Pr65 Gag Research Articles

Mouse APOBEC3 interferes with autocatalytic cleavage of murine leukemia virus Pr180gag-pol precursor and inhibits Pr65gag processing

Mouse APOBEC3 (mA3) inhibits murine leukemia virus (MuLV) replication by a deamination-independent mechanism in which the reverse transcription is considered the main target process. However, other steps in virus replication that can be targeted by mA3 have not been examined. We have investigated the possible effect of mA3 on MuLV protease-mediated processes and found that mA3 binds both mature viral protease and Pr180gag-pol precursor polyprotein. Using replication-competent MuLVs, we also show that mA3 inhibits the processing of Pr65 Gag precursor. Furthermore, we demonstrate that the autoprocessing of Pr180gag-pol is impeded by mA3, resulting in reduced production of mature viral protease. This reduction appears to link with the above inefficient Pr65gag processing in the presence of mA3. Two major isoforms of mA3, exon 5-containing and -lacking ones, equally exhibit this antiviral activity. Importantly, physiologically expressed levels of mA3 impedes both Pr180gag-pol autocatalysis and Pr65gag processing. This blockade is independent of the deaminase activity and requires the C-terminal region of mA3. These results suggest that the above impairment of Pr180gag-pol autoprocessing may significantly contribute to the deaminase-independent antiretroviral activity exerted by mA3.

Full-Length Glycosylated Gag of Murine Leukemia Virus Can Associate with the Viral Envelope as a Type I Integral Membrane Protein.

The glycosylated Gag protein (gPr80) of murine leukemia viruses (MLVs) has been shown to exhibit multiple roles in facilitating retrovirus release, infection, and resistance to host-encoded retroviral restriction factors, such as APOBEC3, SERINC3, and SERINC5. One way in which gPr80 helps MLVs to escape host innate immune restriction is by increasing capsid stability, a feature that protects viral replication intermediates from being detected by cytosolic DNA sensors. gPr80 also increases the resistance of MLVs to deamination and restriction by mouse APOBEC3 (mA3). How the gPr80 accessory protein, with its three N-linked glycosylation sites, contributes to these resistance mechanisms is still not fully understood. Here we further characterized the function of gPr80 and, more specifically, revealed that the asparagines targeted for glycosylation in gPr80 also contribute to capsid stability through their parallel involvement in the Pr65 Gag structural polyprotein. In fact, we demonstrate that sensitivity to deamination by the mA3 and human A3 proteins is directly linked to capsid stability. We also show that full-length gPr80 is detected in purified viruses. However, our results suggest that gPr80 is inserted in the NexoCcyto orientation of a type I integral membrane protein. Additionally, our experiments have revealed the existence of a large population of Env-deficient virus-like particles (VLPs) harboring gPr80 inserted in the opposite (NcytoCexo) polarity, which is typical of type II integral membrane proteins. Overall this study provides new insight into the complex nature of the MLV gPr80 accessory protein.IMPORTANCE Viruses have evolved numerous strategies to infect, spread in, and persist in their hosts. Here we analyze the details of how the MLV-encoded glycosylated Gag (gPr80) protein protects the virus from being restricted by host innate immune defenses. gPr80 is a variant of the structural Pr65 Gag protein with an 88-amino-acid extended leader sequence that directs the protein for translation and glycosylation in the endoplasmic reticulum. This study dissects the specific contributions of gPr80 glycans and capsid stability in helping the virus to infect cells, spread, and counteract the effects of the host intrinsic restriction factor APOBEC3. Overall this study provides further insight into the elusive role of the gPr80 protein.

Recovery of Glycosylated gag Virus from Mice Infected with a Glycosylated gag-Negative Mutant of Moloney Murine Leukemia Virus.

