Ubiquitination of Human Immunodeficiency Virus Type 1 Gag Is Highly Dependent on Gag Membrane Association

Expression constructs.The plasmids pΔR/PR(-), pΔR(PTAP-)/PR(-), and pHA-Ub were described in references 12 and 13. Plasmids encoding green fluorescent protein (GFP)-wild-type VPS4A, GFP-VPS4A-E228Q, GFP-VPS4A-K183Q, CHMP3-YFP, and cyan fluorescent protein CFP-CHMP4C (8, 46) were kindly provided by W. Sundquist. The pNL4-3 plasmid and its G2A mutant [pNL4-3(G2A)] (7) were kindly provided by E. Freed. The pcDNA3-based, Rev-independent Gag expression constructs were described in reference 13. The Gly→Val mutation encoded by codon 2 of the gag open reading frame in pcDNA3-Gag was introduced by mutagenic PCR. In MACA^NTD, pcDNA3-Gag is truncated after the N-terminal CA domain (amino acid 151).

Cell culture, transfections, analysis of virion release, and pulse-chase.293T cells were maintained and transfected using the calcium phosphate precipitation method as described previously (13). Medium was changed 6 h after transfection, and cells were further processed 24 h after transfection. Virion preparations and the pulse-chase experiment were performed as described in reference 12.

ELISA, Western blotting, and antibodies.An enzyme-linked immunosorbent assay (ELISA) for detecting CA was carried out as described previously (18). For improved detection of uncleaved Gag, samples were boiled in protein sample buffer containing 1% sodium dodecyl sulfate (SDS) before ELISA. For Western blotting, protein samples were resolved on 17.5% low-cross-linking polyacrylamide gels and probed with the indicated antibody dilutions. Rabbit anti-GFP (1:5,000), rabbit anti-MA (1:5,000), and sheep anti-CA (used for immunoprecipitation) were raised against recombinant proteins. HIV-1 CA-specific monoclonal antibody (183-H12-5C) was obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH, from Bruce Chesebro (5). Rat anti-hemagglutinin (anti-HA; clone 3F10; 1:500) was purchased from Roche, mouse anti-transferrin receptor (anti-TfR; 1:1,500) from Zymed, and mouse anti-actin (C2; 1:500) from Santa Cruz.

Gag immunoprecipitation.293T cells were lysed in 150 μl boiling SDS lysis buffer (1% SDS, 150 mM NaCl, 50 mM Tris, pH 8.0) and boiled for 10 min at 95°C. The lysate was diluted to 1× radioimmunoprecipitation assay (RIPA) buffer concentrations (0.1% SDS, 1% Triton X-100, 0.5% sodium deoxycholate, 150 mM NaCl, 50 mM Tris, pH 8.0). After sonification, the lysate was cleared by centrifugation for 10 min at 17,000 × g, 30 μl protein A-Sepharose beads (Amersham Biosciences) and 5 μl sheep anti-CA serum were added, and immunoprecipitation was performed for 2 h at 4°C. The amounts of immunoprecipitated Gag from different samples were thoroughly normalized by multiple rounds of ELISA.

Consecutive immunoprecipitations.293T cells were lysed in 0.5 ml cold 1× RIPA buffer containing 5 mM N-ethylmaleimide (NEM) and protease inhibitors. The lysate was cleared by centrifugation for 3 min at 1,000 × g at 4°C, and Gag was immunoprecipitated for 1 to 2 h at 4°C. Sepharose beads were then pelleted (1 min, 80 × g); the supernatant was adjusted to a 1% final SDS concentration, boiled for 10 min at 95°C, and diluted to a 1× RIPA concentration; and the second Gag immunoprecipitation was performed.

Membrane flotation assay.293T cells swollen on ice for 20 min in cold hypotonic buffer (10 mM Tris, pH 8.0, 1 mM Mg₂Cl, protease inhibitors, and 5 mM NEM) were disrupted by Dounce homogenization and adjusted to 40% OptiPrep (Progen). The mixture was overlaid with 28% OptiPrep and buffer and centrifuged at 165,000 × g for 3 h at 4°C. Eight fractions were collected from the top of the gradient.

