morphogenic and biological properties with both B-type and C-type retroviruses, it differed from both (35, 36). Indeed, recent studies on the nucleotide sequence of M-PMV have shown that it originated from a recombination event between the gag-pol region of a B-type retrovirus and the env region of a primate C-type virus (50). M-PMV resembles the B-type retroviruses in that intracytoplasmic A-type particles (ICAPs) are preassembled in the infected cell cytoplasm and then migrate to the plasma membrane, where they are enveloped in a virus glycoprotein-containing lipid bilayer. M-PMV thus provides an excellent system for studying factors involved in retroviral morphogenesis, since the processes of capsid assembly and virus budding are unlinked both temporally and spacially. An analysis of the amino acid sequence of the M-PMV Pr785ag deduced from nucleic acid sequence studies (50) shows the penultimate amino acid to be a glycine. Cotranslational removal of the terminal methionine would thus yield the amino-terminal glycine that has been shown to be required for the amide linkage of myristic acid (48). On the basis of the proposed membrane-binding function of plO, the amino-terminal myristylated protein that is cleaved from Pr789'9 (5), myristic acid would appear to be primarily responsible for the nonpolar interaction between the membrane and ICAPs, since the polypeptide itself is not notably hydrophobic. However, whether myristylation of Pr78gag might also be essential for assembly of the A-type particles was not clear, and one of the objectives of this study was to answer this question.
Corresponding author. 1045
RHEE AND HUNTER
We report here on a point mutation, introduced into the M-PMV DNA genome, that blocks the addition of myristic acid to P078g`". The mutation, which alters the codon for the penultimate glycine to one for valine, does not interfere with A-type particle assembly but completely prevents their transport to and subsequent budding from the plasma membrane. Since the nonmyristylated viral capsids accumulate in the cytoplasm, myristylation appears to play a role in directing the intracellular transport of cytoplasmic poly-
MATERIALS AND METHODS
Oligonucleotide-directed mutagenesis. Oligonucleotide-
directed mutagenesis of the codon for glycine (the second amino acid residue of plO) was performed on an infectious M-PMV proviral DNA genome which had been cloned into pAT153 (2). This clone, which was derived from an integrated provirus, was designated pMPMV 6A/7. To accomplish the mutagenesis, a 2.6-kilobase-pair SphI-BamHI fragment of pMPMV 6A/7 containing the gag gene and the 3' end of the left long terminal repeat was subcloned into the polylinker of M13mpl9. An oligonucleotide 24 residues long with a single-base substitution that converted the codon for glycine to one for valine was used to carry out the sitedirected mutagenesis on single-stranded M13 viral DNA as previously described (56). After mutagenesis, a 0.8-kilobasepair NarI-SstI fragment, which is flanked by restriction enzyme sites unique in the pMPMV 6A/7 genome, was excised from the replicative form of M13 and substituted for the wild-type fragment in the original plasmid. Bacterial colonies derived from cells transformed by plasmid containing the mutated fragment were identified by hybridization to the mutagenic oligonucleotide. The presence of the mutation was confirmed by dideoxy sequencing of the doublestranded DNA. The mutant viral DNA containing a valine codon instead of a glycine codon at amino acid position 2 was
Cell culture and radiolabeling. HeLa cells were used for transfection of wild-type and mutant proviruses. Monolayer cultures of HeLa cells were grown in Dulbecco modified Eagle medium supplemented with 2.5% calf serum and 2.5% fetal calf serum (complete growth medium). Semiconfluent monolayers of HeLa cells were cotransfected with either pMPMV 6A/7 or pA-1 and pPB3, a pKO-neo derived vector (53) that expresses the hygromycin resistance gene, by the calcium phosphate precipitation method of Graham and Van der Eb (25), modified as described by Stow and Wilkie (51). pPB3 was constructed and kindly provided by P. Bird, Health Sciences Center, Dallas, Tex. To determine whether the cell clones selected in medium containing 600 ng of the antibiotic hygromycin B per ml expressed virus structural proteins, each cell clone was pulse-labeled with [3H]leucine, lysed in detergent, and immunoprecipitated with antiserum against M-PMV proteins (5, 6). Cell clones containing integrated wild-type and mutant virus DNA were designated 6A/7 and A-lplO, respectively. Confluent monolayers in 35-mm dishes were incubated for 1 h in leucine-free medium and then pulse-labeled for 15 min at 37°C in 250 [LI of leucine labeling medium containing [3H]leucine (0.8 mCi/ml, 63 Ci/mmol; Amersham Corp.) with occasional rocking. In pulse-chase experiments, pulse-labeled cells were chased for 4 h in complete growth medium. For fatty acid labeling, [9,10-3H]myristic acid (55 Ci/ mmol; Amersham) in toluene was dried and then dissolved in small amount of dimethyl sulfoxide, and complete growth