Two independent pathways for gag gene expression exist in Moloney murine leukemia virus (M-MuLV). One begins with Pr65(gag) that is processed and cleaved into the internal structural proteins of the virion. The other pathway begins with the glycosylated gag polyprotein, gPr80(gag). gPr80(gag) consists of Pr65(gag) plus additional N-terminal residues and it is glycosylated. A glycosylated-gag-negative mutant of M-MuLV (Ab-X-MLV) was previously constructed and shown to replicate in tissue culture. To test for the importance of glycosylated gag in vivo, the Ab-X-MLV mutant was inoculated intraperitoneally into newborn NIH Swiss mice. Mutant-infected mice developed typical lymphoblastic lymphomas at rates comparable to wild-type M-MuLV at either high (2 x 10(4) XC pfu/animal) or low (2 x 10(2) XC pfu/animal) doses. However, when viral protein expression was examined in the resultant tumors, six out of six mice showed evidence of virus that had recovered gPr80(gag) expression. These results suggest that glycosylated gag is important for M-MuLV propagation or leukemogenesis in vivo. Copyright 1994 S. Karger AG, Basel

Effect of heat on viral protein production and budding in cultured mammalian cells.

The life cycle of enveloped viruses is intimately associated with, and influenced by, host cell membrane organization, which is altered by hyperthermia. Hyperthermia-modified Moloney murine leukaemia virus (M-MuLV) release, protein production and intracellular protein processing in a chronically infected cultured murine cell line, C9CL98 (C9). Both 44 degrees C/45 min and 42.8 degrees C/135 min substantially decreased cell-free viral env protein 8-48 h postheating, but virus release and cellular viral protein content increased following 42.8 degrees C/25 min. Proteolytic processing of viral Pr65 gag precursor to p30 gag protein, normally observed within unheated C9 cells, was blocked for at least 8 h after 44 degrees C/45 min. Virus released from heated C9 cells was as infectious to NIH/3T3 cells as was virus from control cells. Cells surviving exposure to 42.8 degrees C/135 min became thermotolerant to decreased virus release from a second heating if delivered 10-48 h after the initial heating. The mechanism by which virus release is blocked after hyperthermia remains to be elucidated.

The HIV-1 Gag precursor Pr55 gag synthesized in yeast is myristoylated and targeted to the plasma membrane

Myristoylation of the Pr65 gag protein from Moloney murine leukemia virus has been shown to be essential for virus particle formation [Rein et al., Proc. Natl. Acad. Sci. USA 83 (1986) 7246–7250], and by analogy, myristoylation of the human immunodeficiency virus (HIV) Gag precursor could possibly play a similar role. We have investigated the expression and myristoylation of the complete HIV Gag precursor Pr55 gag in yeast, the subcellular localization of that protein, and the contribution of the myristoyl-glycine residue to this localization. Immunogold labelling of myristoylated Pr55 gag with antibodies directed against HIV Gag products was apparent in the vicinity of the plasma membrane. On the contrary, non-myristoylated derivatives of Pr55 gag were only detected in relatively well-defined regions of the cytoplasm. These results show that targeting and accumulation of the HIV Gag precursor, Pr55 gag, at the plasma membrane occurs in yeast in the absence of other viral components and requires the N-myristoyl-glycine residue.

CUG initiation codon used for the synthesis of a cell surface antigen coded by the murine leukemia virus

Murine leukemia virus (MuLV) codes for two precursors of the group-specific antigens, Pr65 gag and Pr75 gag, in vivo. While Pr65 gag is the precursor to the virion structural proteins, Pr75 gag undergoes glycosylation and is found on the surface of the infected cell as gp85 gag, and it is thought to play a role in virus maturation and spread. Pr65 gag synthesis starts at an AUG codon within a favourable initiation context (AAUAUGG at positions 618 to 624). The gp85 gag start codon is upstream but its precise location is not known. To map the initiation codon of gp85 gag, we used deletion and site-directed mutagenesis of the leader sequence of MuLV RNA and in vitro translation of the RNAs. Synthesis of the MuLV gp85 gag protein appears to be initiated at a CUG codon located within a favourable context (ACCCUGG at positions 354 to 359 for Moloney-MuLV). The possible function of gp85 gag was investigated by expressing Moloney-MuLV and Friend-MuLV proviral DNA and mutants deficient for gp85 gag synthesis in mouse and rat cells. The results indicate that the gp85 gag protein probably facilitates the spread of virus infection in tissue culture.

