Cumulative Mutations of Ubiquitin Acceptor Sites in Human Immunodeficiency Virus Type 1 Gag Cause a Late Budding Defect
http://www.100md.com
《病菌学杂志》
Abteilung Virologie, Universittsklinikum Heidelberg, D-69120 Heidelberg, Germany
ABSTRACT
The p6 domain of human immunodeficiency virus type 1 (HIV-1) Gag has long been known to be monoubiquitinated. We have previously shown that the MA, CA, and NC domains are also monoubiquitinated at low levels (E. Gottwein and H. G. Krausslich, J. Virol. 79:9134-9144, 2005). While several lines of evidence support a role for ubiquitin in virus release, the relevance of Gag ubiquitination is unclear. To directly address the function of Gag ubiquitination, we constructed Gag variants in which lysine residues in the NC, SP2, and p6 domains were mutated to arginine either in individual domains or in combination. Using these mutants, we showed that in addition to MA, CA, NC, and p6, SP2 is also mono- or diubiquitinated at levels comparable to those of the other domains. Replacement of all lysine residues in only one of the domains had minor effects on virus release, while cumulative mutations in NC and SP2 or in NC and p6 resulted in an accumulation of late budding structures, as observed by electron microscopy analysis. Strikingly, replacement of all lysine residues downstream of CA led to a significant reduction in virus release kinetics and a fivefold accumulation of late viral budding structures compared to wild-type levels. These results indicate that ubiquitination of lysine residues in Gag in the vicinity of the viral late domain is important for HIV-1 budding, while no specific lysine residue may be needed and individual domains can functionally substitute. This is consistent with Gag ubiquitination being functionally involved in a transient protein interaction network at the virus budding site.
INTRODUCTION
Retroviral Gag proteins drive the essential stages of virus assembly, including Gag membrane targeting, RNA encapsidation, virus bud formation, and release of the newly assembled virion by a membrane fission event (2, 16). Cleavage of Gag by the viral protease is closely coupled to virus release and, in the case of human immunodeficiency virus type 1 (HIV-1), results in the release of the Gag domains MA, CA, NC, and p6 as well as of two spacer peptides, SP1 and SP2. The late stages of virus assembly involve the action of so-called late (L) domains within Gag (3, 16). Late domains interact with host cell components of the cellular multivesicular protein-sorting pathway which ultimately mediate the release of the virus from the cell. Well-characterized late domain motifs include the PPXY motif of e.g., Rous sarcoma virus (RSV) and murine leukemia virus, which recruits members of the Nedd4-like ubiquitin ligase family, and the P(T/S)AP motif of, e.g., HIV-1, which functions by interaction with TSG101. Late domain mutation results in the accumulation of virus buds that fail to detach from the host membrane.
Ubiquitin (Ub) has been considered to play a role in retrovirus release, mainly because (i) the late motif PPXY recruits an E3 Ub ligase (12, 14), (ii) the Gag proteins of several different retroviruses are monoubiquitinated at low levels (18-20), (iii) the presence of different late domain motifs within Gag alters the level of Gag ubiquitination (6, 15, 27, 28), (iv) unconjugated Ub is present in the virions of many retroviruses (19, 23), and (v) expression of Ub mutants interferes with virus release (28). Furthermore, the release of many but not all retroviruses is sensitive to proteasome inhibitors which may be caused by disturbing Ub turnover (20-22, 24, 25). The relevance and function of Gag ubiquitination is currently unknown, however. It may be functionally important for budding, but it could also be a bystander effect of recruiting a Ub ligase required for ubiquitination of cellular proteins or may even be nonspecific.
Lysine residues in the proximity of the late domains of HIV-1 (p6) and murine leukemia virus (p12) have been shown to be monoubiquitinated, but mutation of these Ub acceptor residues had no effect on virus release (18). In other systems, however, alternative lysine residues were ubiquitinated when the primary acceptor site had been mutated (10), and this could also apply for retroviruses. In the case of RSV, combined mutation of five lysine residues close to the viral late domain caused a release defect which could be overcome by reintroducing one or more lysine residues into the same or a downstream region (26). If these residues are indeed subject to transient ubiquitin modification, this result would suggest that Gag ubiquitination in the vicinity of the viral late domain is functionally relevant. However, RSV contains a PPXY late domain recruiting a Nedd4 ubiquitin ligase, and it is not clear whether the same would also hold true for other viruses, like HIV, which do not contain a PPXY motif. We recently observed that HIV-1 Gag is monoubiquitinated in several domains outside p6 which may substitute in the case of p6 mutation (6). Here, we extend this analysis and show that combined mutation of all Ub acceptor residues downstream of the CA domain of HIV-1 Gag causes a late budding defect. These observations strongly suggest that ubiquitination of the C-terminal region of Gag contributes to HIV-1 budding, while individual ubiquitination sites are not important and ubiquitination of different domains may functionally substitute for each other.