medium was added as previously described (11). Labeling was carried out for 2 h at 37°C in [3H]myristic acid labeling medium (0.5 mCi/ml). Antisera and immunoprecipitation. Antiserum to M-PMV p27 was obtained from the Division of Cancer Cause and Prevention, National Cancer Institute, Bethesda, Md., and antiserum to gp7O was raised in rabbits by inoculation of gel-purified protein (13, 15). Immunoprecipitation of viral proteins was carried out as described previously (5). Gel electrophoretic analysis and electrophoretic blotting of polypeptides. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was carried out to separate virus polypeptides as described previously with either a 10% or a 14% resolving gel (5). Proteins in polyacrylamide gels were electrophoretically transferred to nitrocellulose at 0.2 A for 5 h at 4°C as previously described (16). After transfer, nitrocellulose blots were soaked in bovine lacto transfer technique optimizer (5% nonfat dry milk in phosphatebuffered saline with 0.01% sodium azide; BLOTTO) (30) to saturate additional protein-binding sites and then reacted with goat anti-p27 antiserum, appropriately diluted in BLOTTO, for 16 h at 4°C. Bound antibody was detected by incubating blots with 125I-labeled protein A, prepared as described previously (26), for 1 h at 37°C. Virus and ICAP preparation. To quantitate virions released from the 6A/7 and A-lplO clones, cells were labeled for 48 h with [3H]leucine (1 mCi/ml) in complete growth medium. Culture fluids were harvested from the cells twice, at 24-h intervals, and then pooled. Viruses were pelleted and purified through a 24 to 48% (wt/wt) linear sucrose gradient as described previously (5). For separation of ICAPs from the bulk of the cellular proteins, unlabeled cell monolayers were lysed in Triton X-100 lysis buffer (0.25 M sucrose, 1.0 mM EDTA, 10 mM Tris [pH 7.5], 0.14 M NaCl, 0.5% Triton X-100). Nuclei were pelleted by centrifugation at 5,000 rpm in a Beckman JA20 rotor for 10 min, and then ICAPs were pelleted from the clarified supernatants by centrifugation at 30,000 rpm for 100 min in a Beckman SW41 rotor. Pellets were dissolved directly in 200 ,ul of PAGE sample buffer and applied to a 10% polyacrylamide gel for immunoblotting. RT assay. Infectivity of proviral clones was assessed indirectly by measuring the spread of virus through the culture in a reverse transcriptase (RT) assay. This assay was also used to determine release of RT-containing virion from cell clones. Culture fluid from each cell clone was clarified, and then the viruses pelleted by centrifugation for 2 h at 40,000 rpm in a Beckman SW41 rotor at 4°C. Pelleted viruses were disrupted in 20 [LI of lysis buffer (0.05% Triton X-100, 100 mM KCl, 2 mM dithiothreitol, 50 mM Tris hydrochloride [pH 8.0]) for 15 min on ice. The RT activity was measured with some modifications as described previously (42). A 4-plI portion of the disrupted virus was incubated with 16 [L of RT mixture which contained poly(rA) as synthetic RNA template and oligo(dT) as primer. The extent of DNA synthesis was determined by spotting the reaction mixture onto a DE81 filter disk wetted with 0.5 M phosphate buffer (pH 6.5) and then washing thoroughly in the same buffer. The incorporated [3H]TTP (57 Ci/mmol; Amersham) on the dried disk was quantitated by being counted in an aqueous scintillation cocktail (Budget Solve; Research Products Int. Corp.). Electron microscopy. Cells were prepared for electron microscopy as described previously (18). The cell clones grown in 12-well plates were fixed for 1 h at room temperature with 1% glutaraldehyde and then washed in phosphate-
buffered saline. After postfixation with 1% osmium
M-PMV MYRISTYLATION MUTANT
VOL. 61, 1987
A " T
GATC GATC FIG. 1. DNA sequence of the mutant and wild-type gag gene. Oligonucleotide-directed mutagenesis was carried out on a subcloned SphI-BamHI fragment of pMPMV 6A/7 by using a 24-nucleotide synthetic oligonucleotide with a single-base substitution of G to T at nucleotide 2 of the glycine codon GGG. The presence of the mutation in mutant DNA pA-1 was confirmed by dideoxy sequencing. (A) Wild-type pMPMV 6A/7 shows the ATG initiation codon followed by the GGG codon for the second amino acid, glycine. (B) In mutant DNA pA-1, the ATG codon is followed by the GTG codon for valine.