A late budding arrest leads to a specific increase in detergent-resistant, membrane-associated Gag.We initially sought to test whether arresting HIV-1 budding at different stages might reveal differences in the ubiquitination status of Gag during the budding process. Therefore, 293T cells were transfected with the full-length Gag expression plasmid pΔR/PR(-) or a PTAP-mutated variant [pΔR(PTAP-)/PR(-)] or were cotransfected with pΔR/PR(-) and an expression plasmid for wild-type or dominant-negative VPS4A. Cells were lysed in cold RIPA buffer 24 h after transfection, and lysates were cleared by centrifugation at 17,000 × g. Surprisingly the amounts of Gag detected in the lysates were approximately equal for all samples (white bars in Fig. 1A), although a blocking of virus release should lead to an accumulation of intracellular Gag. To test whether the insoluble material obtained after clearing the lysates contained residual Gag, the pellets were solubilized by boiling them in SDS buffer and their Gag content was determined after dilution to RIPA conditions (Fig. 1A, black bars). In the case of wild-type Gag, more than 50% of total intracellular Gag was recovered in the RIPA-insoluble fraction. The amount of RIPA-insoluble Gag was further increased by a factor of 3 to 4 when budding was arrested by PTAP mutation or cotransfection of dominant-negative VPS4A, whereas the amount of Gag in the soluble fraction remained largely constant. The same effect was observed when virus release was blocked by dominant-negative CHMP4C or small interfering RNA-mediated knockdown of TSG101 (not shown). This suggests that RIPA-insoluble Gag mainly represents protein accumulating in virus buds when release is inhibited.

• Open in new tab
• Download powerpoint

FIG. 1.

Analysis of HIV-1 Gag solubility and immunoprecipitation. (A) 293T cells transfected with wild-type or PTAP-defective pΔR/PR(-) or cotransfected with pΔR(WT)/PR(-) and VPS4A expression constructs were lysed in RIPA buffer, and the insoluble fraction was removed by centrifugation. The pellet was dissolved by boiling it in SDS buffer and subsequently diluted to the same volume as the supernatant. The Gag contents in both fractions were determined by ELISA. Error bars represent standard deviations for three independent experiments. (B) Cells transfected with the same plasmids as those for panel A were lysed in RIPA buffer, and Gag was immunoprecipitated (IP) directly from the lysate (white bars) or following denaturation of the unbound supernatant (black bars). The amounts of precipitated Gag were determined by ELISA. Error bars represent standard deviations for three independent experiments. (C) pΔR/PR(-)-transfected cells were labeled with [³⁵S]methionine for 7 min and lysed in cold RIPA buffer after the indicated chase periods. Two successive Gag immunoprecipitations with an in-between denaturation step were performed as described for panel B. In parallel, virions were purified from the cell culture supernatant by centrifugation through sucrose cushions and Gag was immunoprecipitated after denaturation. All samples were analyzed by SDS-polyacrylamide gel electrophoresis (PAGE) and autoradiography, and the Gag signal of each sample was quantified using Quantity One software (Bio-Rad). A representative example from two independent experiments is shown. (D) Cells transfected with pΔR/PR(-) were subjected to membrane flotation by density gradient centrifugation. The membrane-bound and soluble fractions were divided into two aliquots, and one was left at 4°C and adjusted to 1× RIPA buffer (white bars), while the other was adjusted to 1% SDS, incubated for 10 min at 95°C, and then diluted to 1× RIPA buffer (black bars). Gag was immunoprecipitated from the four samples, and the amounts of precipitated Gag were determined by ELISA. Error bars represent standard deviations for three independent experiments.

This hypothesis was consistent with a previous study by Ono et al. (32), who showed that immunoprecipitation of membrane-bound HIV-1 Gag requires denaturation, presumably due to epitope masking in higher-order Gag multimers. We therefore performed consecutive Gag immunoprecipitations from cell lysates: Gag was first immunoprecipitated under nondenaturating conditions, followed by denaturation of the supernatant and a second Gag immunoprecipitation. A considerable fraction of Gag was resistant to immunoprecipitation under native conditions and could be immunoprecipitated only after denaturation (Fig. 1B). Consistent with the result presented in Fig. 1A, induction of a budding arrest by either PTAP mutation or coexpression of dominant-negative VPS4A led to a two- to threefold increase in the amount of Gag in the fraction requiring denaturation for immunoprecipitation, whereas wild-type VPS4A had no effect. This result supported the suggestion that multimerized Gag in the process of assembly and budding requires denaturation to become accessible to CA antibodies.