MATERIALS AND METHODS
Expression constructs, cell culture, and transfections. Hemagglutinin (HA)-Ub was expressed from the previously described plasmid pHA-Ub (6). Lysine mutants are based on the previously described plasmids pNL4-3, pR, and pR/PR(–) (1, 6). All changes were introduced by PCR using mutagenic primers. For mutation of the 5' NC coding region (upstream of the ApaI site), unique SpeI and ApaI or SphI and ApaI sites were used in the case of pNL4-3 or pR, respectively. In all other cases, unique ApaI and SdaI sites were used for introduction of mutated inserts. Detailed cloning procedures and primer sequences are available on request. 293T, HeLaP4, and TZM cells were cultivated as previously described (5, 17). For transfection of HeLaP4 cells, FuGENE6 (ROCHE) was used as instructed. 293T cells were transfected using the calcium phosphate precipitation method. In each case, the culture medium was changed 3 to 6 h after transfection.
Infectivity assay. For infectivity assays, 3 x 105 293T cells were transfected with 2 μg plasmid. One day after transfection, media were harvested and cleared by centrifugation (400 x g, 5 min). Aliquots were adjusted to 0.25% Triton X-100 and subjected to CA antigen enzyme-linked immunosorbent assay (ELISA) as described previously (13). Virus titration on TZM cells was carried out as described previously (17).
Virion preparation and analysis of ubiquitination. Culture media were harvested, cleared by low-speed centrifugation, and filtered through 450-nm filters. Virus was recovered by ultracentrifugation through sucrose cushions (20% [wt/vol] in phosphate-buffered saline [PBS]) at 4°C (Beckman SW28 with 8 ml sucrose, 103,900 x g, 60 min [for detection of Ub conjugates], or Beckman SW60 with 1 ml sucrose, 260,840 x g, 30 min [in the case of pulse-chase experiments]). For external digestion of viral particles with trypsin, sucrose pellets were resuspended in 0.1 mg/ml trypsin (ICN) in PBS and incubated at 30°C for 30 min. The reaction was stopped by addition of protease inhibitors. When virus was lysed with 0.1% Triton X-100 before trypsinization, a complete loss of the viral CA band was observed by Western blotting. For detection of cell-associated Gag-Ub conjugates, cells were lysed in a buffer containing 1% sodium dodecyl sulfate (SDS) at 95°C for 10 min and adjusted to radioimmunoprecipitation assay (RIPA) buffer as described previously (6). Cell lysates were sonicated, cleared by high-speed centrifugation, and subjected to immunoprecipitation with anti-CA antiserum. Immunoprecipitates were normalized for their Gag content (determined by ELISA), and equal amounts were subjected to Western blotting with anti-HA and anti-CA antibodies. CA antigen ELISA of immature Gag was carried out as previously described (6, 13).
Pulse-chase experiments. HeLaP4 cells were transfected with the desired plasmid. Twenty-four hours after transfection, cells were detached using PBS containing 15 mM EDTA (10 min, 37°C), collected by centrifugation, washed with PBS, and starved for 20 min in 1 ml methionine-free medium (ICN) containing 2% fetal calf serum. Cells were collected by centrifugation and pulse labeled for 25 min with [35S]methionine (0.8 mCi; SJ-5050; Amersham) in a small volume. The pulse period was ended by adjusting the medium to 1 mM unlabeled methionine (Sigma), followed by removal of the labeling medium. The cells were washed once with Dulbecco's modified Eagle medium (DMEM) containing 0.4 mM methionine and divided into aliquots, which were chased in DMEM containing 0.4 mM methionine for the indicated period of time (in 6-well dishes). After the chase period, medium was removed and cells that had not reattached were recovered by centrifugation (400 x g, 5 min). The supernatant virion fraction was diluted with PBS and filtered through 450-nm-pore-size filters. Cells that had reattached during the chase period (the large majority) were treated with PBS/15 mM EDTA for 5 min and combined with the pellet of unattached cells. Cells were recovered by centrifugation as described above and lysed in RIPA buffer. Cell lysates were cleared by centrifugation (16,000 x g, 15 min, 4°C). Virions were pelleted through sucrose cushions and lysed in RIPA buffer. The cleared cell lysates and virion lysates were immunoprecipitated with sheep anti-CA antiserum, and immunoprecipitates were resolved by SDS-polyacrylamide gel electrophoresis. Gels were fixed, incubated with Amplify (Amersham), and dried. Labeled viral proteins were quantified using a Bio-Rad Personal FX Phosphorimager and Bio-Rad Quantitiy One software.