tetroxide, cells were embedded in an epoxy resin mixture, sectioned, and stained with uranyl acetate and lead nitrate. All preparations were examined with a Philips 301 electron microscope. RESULTS Mutagenesis of the amino-terminal glycine of gag protein plO. To remove the amino-terminal glycine residue that is conserved in all myristylated proteins, we replaced the codon for glycine with one for valine by using oligonucleotide-directed mutagenesis. Mutagenesis was carried out on a subcloned fragment from an infectious M-PMV proviral DNA pMPMV 6A/7 as described in Materials and Methods. The mutated fragment was recloned into pMPMV 6A/7, and the presence of the mutation was confirmed by sequencing. The sequence of the gag gene in wild-type pMPMV 6A/7 starts at the ATG initiation codon, which is followed by the GGG codon for glycine (Fig. 1A). In the mutant DNA pA-1, the ATG was followed by the GTG codon for valine instead of GGG for glycine (Fig. 1B). To determine the effects of this mutation on virus replication, we transfected HeLa cells with the mutated provirus (pA-1) or pMPMV 6A/7 and then measured RT activities in the material pelleted from culture fluids on days 3, 5, and 7 after transfection. At 3 days after transfection, a rapid increase in RT activity was observed in the wild-type genome-infected HeLa cells; in contrast, no increase was observed in the mutant genome-infected cells (Fig. 2). Since the increase in RT activity with time in transfected cells
reflects the spread of infectious virus through the culture, we can conclude that the mutation of glycine to valine at residue 2 of the gag gene product renders the genome noninfectious. To pinpoint the defect in virus replication, cell lines expressing the wild-type and mutant genomes were established. HeLa cells were cotransfected with pMPMV 6A/7 or pA-1 and a plasmid that confered hygromycin resistance. After drug treatment, colonies were assessed for M-PMV protein synthesis. Two cell lines, expressing either the mutant genome (line A-lplO) or wild-type genome (line 6A/7) to equivalent levels were selected for further studies. The mutant gag polyproptein precursors are not myristylated. To determine whether the mutation blocked the covalent addition of myristic acid to the gag-related precursors, cells of lines A-lplO and 6A/7 and uninfected HeLa cells were cultured for 2 h at 37°C in [3H]myristic acid labeling medium (0.5 mCi/ml) or pulse-labeled for 15 min in [3HJleucine labeling medium (0.8 mCi/ml). The cells were lysed and then immunoprecipitated with anti-p27 antiserm to identify M-PMV gag-related polyproteins. M-PMV-infected cells produce three precursor polyproteins related to the gag gene (5). The major precursor polyprotein is a molecule of 78 kilodaltons (kDa), Pr78 ,a which is the precursor of internal structural proteins (5). The two other precursors, presumably synthesized via a frame-shifting mechanism are P95, the viral protease precursor (5, 50), and Prl80, the precursor to the RT (14). In both wild-type- and mutant-infected cells pulse-labeled with [3H]leucine, all three precursors can be clearly identified (Fig. 3A, lanes 2 and 3). As would be predicted from the results of Schultz and Oroszlan (47), [3H]myristic acid labeled Pr78gag and P95 in 6A/7 cells (Fig. 3B, lane 2), whereas, in contrast, no myristic acid-labeled protein could be immunoprecipitated from A-lplO cells (Fig. 3B, lane 3). This result provides formal proof that an amino-terminal glycine is essential for myristic acid addition; the valine residue cannot substitute for glycine in this role. It should be noted that in addition to the major gag precursor products, bands of 68 and 85 kDa can be immunoprecipitated from leucine-labeled mutant- and wildtype-infected cells (Fig. 3A, lanes 2 and 3). While these represent only minor products in 6A/7 cells (lane 2), they are of equivalent intensity to Pr78gag and P95 in A-lplO cells (lane 3). A review of the M-PMV genomic sequence (50) 60000 -
50000 40000 30000 20000
20000-/ 10000 0 0
4 Days post Infection
FIG. 2. Determination of the infectivity of mutant genome pA-1.