To test whether the Gag portion that requires denaturation for immunoprecipitation takes part in virus formation, we performed a pulse-chase experiment. Cells transfected with pΔR/PR(-) were pulse labeled for 7 min, and Gag amounts in the soluble, denaturation-dependent, and virus-associated fractions were quantified after different chase periods (Fig. 1C). Soluble Gag displayed the highest value at the earliest time point measured (10 min), while epitope-masked Gag, requiring denaturation for immunoprecipitation, reached its maximum after 25 min of chase. Subsequently, both fractions showed similar rates of decline concomitant with an increase of extracellular particulate Gag. This result suggests that the detergent-resistant fraction consists, at least in part, of Gag multimers that serve as assembly intermediates in the process of virus release.

To further validate our conclusion that the Gag fraction requiring denaturation for immunoprecipitation represents assembling molecules and is not due to cytoplasmic, nonproductive Gag aggregates, we analyzed whether it is membrane bound. Lysates from transfected cells were subjected to membrane flotation, and the top (membrane) and bottom (soluble) fractions were collected and divided into two aliquots. One of these was kept at 4°C, while the other was denatured, and Gag was subsequently immunoprecipitated from both samples (Fig. 1D). We observed that denaturation of the membrane-bound fraction increased Gag recovery approximately fivefold, while there was no effect on Gag recovery from the soluble fraction. Taken together, our results and those reported by Ono et al. (32) revealed that the epitope-masked fraction of Gag, requiring denaturation for immunoprecipitation, most likely represents assembly intermediates and assembled structures. Collectively, the results indicate that immunoprecipitation of immature HIV-1 Gag is strongly affected by loss of assembly intermediates following centrifugation (Fig. 1A) and by inefficient binding of antibodies to multimerized Gag (Fig. 1B). Both obstacles can be overcome by vigorous lysing conditions. An additional advantage of this procedure is the rapid inactivation of potential Gag-modifying enzymes, such as Ub hydrolases.

A late budding arrest leads to increased HIV-1 Gag ubiquitination.Having optimized the experimental conditions for HIV-1 Gag recovery by immunoprecipitation, we now investigated the influence of the ESCRT pathway on the ubiquitination status of Gag. According to the model of ESCRT action, PTAP mutation should impair the entry of Gag into the ESCRT pathway, while dominant-negative CHMP proteins should arrest virus budding at an intermediate stage, and dominant-negative VPS4A should block the disassembly of the ESCRT complex. 293T cells were cotransfected with pΔR/PR(-), pHA-Ub, and a construct expressing either wild-type or dominant-negative VPS4A, dominant-negative CHMP3-YFP, or CFP-CHMP4C (8, 46). Gag was immunoprecipitated from cell lysates obtained under denaturing conditions, and equal amounts of Gag (Fig. 2A, lower) were then subjected to Western blotting with an anti-HA antibody to determine the extents of Gag ubiquitination (Fig. 2A, upper). Typically, different isoforms of mono- and diubiquitinated Gag appeared as ladders of bands as previously described (13).

• Open in new tab
• Download powerpoint

FIG. 2.

Analysis of Gag-Ub levels for budding-arrested HIV-1. (A) 293T cells were cotransfected with pHA-Ub together with either pΔR(PTAP-)/PR(-) or wild-type pΔR(WT)/PR(-) and expression plasmids for wild-type (WT) or dominant-negative ESCRT proteins, as indicated above each lane. As a control, HA-Ub was coexpressed with pΔR(WT)/PR(-) alone (lane 1). Cells were lysed by boiling them in SDS lysis buffer, diluted into RIPA buffer, and subjected to Gag immunoprecipitation. Comparable amounts of Gag were analyzed by Western blotting using antisera directed against HA (top) or CA (bottom). Please note that minor differences in the HA-Ub signal intensities (lanes 2 and 4 to 7) are below the sensitivity of the method and were not reproducible. Gag and Gag-HA-Ub conjugates are identified at the right, and molecular mass standards are shown on the left. (B) Lysates from the same transfection as that for panel A were tested for Gag (top) or GFP (middle) expression, and particle preparations from the same transfections were analyzed by anti-CA immunoblot (bottom). Molecular mass standards (in kDa) are indicated on the left. YFP, yellow fluorescent protein.