Processed enzymatically active protease (p15 gag) of avian retrovirus obtained in an E. coli system expressing a recombinant precursor (Pr25 lac-Δgag)

Processing proteases of avian and mammalian retroviruses cut the polyprotein precursors encoded by the retroviral genes into mature functional proteins. Retroviral processing proteases are still a rather poorly characterized group as to their relation to other proteases, specificity, and mechanism of enzymatic action. In avian retroviruses the generation of the processing protease itself comprises a processing cleavage event — the protease p15 gag is cut off the carboxy-terminus of a gag polyprotein precursor, Pr76 gag . We report here that direct and efficient production of the avian retrovirus processing protease p15 gag (required for structure-function studies and rational design of inhibitors) was obtained in an E. coli system, where massive expression of a size-reduced, recombinant precursor (Pr25 lac-Δgag was accompanied by its structurally accurate processing.

Characterization of a virus-specific proteolytic activity processing the gag precursor of the simian sarcoma-associated virus

The proteolytic processing of the gag precursor polypeptide pr65 gag of simian sarcoma-associated virus (SSAV) has been studied in vivo and in vitro. In SSAV-infected cells (i.e., in vivo) proteins of 52 and 38 kDa and the viral protein p30 could be immunoprecipitated with anti-p30 serum. This cleavage pattern is only in part imitated by in vitro cleavage of the isolated pr65 gag with avian myeloblastosis virus (AMV) protease pl5. However, in vitro incubation of isolated pr65 gag with detergent-disrupted SSAV particles generated products identical in size to those found in vivo, i.e. proteins of 52 and 38 kDa and p30. The extent of cleavage is dependent on the concentration of the disrupted virions added to the incubation mixture. Studies with protease inhibitors suggest that the SSAV enzyme is a serine-type protease like that of other mammalian retroviruses and unlike the protease of avian viruses. The SSAVprotease activity eluted from a molecular sieve column in a range of about 10–15 kDa reflecting the molecular weight of the murine leukemia virus (MuLV) protease ( M r = 13.5K). Thus, it appears that there is a close similarity between the proteolytic enzymes present indifferent mammalian retroviruses such as MuLV and SSAV.

Expression of avian sarcoma virus genes in tumor cells derived from different periods of tumor growth

Avian sarcoma virus (ASV)-induced tumor cells that are derived from regressing neoplasms contain about 75% less pp60 src kinase activity than do progressively-growing tumor cells. Paradoxically, the polyadenylated RNAs extracted from ‘regressor’ cells found in regressing neoplasms are much more productive than those from progressively-growing tumors in a reticulocyte lysate in vitro translation system, as determined both by time course and concentration curve studies. Electrophoretic analysis of polypeptides corresponding to known ASV proteins further suggests that the translation product from the mRNAs of regressing tumor cells contains much more viral protein than does that derived from progressively-growing sarcoma cells. The identity of two of these viral proteins was confirmed by immunoprecipitation of the respective translation products, using a tumor-bearing rabbit serum with specificity for pp60 src and Pr76 gag precursor polypeptides, and showed that much more of pp60 src and Pr76 gag were produced by the mRNAs from regressor than from progressor cells. We further found that ASV genes including genomic and subgenomic species (8.6, 5.0 and 3.2 kb) were present in greater amounts in regressing as opposed to progressively-growing tumor cells. These results support the notion that viral RNAs are accumulated in ‘regressor’ tumor cells.