Infectivity of mutant genome pA-1 was determined by measuring RT activity in culture fluids of HeLa cells transfected with pA-1 at days 3, 5, and 7 posttransfection. Wild-type-transfected cells (-) show an increase in RT activity 3 to 5 days after transfection, reflecting the rapid spread of infectious viruses through the culture. In contrast, in mutant-infected cells (*), no RT activity above that detected with uninfected cells (O) was observed.
RHEE AND HUNTER
Pr180 P 95 Pr78
A 2 3
B 2 3
FIG. 3. Immunoprecipitation of gag precursors labeled with
[3H]myristic acid or [3H]leucine. Cells of lines A-1p1O and 6At7, expressing the mutant pA-1 genome or the wild-type pMPMV 6A/7 genome, respectively, were labeled with [3H]myristic acid for 2 h or pulse-labeled with [3H]leucine to determine the level of myristic acid addition to the mutant gag precursors. The labeled cells were immunoprecipitated with anti-p27 antiserum to identify the gagrelated polyproteins, which were separated by 10% SDS-PAGE. The three gag precursors, Prl80, P95, and Pr78ag, were detected in [3H]leucine-labeled wild-type-infected (panel A, lane 2) and mutantinfected (panel A, lane 3) cells. [3H]myristic acid labeled p95 and Pr78gag in wild-type cells (panel B, lane 2), but no proteins were labeled with [H]myristic acid in mutant cells (panel B, lane 3). Uninfected HeLa cells were labeled with [3H]leucine (panel A, lane 1) and [3H]myristic acid (panel B, lane 1) as a control.
revealed a second, in-frame, initiator codon approximately 300 base pairs (100 amino acids) downstream; initiation at this ATG could thus generate the 10-kDa smaller products that were observed (consistent with this, bands of 180 and 170 kDa are seen in mutant-infected cells). Thus it appears likely that the mutation of the glycine codon (GGG) to that for valine (GTG) causes approximately 50% of the ribosomes to skip the normal gag ATG. The basis for this result is currently under investigation. The mutant Pr78gag polyprotein is not cleaved to virion structural proteins. The M-PMV Pr78gag polyprotein is known to be cleaved to yield the six nonglycosylated internal structural proteins (plO, ppl6-18, p12, p27, p14, and p4) of the virion during a final stage of assembly (5, 29). The M-PMV env gene encodes an 86-kDa precursor, Pr86env, which is cleaved to two glycosylated proteins, gp7O and gp2O, found on the surface of the viral envelope (6). To determine whether the nonmyristylated precursor proteins were processed in a normal manner, a pulse-chase experiment was performed. Figure 4A (lanes 2 and 3) shows three prominant gag-related polyproteins, Prl80, P95, and Pr78gag, and the env precursor, Pr86env, immunoprecipitated with antiserum against M-PMV proteins from the pulselabeled wild-type 6A/7 and mutant A-lplO cells. After a 4-h chase in 6A/7 cells, the intensity of three gag-related bands decreased with a concomitant appearance of proteins of 27 and 16 kDa, corresponding to the major internal and phosphorylated structural proteins p27 and ppl6 of the
virion (Fig. 4B, lane 5). The env precursor was processed to two glycosylated proteins, gp7O and gp2O. From these results, it can be concluded that a 4-h chase is sufficient for the majority of the polyproteins synthesized in the pulse to be cleaved. Although some reduction in intensity of the three gag-related polyproteins in A-lplO cells was observed, no new bands corresponding to p27 and ppl6 appeared (lane 6). This indicates that in the absence of myristic acid addition, the precursor cleavage function of the viral protease is defective. The processing of Pr87env is shown more clearly in Fig. 4C and D. In both 6A/7 and A-lplO cells, Pr86env, immunoprecipitated with anti-gp7O antiserum after a pulselabel (Fig. 4C), was processed after a 4-h chase to a diffuse band of 70 kDa corresponding to gp7O (Fig. 4D). Thus, the env gene products are synthesized and processed normally into the mature glycosylated products of M-PMV in both wild-type- and mutant-infected cells. It should be noted, however, that the turnover of gp7O in A-1-10 cells appears to be faster than that of wild-type-infected cells (Fig. 4D), a point that will be developed in the Discussion. To determine whether the nonmyristylated Pr78gag accumulates in the uncleaved form, total cellular proteins from lysates of both 6A/7 and A-lplO cells were separated by SDS-PAGE and electrophoretically blotted onto nitrocellulose for Western blot analysis by using anti-p27 antiserum (Fig. SA). The A-lplO cell lysate (lane 3) shows two clear bands corresponding to the gag-related polyproteins Pr78gag and the smaller 68-kDa gag product mentioned above but no bands corresponding to the cleaved product, p27. On the A 1