Consistent with earlier studies, the ubiquitination level of PTAP mutant Gag (Fig. 2A, lane 2) was strongly increased compared to that of wild-type Gag (lane 1). Induction of a budding arrest by coexpression of dominant-negative VPS4A variants (lanes 4 and 5) or CHMP fusion proteins (lanes 6 and 7) led to similar increases in the amounts of Gag-Ub conjugates. In contrast, coexpression of wild-type VPS4A had only a slight effect (lane 3), which is consistent with its minor effect on virus release (8). Virus release was significantly inhibited upon late-domain mutation and coexpression of dominant-negative ESCRT proteins (Fig. 2B, bottom panel). As a control for expression, cell lysates were also probed with antibodies against CA or GFP (Fig. 2B, upper panels). In conclusion, induction of a late budding arrest at different stages consistently led to an increase in Gag-Ub levels, independent of the presence or absence of an intact late-domain motif.

Multimerized, detergent-insoluble Gag is ubiquitinated at higher levels than soluble Gag.Next, we investigated the ubiquitination levels of multimerized, membrane-bound Gag and soluble Gag. HA-tagged Ub was coexpressed with Gag, and consecutive Gag immunoprecipitations, with denaturation of the remaining protein after the first immunoprecipitation, were performed as described above. Similar amounts of Gag from both immunoprecipitations were subsequently analyzed by anti-HA blotting (Fig. 3A, left). At the conditions used, ubiquitination of soluble Gag was undetectable, while significant amounts of ubiquitinated Gag were found in the denaturation-dependent fraction. This still represented only a minor fraction of total Gag, since Ub-modified Gag products were not seen on the anti-CA blot (Fig. 3A, lower-left).

• Open in new tab
• Download powerpoint

FIG. 3.

Comparison of Gag ubiquitination levels in different cellular fractions. (A) 293T cells cotransfected with pΔR/PR(-) and pHA-Ub expression plasmids were lysed in cold RIPA buffer at 24 h posttransfection. Detergent-resistant Gag was separated from soluble Gag either by two consecutive Gag immunoprecipitations (IP) as described for Fig. 1B (left) or by pelleting the RIPA-insoluble fraction for 10 min at 17,000 × g (right). Equal amounts of Gag were analyzed by Western blotting using antisera against HA (top) or CA (bottom). Gag and Gag-HA-Ub conjugates are identified on the right, and molecular mass standards (in kDa) are shown on the left. (B) Cells were cotransfected with pΔR(WT)/PR(-) and pHA-Ub, lysed by Dounce homogenization, and subjected to membrane flotation by density gradient centrifugation. Fractions were collected successively from top to bottom, and equal volumes of each fraction were subjected to SDS-PAGE. Each fraction was analyzed for Gag content, membrane-bound TfR, and soluble actin by immunoblot analysis (lower panels). Fractions 1 (membrane associated) and 6 (soluble) were denatured, followed by Gag immunoprecipitation. Equal amounts of Gag were analyzed by Western blotting using antisera against HA and CA (top panels). (C) Cells were cotransfected with pHA-Ub and either the wild-type (WT) HIV-1 proviral plasmid pNL4-3, the membrane-binding-deficient variant pNL4-3(G2A), the variant pNL4-3Δp17 targeted to the endoplasmic reticulum (6), or the pNL43/78VE mutant targeted to the MVB (7, 31). Transfections were performed in the presence of 5 μM of the HIV protease inhibitor indinavir (45) to prevent Gag processing. Cells were lysed under denaturing conditions, Gag was immunoprecipitated, and equal amounts of Gag were analyzed by Western blotting using antisera against HA (top) and CA (bottom).

We also analyzed the level of Gag ubiquitination in the RIPA-soluble and RIPA-insoluble fractions (Fig. 3A, right). This experiment demonstrated that the Gag protein recovered from the RIPA-insoluble fraction was specifically ubiquitinated, while no Gag ubiquitination was detectable for RIPA-soluble Gag. Taken together, the results obtained for denatured and RIPA-insoluble HIV-1 Gag suggest that ubiquitination of Gag occurs during the process of multimerization and assembly, thus explaining why arresting the budding process at any stage would result in elevated levels of ubiquitinated Gag.

HIV-1 Gag ubiquitination is dependent on membrane association.Our group previously compared the levels of Gag ubiquitination in the membrane-bound and soluble fractions by membrane flotation experiments and subsequent Gag immunoprecipitation under nondenaturing conditions. In these experiments, we did not detect any difference in the relative levels of Gag ubiquitination (13). Based on the observation that ubiquitinated Gag is mainly found in the antibody-inaccessible fraction, we repeated this experiment but denatured the samples prior to immunoprecipitation. Cell lysates of 293T cells expressing HA-Ub and Gag were subjected to membrane flotation, and individual fractions were analyzed by Western blotting. Gag was detected in the top and bottom fractions of the gradient, while control proteins were distributed as expected (Fig. 3B, lower panels). Subsequently, samples from the top and bottom factions were denatured, Gag was immunoprecipitated, and equal amounts of Gag were analyzed for ubiquitination by anti-HA immunoblotting. This experiment revealed virtually no ubiquitination in the bottom fraction of the gradient, while a strong HA signal was detected for Gag polyproteins recovered from the membrane fraction (Fig. 3B, upper panel).