Inhibition of cleavage of moloney murine leukemia virus gag and env coded precursor polyproteins by cerulenin

Cerulenin, an inhibitor of de novo fatty acid (and cholesterol) biosynthesis, has been shown to significantly decrease (>75%) the amount of Moloney murine leukemia virus (M-MuLV) released into the culture medium of chronically infected mouse fibroblasts (I. Katoh, Y. Yoshinaka, and R. B. Luftig, 1986, Virus Res., in press). In order to clarify the mechanism by which this decrease in virus production occurs, we analyzed the kinetics of gag and env coded protein synthesis in M-MuLV infected, cerulenin-treated cells by immunoprecipitation with monospecific antisera to p30, p12, p10, gp70, and p15(E). We found that in pulse (15 min-2 hr)-chase (0–4 hr) experiments the cleavage of not only Pr65 gag to p30 and other gag coded proteins but Pr80 env to gp70 and Pr15(E) as well, was greatly reduced by cerulenin treatment. Further, since the total amount of label in the Pr65 gag and Pr80 env bands remained about the same or was slightly decreased in 2-hr pulsed, cerulenin-treated cells, this suggests that cerulenin decreases virus production, in part, by inhibiting the cleavage of both precursor gag and env coded polyproteins during virus assembly and budding at the cell membrane. We also observed that at longer chase periods (4 hr), the effect of cerulenin could be partially overriden in that minor amounts of cleaved gag and env coded polyproteins were produced and assembled into virion particles. However, these particles contained abnormally large amounts of the uncleaved precursor Pr65 gag , suggesting that maturation was incomplete. The above results suggest two independent, but not exclusive, possible mechanisms of cerulenin action to block M-MuLV production, viz. (i) cerulenin decreases the pool of fatty acids, thereby inhibiting fatty acid acylation of Pr65 gag , as well as Pr80 env , and thus preventing the interaction between gag (the p15 antigenic determinant on Pr65 gag ) and env [the p15(E) antigenic determinant of Pr15(E)] coded gene products at the cell membrane needed for efficient virus assembly (M. Satake and R. B. Luftig, 1983, Virology 124, 259–273), and (ii) cerulenin inhibits one or more proteolytic enzymes responsible for the cleavage of Pr65 gag and Pr80 env .

The effect if cerulenin on the synthesis of the precursor gag polyprotein in defective murine leukemia and sarcoma virus producing cell lines

The effect of cerulenin, an inhibitor of de novo fatty acid (and cholesterol) biosynthesis, on the synthesis of the precursor gag polyprotein, Pr65 gag in a defective murine leukemia virus (334C) producing cell lines (MJD-54, clone 2) the amount of the Pr65 gag remaining producing hamster cell line (HTG-2) was examined. In contrast to Moloney murine leukemia virus (M-MuLV) producing cell lines (MJD-54, clone 2) the amount of the Pr65 gag remaining in the presence of cerulenin (20 μg/ml) was greatly reduced in both defective virus-infected cells. This effect appears specific for the Pr65 gag polyprotein, since the env precursor polyprotein Pr80 env was normally synthesized and remained undegraded in cerulenin-treated 3JE-infected cells. Thin-section electron micrographs showed an increased accumulation of virion particles in vesicles of treated HTG-2 cells.

The effect of cerulenin on Moloney murine leukemia virus morphogenesis

Cerulenin is an antibiotic that interferes with fatty acid synthesis in eukaryotic cells. It had been shown by Schultz and Oroszlan (1983), that murine leukemia virus (MuLV) Pr65 gag, the polyprotein precursor to the virion core proteins contains the fatty acid myristate at its NH 2 terminus. We showed that when 20 μg/ml of cerulenin is added for 3 h to mouse fibroblasts chronically infected with Moloney (M)-MuLV it causes a > 4-fold decrease in virus production. This is accompanied by an accumulation of uncleaved Pr65 gag in the infected cells. Further, thin-section electron micrographs of cerulenin-treated cells show a 2-fold increase in the number of nascent-budding forms, as well as the appearance of aberrant viral forms at the cell membrane. This suggests that the failure to add myristic acid to Pr65 gag prevents their proper assembly into viral particles.