D 10 11 12
Prl188 P 95 Pr86 Pr78
FIG. 4. Processing of M-PMV precursors in mutant cells. The processing of gag and env precursors, Pr78 9 and Pr86env, in mutant-infected A-1p1O cells and wild-type-infected 6A/7 cells was determined by pulse-chase experiments. Polyproteins immunoprecipitated from pulse-labeled cells with anti-M-PMV antiserum (A) and anti-gp7O antiserum (C) can be seen in both 6A/7 cells (lanes 2 and 8) and A-1p1O cells (lanes 3 and 9) as predicted. After a 4-h chase (B and D), new bands corresponding to p27, gp2O, and ppl6 can be clearly seen in immunoprecipitates of wild-typeinfected cells with anti-MPMV antiserum (lane 5). A diffuse band of gp7O can be seen in both anti-MPMV (lane 5) and anti-gp7O precipitated 6A/7 cells. In the mutant-infected A-lplO cells, the cleaved gag products p27 and ppl6 are not observed but the mature env gene products gp20 and gp7O can be clearly seen (lanes 6 and 12). HeLa cells were labeled and immunoprecipitated as a control (lanes 1, 4, 7, and 10).
M-PMV MYRISTYLATION MUTANT
VOL. 61, 1987
other hand, bands corresponding to p27 and Pr78gag were clearly observed in 6A/7 cell lysates (lane 2). These results are consistent with those of the pulse-chase experiment that showed that the nonmyristylated gag precursor is not cleaved to the virus structural proteins. Although in Fig. 5A no obvious accumulation of precursors is seen, in other experiments the intensity of the bands in mutant-infected lysates was three- to fivefold that of bands from wild-type infected cells (data not shown). The fact that the Pr78rar band detected in 6A/7 lysates was less intense than the p27 band supports our previous data that the gag polyproteins of wild-type-infected cells are rapidly cleaved to the final structural proteins (5). Mutant A-lplO cells do not release virus particles. The experiments described above demonstrate that cells transfected with the mutant pA-1 genome do not release infectious virus. Nevertheless, it was important to determine whether noninfectious virions were assembled and shed from mnutant-infected cells. In initial experiments, we tested a pelleted sample of culture fluid to determine whether the A-lplO cell line released RT-containing particles into the medium. Virions were concentrated from the culture media of either 6A/7 or A-lplO cells and tested for enzyme activity by using an exogenous poly(rA)-oligo(dT) template. No RT activity was obtained with the culture fluids from mutant A-lplO cells, whereas high levels of incorporation were A 1
P 95 Pr78 Pr68
FIG. 5. Western blot analysis of cell lysates and ICAP preparations. To determine whether the nonmyristylated Pr78gag accumulated in an uncleaved form in mutant-infected cells, Western blot analyses of total cellular proteins from cell lysates (A) were performed with anti-p27 antiserum. Immunoblots of uninfected, 6A/7, and A-lplO cell lysates are shown in lanes 1, 2, and 3, respectively. No p27 band can be observed in immunoblots of A-lplO cell lysates, while this cleavage product is the major M-PMV product associated with wild-type-infected 6A/7 cells. The 68-kDa p27-related polypeptide observed in lane 3 appears to be an alternate initiation product of the mutant gag gene (see text). To determine whether the nonmyristylated gag polyproteins preassemble ICAPs, crude ICAPs from 0.5% Triton X-100-treated cell lysates were pelleted through a 30% sucrose cushion, electrophoresed by SDS-PAGE, and then immnunoblotted with anti-p27 antiserum (B). Mutant-infected A-lplO cells (lane 6) as well as wild-type-infected 6A/7 cells (lane 5) show assembly of ICAPs composed of uncleaved gag polyproteins in the cytosol. Much more intense bands of P95, Pr78gag, and Pr68gag can be observed in the pelleted material from mutant cells than that from wild-type cells.