If Gag ubiquitination depends on membrane association, a membrane-binding-deficient Gag polyprotein should be ubiquitinated to a lesser extent than wild-type Gag. To test this hypothesis, we made use of a nonmyristoylated (G2A)Gag variant whose membrane binding is reduced to 2% compared to that for wild-type Gag (30). Total Gag was immunoprecipitated from denatured cell lysates of transfected 293T cells and analyzed for ubiquitination by HA immunoblot analysis. This experiment showed a strongly reduced level of ubiquitination for membrane-binding-deficient (G2A)Gag compared to that for wild-type Gag (Fig. 3C).

To test whether ubiquitination of Gag takes place specifically at the plasma membrane, where assembly and budding of HIV Gag normally occur, we tested two Gag variants with altered membrane targeting properties. Deletion of the globular head of MA (pNL43Δp17) and a point mutation within the HIV-1 MA domain (pNL43/87VE) had previously been shown to retarget Gag to intracellular membranes of the endoplasmic reticulum and the MVB, respectively (6, 7, 31). Both mutant polyproteins exhibited highly elevated levels of ubiquitinated Gag but also showed different patterns of ubiquitination (Fig. 3C). While wild-type Gag and budding-arrested wild-type Gag underwent predominant mono- and diubiquitination, a strong signal for polyubiquitinated Gag was observed for the retargeted Gag variants, precluding direct comparison (Fig. 3C). Taken together, the results of these experiments clearly show that HIV-1 Gag ubiquitination depends on membrane association. Due to the dramatic increase of Gag polyubiquitination for the targeting mutants, however, it is currently not clear whether the relevant machinery is confined to the plasma membrane or is present at intracellular membranes as well.

Relative ubiquitination levels of truncated HIV-1 Gag proteins correlate with their membrane-binding capacities.Previously, we have shown that all lysine-containing domains of HIV-1 Gag are monoubiquitinated (13). We now reinvestigated the ubiquitination of MA, CA, MACA (i.e., Gag truncated after the CA domain), and MACA^NTD (i.e., Gag truncated after the N-terminal CA domain) under denaturing conditions. The different proteins were immunoprecipitated from denatured lysates of transfected 293T by cells by using antibodies against CA or MA. The precipitates were adjusted to give approximately equal signal intensities in anti-MA and anti-CA blots (Fig. 4A, right) and were analyzed for ubiquitination by anti-HA immunoblotting (Fig. 4A, left). In agreement with previous data (13), equal ubiquitination signals were observed for full-length Gag and MACA, with much weaker signals for MA and MACA^NTD and virtually no detectable ubiquitination of CA (Fig. 4A, left). The background signal was derived from the rabbit antibody heavy chain used for immunoprecipitation of MA and MACA^NTD.

• Open in new tab
• Download powerpoint

FIG. 4.

Ubiquitination and membrane association of truncated Gag proteins. 293T cells were transfected with pcDNA3-based expression vectors for Rev-independent production of HIV-1 Gag or Gag subdomains as indicated above each lane. (A) Cells were lysed by boiling them in SDS lysis buffer and diluted to RIPA buffer conditions, and HIV-1 specific proteins were immunoprecipitated (IP) with rabbit antiserum against MA (lanes 1 and 3) or sheep antiserum against CA (lanes 2, 4, and 5). Precipitated proteins were normalized to yield comparable signal intensities and analyzed by Western blotting (WB) using a mixture of antisera against MA and CA (right). Ubiquitination was analyzed by Western blotting using antiserum against HA (left). The asterisk marks the background signal of the heavy chain of the rabbit antibody used for the immunoprecipitation of MA and MACA^NTD, which lack the C-terminal CA domain recognized by the sheep anti-CA antiserum. Please note that the ubiquitinated forms of MACA^NTD migrate very close to the heavy chain. (B) The same lysates as those for panel A were subjected to membrane flotation by density gradient centrifugation. Fractions were collected from top to bottom, and equal volumes were analyzed by Western blotting using antisera against MA (panels MA and MACA^NTD) or CA (all other panels). Antibodies against membrane-bound TfR and soluble actin were used as controls. Molecular mass standards are given in kDa. (C) Top: the pixel densities of the HA-Ub Western blot signals of Gag and the different Gag subdomains (A, left panel) were measured (Quantity One software; Bio-Rad) and divided by the signal intensities of the loading control (A, right panel). Bottom: the membrane association levels of the different Gag subdomains were displayed relative to that for wild-type Gag by measuring the pixel intensities of the Western blots shown in Fig. 4B. The band intensity in the floating fraction (fraction 2) was divided through the sum of the signal intensities of the bands in the bottom fractions (fractions 6 to 8).