Detection of the myristylated gag-raf transforming protein with raf-specific antipeptide sera

The post-translational modifications of the gag-raf fusion proteins of the 3611 murine sarcoma virus (MSV) have been examined by inhibiting glycosylation with tunicamycin and by in vivo labeling with [ 3H]myristic acid. The results show that P75 gag-raf is myristylated but not glycosylated and that P90 gag-raf is glycosylated but not myristylated (and is now termed gp90 gag-raf). gP90 gag-raf expression appeared to become lost during passage of the transformed cells, and consequently does not appear to be necessary for the maintenance of transformation. raf-specific sera for detecting gag-raf fusion proteins have been obtained from synthetic peptides made from different regions of the predicted v- raf sequence. Immunoprecipitation of P75 gag-raf with raf-specific sera directly confirmed the deduced v- raf sequence. The fact that P759 gag-raf is both myristylated and precipitated by antiserum to a predicted carboxyl-terminal peptide of the v- raf gene established that the mature protein represents the entire coding region. The gP90 gag-raf thus appears to be a glycosylated form of P75 gag-raf specified by the gag sequences of the fusion protein, in analogy with Pr65 gag and gPr80 gag of murine leukemia viruses. Antiserum to the carboxyl-terminal P75 gag-raf peptide was the most efficient in immunoprecipitation, and will be useful for detecting the product of the c- raf gene.

Influence of base-pairing in the leader region on in vitro translation of Rous sarcoma virus RNA

The capacity of the leader region of Rous sarcoma virus (RSV) RNA to act as a regulator of viral protein synthesis was tested in vitro. When DNA/RNA hybrids of sufficient length (> 90–100 nucleotides) are created within the leader, synthesis of Pr76 gag is inhibited. The inhibition is dependent upon the length of the hybrid rather than its position with the exception that encumberance of the 5'-terminal 33 nucleotides does block translation. These findings demonstrate that physical alteration of the non-coding leader structure directly affects downstream initiation of protein synthesis. It is thus likely that biochemical or physiologic changes in vivo which alter the structure of the leader may affect regulation of viral protein synthesis.

Presence of circulating antibodies against gag-gene MuLV proteins in patients with autoimmune connective tissue disorders

An immunoblotting procedure using viral proteins from purified murine sarcoma virus or MSV-(MLV) has been developed to characterize antiviral antibodies in sera from patients with autoimmune connective tissue disorders. Fifty-eight sera with anti-Sm, anti-RNP, anti-SS-B (La), and other undefined specificities were found to react with several major viral polypeptide bands. Most of them corresponded to gag-gene-encoded products: pr65 gag, p40 gag, p30, p15, p12 and p10. Other bands with molecular weights averaging 90K, 60K, 45K, and 28K were recognized by a few sera. Immunological specificity of the reaction was assessed by reproducing the tests with IgG purified from sera and from corresponding F(ab′)2 fragments. Moreover, the specificity of the reaction with gag proteins was confirmed by repeating the tests with p30 and p15 prior purified by immunoprecipitation with anti-p30 and anti-p15 goat sera. Furthermore, the gag polypeptides were recognized by human sera by replacing MSV-(MLV) by three other murine retroviruses of different origin. An indirect confirmation of these results was obtained by applying this method to sera of MRL lpr/lpr mice which develop an autoimmune syndrome comparable to that of human systemic lupus erythematosus. In agreement with previously published results ( C. Rordorf, C. Gambke, and J. Gordon (1983), J. Immunol. Methods 59, 105–112), we found that anti- gag-gene antibodies were present in the sera of individual mice. Patterns of reactivity were found to vary with the age of the animals. No retroviral polypeptide was significantly detected in the great majority (80%) of sera from normal donors. However, 5 out of 25 sera showed faint bands although to a lesser extent than pathological sera. These five sera also reacted with HeLa cell purified HnRNPs, suggesting that their normal status should be reconsidered.