1.1 8 -E
St 14 X
-1 12 0
-1 10 o
U) p 27 6
FIG. 6. Determination of virus release from mutant-infected A-lplO cells. (A) Viruses released from A-lplO and 6A/7 cells labeled with [3H)leucine were quantitated after centrifugation through a 24 to 48% linear sucrose gradient. Radiolabeled particles which were released from wild-type-infected cells sedimented to a region of the gradient with a density of 1.16 g/ml (O). No corresponding peak is observed in sucrose gradieqts of A-1p1O cell supernatants (*). (B) Western blot analysis of potential viruses pelleted from the culture fluids of uninfected (lane 1), 6A/7 (lane 2), and A-lplO (lane 3) cells, confirms the results shown in panel A. An intense p27 band can be seen in the lane correspondihg to 6A/7 supernatants, while no M-PMV p27-related proteins could be seen in supernatants fron mutant-infected A-1p1O cells.
observed with 6A/7 culture fluids (data not shown). The RT assay could detect only enzyme-containing virus released from the cell clones, and since it was possible that myristylation was required for incorporation of the enzyme precursor into virions, we tested cells for release of [3H]leucine-labeled particles, as described in Materials and Methods. A distinct peak of radioactivity was obtained at a density of 1.16 g/ml from 6A/7 cells (Fig. 6A), consistent with the patterns previously observed (14). No corresponding peak could be seen in sucrose gradients of A-lplO supernatant fluids, indicating that noninfectious virus is not released from the cells. This result was confirmed by Western blot analysis. Potential virus was pelleted from 200 ml of culture medium from 6A/7 and A-lplO cells, electrophoresed by SDS-PAGE, and immunoblotted with anti-p2t anstiserum after transfer to a nitrocellulose membrane (Fig. 6B). Neither a virus-specific p27 band nor the gag-related precursors Pr78gag or P95 were seen in mutant A-lplO supernatants (lane 3), whereas an intense p27 band was observed with supernatants from the wild-type 6A/7 cells (lane 2). These data suggest that myristylation of gag precursor polyproteins is a critical step in virus morphogenesis. Myristic acid modification of the gag precursors is not required for assembly of ICAPs. M-PMV preassembles an ICAP, which then migrates to the plasma membrane to acquire its envelope during budding from the cell. We therefore prepared crude ICAPs from 6A/7 and A-lplO cells by pelleting 0.5% Triton X-100 lysates through a 30% sucrose cushion. The polypeptides in the pellets were separated on a 10% polyacrylamide gel, transferred to nitrocellulose, and probed with anti-p27 antibody. The results of this experiment (Fig. 5B) show bands of 95, 78, and 68 kDa in the wild-type-infected cell lane (lane 5) and much more intense bands in equivalent positions in the A-1p1O lane (lane 6). Since equivalent amounts of pelleted protein were loaded in each lane, these results suggest that the nonmyristylated precursor proteins were accumulating in mutant cells in the form of ICAPs. To extend this observation, we examined
RHEE AND HUNTER