It appears unlikely that the different number of Ub acceptor lysines accounts for the observed difference, since MACA yielded a much stronger signal than the combined signals for MA and CA. We therefore investigated whether the ubiquitination levels of truncated Gag proteins correlated with their respective membrane-binding capacities. Previous studies had shown much weaker membrane binding for MA than for Gag with approximately 40% of full-length Gag, but only 3% of MA recovered in the membrane fraction (29). This difference in membrane binding is believed to be due to a myristoyl switch masking the hydrophobic properties of MA (41, 49). To correlate the ubiquitination levels of truncated Gag proteins with their respective membrane associations, membrane flotations were performed in parallel to the experiment described above. All MA-containing Gag derivatives were detected in the membrane fraction, albeit to various extents, while CA was completely absent from the membrane fraction (Fig. 4B). Membrane association levels were approximately equal for full-length Gag and MACA, while MA and MACA^NTD showed only weak membrane association. Comparing the relative intensities of the HA-Ub immunoblot signals from Fig. 4A (Fig. 4C, top) with the ratio of membrane-bound to soluble Gag-derived proteins in Fig. 4B (Fig. 4C, bottom) revealed a very similar pattern. These results strongly support our conclusion that Gag ubiquitination is directly linked to membrane association.

The effect of a late-domain mutation on HIV-1 Gag ubiquitination depends on membrane binding.Our collective results indicate that the increased ubiquitination of HIV-1 carrying a mutation in the PTAP motif is due to the increased amount of Gag arrested in the budding process. This conclusion predicts that a PTAP mutation would have no effect on ubiquitination if introduced into a membrane-binding-defective Gag variant. To test this, the PTAP motif was mutated in the context of a myristoylation-deficient Gag protein [(G2V)Gag(PTAP-)] and the effects of the PTAP mutation on Gag ubiquitination were compared for myristoylated and nonmyristoylated Gag. Figure 5A shows that the membrane association of Gag is strongly enhanced upon introduction of a late-domain mutation into wild-type Gag. Very weak membrane association was observed for (G2V)Gag, and this was not affected by the late-domain mutation. A pronounced increase in the ubiquitination of Gag was observed for a PTAP-defective variant of Gag (Fig. 5B, upper panels). Importantly, this effect was completely lost in the context of the G2V variant, which exhibited weak ubiquitination levels, irrespective of a functional late domain. As expected, both PTAP mutation and lack of myristoylation led to severe reductions of virus release (Fig. 5B, lower two panels). This result further strengthens the conclusion that mutation of the PTAP motif influences Gag ubiquitination through alteration of the ratio of membrane-bound to soluble Gag.

• Open in new tab
• Download powerpoint

FIG. 5.

The effect of a PTAP mutation on HIV-1 Gag ubiquitination depends on membrane association. 293T cells were cotransfected with pHA-Ub and pcDNA3-based expression vectors coding for wild-type (WT) or budding-deficient (PTAP-), Rev-independent HIV-1 Gag, either with or without mutations in the myristoylation signal (G2V). (A) Half of the cells were lysed by homogenization and subjected to membrane flotation by density gradient centrifugation. Gradient fractions were collected from top to bottom, subjected to SDS-PAGE, and analyzed by anti-CA Western blotting. (B) The other half of the cells were lysed by boiling them in SDS lysis buffer and diluted to RIPA buffer, and Gag was immunoprecipitated (IP). Equal amounts of Gag were analyzed by Western blotting (WB) using antisera against HA (top panel) or CA (second panel). As a control for the efficiency of Gag release, virus particle preparations from the culture supernatants of the same transfections (third panel) and the Gag contents of the cell lysates (bottom panel) were directly analyzed by Western blotting with antiserum against CA. Gag and Gag-HA-Ub conjugates are identified on the right, and molecular mass standards (in kDa) are shown on the left.