Rous sarcoma virus encodes a transcriptional activator

Rous sarcoma virus expresses a transcriptional activator that affects the LTR as well as other promoters. We discern this activity as a stimulation of the transient expression of an LTR-promoted hybrid transcriptional unit and also of the rat preproinsulin II gene in transfected NIH 3T3 cells. We map the activity to an alternate reading frame in the p19-p10 region of the gag gene and identify a mRNA whose spliced structure would direct translation of this reading frame from the Pr76 gag initiation codon. This mRNA probably differs from genomic RNA only by the 282 nucleotide splice. The predicted translation product is a 124 residue polypeptide; the first six amino acids arise from gag. The target for the action of this transcriptional modulator at the LTR lies between 111 and 620 nucleotides upstream of the cap site.

Biological and biochemical characterization of a new isolate of feline sarcoma virus: Theilen-pedersen (TP1-FeSV)

A new feline sarcoma virus designated Theilen-Pedersen (TP1-FeSV) has been isolated from a spontaneous, slowly growing fibrosarcoma of a domestic short-haired 4-year-old castrated cat. The virus codes for a gag-onc fusion protein of 83,000 molecular weight phosphorylated in vivo at serine, threonine, and tyrosine residues. Cells transformed in vitro with TP1-FeSV exhibit five- to 10-fold elevated levels of phosphotyrosine over FeLV-infected cells. The gag-onc polyprotein has associated with it a tyrosyl protein kinase activity which in vitro results in autophosphorylation of the molecule at tyrosine residues. The fusion protein cannot be labeled metabolically with [ 3H]glucosamine and tunicamycin has no effect on the electrophoretic mobility of the in vivo [ 32P]orthophosphate-labeled fusion protein. The fusion protein, in common with the gag precursor Pr65 gag, can be metabolically labeled with palmitic acid.

Murine retrovirus Pr65 gag forms a 130K dimer in the absence of disulfide reducing agents

Gazdar-murine sarcoma virus (Gz-MSV) particles, obtained from tissue culture fluids of chronically infected HTG-2 hamster cells are immature in morphology and contain uncleaved Pr65 gag as the predominant protein (>95% Coomassie blue stain) ( A. Pinter and E. deHarven, 1979, Virology, 99, 103–110; Y. Yoshinaka and R. B. Luftig, 1982, Virology, 118, 380–388). When Gz-MSV particles are disrupted in 1% sodium dodecyl sulfate (SDS) and then analyzed by SDS-polyacrylamide gel electrophoresis (PAGE) in the absence of reducing agents, such as β-mercaptoethanol (β-MSH) almost half of the Pr65 gag Coomassie blue-stained band is detected as a band at a M r of 130K. Electrophoretic blotting studies with monospecific antisera against MuLV p30, p15, p12, and p10 showed that the 130K band cross-reacted with all four antigens suggesting that it was a dimer of Pr65 gag Two-dimensional (2D) SDS-PAGE where the first dimension was run under nonreducing conditions and the second with β-MSH, supported the contention that the 130K band was a dimeric complex of Pr65 gag One also saw minor amounts of a 260K and higher polymeric forms of Pr65 gag on the SDS gels, suggesting that polymeric forms may exist as well. When 32P-labeled Gz-MSV particles obtained by in vivo labeling of infected HTG-2 cells with [ 32P]PPi were electrophoresed on SDS-PAGE, only 10% of the 32P label was detected at the 130K position. In contrast, 30% of the Coomassie blue-stained Pr65 gag material was found at 130K on the 2D gels. This suggests that unphosphorylated Pr65 gag is more likely to participate in dimer formation than phosphorylated Pr65 gag Pr65 gag of Moloney murine leukemia virus (M-MuLV), which is present as a minor (5% of stain) protein band on SDS-PAGE also showed 130K dimers. Further, in β-MSH-deficient SDS preparations of Gz-MSV, electrophoresed after trypsin treatment, a 32K band that stained with p15, but not p10, p12, nor p30, antisera was observed. If β-MSH was added, this band was no longer present. Thus Pr65 gag dimerization in immature MuLV particles appears to at least involve the p15 region of the polyprotein. Since p15 is an extremely hydrophobic protein, formation of Pr65 gag dimers may occur when virion precursor proteins are brought to the cell membrane during virus assembly.

Impaired cleavage of the joint gag-pol polyprotein precursor and virion assembly in a temperature-sensitive mutant of rous sarcoma virus

A temperature-sensitive coordinate mutant tsLA83 of Prague (PR-B) strain of Rous sarcoma virus at the nonpermissive temperature (41°) produces noninfectious virus particles (NI-LA83) which contained only 3% of the reverse-transcriptase activity present in infectious virions. Analyses of [ 35S]methionine-labeled NI-LA83 showed the presence of all of the viral proteins except reverse transcriptase. Pulse-chase analyses of the virus-specified proteins in cells infected with LA83 or PR-B showed that the gag and glycoprotein precursors, Pr76 gag and gPr95 env , respectively, were processed at both 35 and 41°. The reverse-transcriptase precursor, Prl80P gag-pol however, was not processed in LA83-infected cells at 41°. In contrast, cells infected with LA83 or PR-B at 35° as well as with PR-B at 41° showed normal cleavage of Pr180 gag-pol . A shiftdown of LA83-infected cells at 41° to the permissive temperature 35° resulted in the normal processing of Pr180 gag-pol , and production of infectious virus containing reverse transcriptase. Electron microscopic analysis showed that at 41° cells infected with LA83 showed a large number of budding structures but fewer released particles. A shiftdown from 41 to 35° resulted in an increase of virus particles with a concomitant decrease in budding structures suggesting that the processing of reverse-transcriptase precursor is related to virion assembly.

Cytoskeleton-associated Pr65 gag and assembly of retrovirus temperature-sensitive mutants in chronically infected cells

Certain temperature-sensitive ( ts) mutants of murine leukemia virus (MuLV) were observed to be defective in virus assembly. These mutants also accumulated intracellular core protein precursor, Pr65 gag , at 39°, the nonpermissive temperature. At 39°, virions released from cells infected with the various is mutants also contained elevated levels of Pr65 gag relative to virions released at 33°, the permissive temperature. Detergent extraction of pulse-labeled cells with Nonidet P-40 (NP-40) generated an NP-40-insoluble cytoskeleton-enriched fraction. Reextraction of this fraction with deoxycholate followed by gel electrophoresis of solubilized, immunoprecipitated viral proteins showed that in Moloney MuLV (Mo-MuLV) ts3-infected cells, and in Rauscher MuLV (R-MuLV) ts17- and ts24-infected cells, increased amounts of intracellular viral Pr65 gag rapidly become associated with the cytoskeleton-enriched fraction during pulse labeling at nonpermissive temperature. Furthermore, examination of cell extracts from chase-incubated cells infected with these ts mutants revealed that Pr659 gag accumulated in the cytoskeleton-enriched fraction at 39° but not at 33°. During steady-state labeling, as much as half of the intracellular Pr65 gag becomes associated with the cytoskeleton-enriched fraction (i.e., is not solubilized by NP-40) at 39°. At permissive temperature only 10–15% of the intracellular Pr65 gag is cytoskeleton associated. In contrast, cells infected with R-MuLV ts25 or ts26 showed little or no preferential localization of Pr65 gag in the cytoskeleton-enriched cell fraction during a short pulse at 39°, but Pr65 gag accumulated in both the NP-40-soluble and -insoluble fractions during a chase incubation relative to the condition at 33°. Based upon these and previous results (Edbauer and Naso,1983), models for retrovirus assembly are described in which the association of Pr65 gag with the cell membrane and cytoskeleton plays a critical role in virus assembly, budding, and postbudding maturation.