Electron microscopy (EM) and statistical analysis. For electron microscopy, 6 x 105 HeLaP4 cells were transfected with 4 μg plasmid DNA. Forty-eight hours after transfection, the medium was aspirated and cells were fixed with 1% glutaraldehyde in 60 mM piperazine-N,N'-bis(2-ethanesulfonic acid), 25 mM HEPES, 2 mM MgCl2, 19 mM EGTA (pH 6.9), which was replaced with fresh fixative after 5 min. After incubation for 30 min at room temperature, samples were stored at 4°C until further processing. The cells were washed 6 times for 5 min each with 100 mM sodium cacodylate (pH 7.2), transferred to ice, scraped into 1 ml 100 mM sodium cacodylate (pH 7.2), and collected by low-speed centrifugation. Cells were resuspended in 2% low-melting-point agarose (37°C; Serva), centrifuged at low speed, and placed on ice. The cell-containing pellet was cut into cubes of less than 1 mm2, which were transferred into 100 mM sodium cacodylate. Embedded cells were incubated with 1% osmium tetroxide and 1.5% potassium hexacyanoferrate in 100 mM cacodylate buffer for 1 h on ice. Cells were washed with water and processed through a graded series of ethanol. After two brief incubations with propylene oxide, cells were stepwise infiltrated with epon, which was polymerized for 2 days at 60°C. We obtained 50-nm sections using a Leica Ultracut S. Sections were stained with 3% uranyl acetate in methanol, followed by Reynold's lead citrate staining. Digital micrographs were taken using a Philips CM120 microscope and Gatan DigitalMicrograph software.
To characterize and quantify the budding phenotypes, transfected cells were identified by the presence of viral profiles at or close to the plasma membrane. For each transfected cell, all virus-related structures were photographed, irrespective of the number of viral particles present (except in cases where the number of viruses exceeded approximately 50, where only a few pictures were taken). To cope with potential experimental variation, cells from three different transfections were analyzed. For each of these, two epon blocks were cut and analyzed. For each block, images from at least 20 individual cells were collected. A second method was chosen to quantify the amount of budding profiles per micrometer of plasma membrane in the case of wild-type (wt) and NCSP2p6(KR) mutant-expressing cells. For this purpose, transfected cells were identified and photographed in a systematic random way. Markers for the negative size imprinted on the microscope screen were used to identify the size of one field of view. The top edge of the cell was oriented into that field. Moving clockwise around the cell, every third field of view was photographed independent of the presence or absence of viral profiles in the field. The plasma membrane length in those pictures was determined by intersection counting as described previously (7). The number of budding profiles per picture was counted, and the total number of budding profiles was related to the length of plasma membrane analyzed.
RESULTS
To address a potential role of HIV-1 Gag ubiquitination in virus release, we mutated a subset of lysine codons in gag. HIV-1 Gag (NL4-3 strain) contains 38 lysine residues (for an overview, see Fig. 1A). Of these, 13 are located in MA (6 in the domain important for membrane binding and targeting of Gag and 7 in the C-terminal part of MA between residues 95 and 114), 11 lie in CA, 10 in NC, and 2 each in SP2 and p6. For this study, the lysine residues in the NC, SP2, and p6 domains were targeted due to their proximity to the described HIV-1 Gag late domain motifs (16).
A series of plasmids carrying Lys-to-Arg mutations was constructed in the context of the previously described HIV-1 expression plasmids pR and pNL4-3 (1, 6). The resulting plasmids were termed according to the mutated region and the type of mutation. In the NC(KR) mutant, all 10 Lys codons in NC were mutated to Arg; in the SP2(KR) mutant, the 2 Lys codons in SP2 were mutated to Arg; and in p6(KR) the 2 Lys codons in p6 were mutated to Arg. Lys mutations in two domains were combined to give the constructs NCSP2(KR), NCp6(KR), and SP2p6(KR). Finally, a mutant that lacks all lysine codons downstream of CA [NCSP2p6(KR)] was constructed.
Detection of ubiquitinated viral proteins in mature HIV-1. We first examined the pattern of mature ubiquitinated Gag proteins within wt and mutant virions produced from pR-derived plasmids. pR is a full-length HIV-1 construct that results in the production of noninfectious virus due to mutations in the reverse transcriptase active site and in the region encompassing the primer binding site. 293T cells were transfected with either pHA-Ub alone or pR alone or were cotransfected with pHA-Ub and wt or lysine mutant pR constructs. Forty-eight hours after transfection, virus was harvested and pelleted through sucrose cushions. To remove soluble contaminants, the virion preparations were subjected to trypsin digestion. Equal amounts of virus (as determined by ELISA) were resolved by SDS-polyacrylamide gel electrophoresis and analyzed by Western blotting using anti-HA antibodies (Fig. 1B, upper panel).
When pHA-Ub or pR was expressed alone, no HA-reactive bands were detected by Western blotting (Fig. 1B, lanes 1 and 2). Upon cotransfection of pHA-Ub and pR, bands for several ubiquitinated proteins were observed (lanes 3 and 4). One of these bands was lost upon trypsin treatment (compare lane 4 to lane 3). This contaminating band was sometimes also found in the supernatant of cells expressing HA-Ub alone. Ubiquitinated MA and CA proteins were assigned based on their comigration with immunoprecipitated HA-Ub-conjugated MA and CA expressed from pcDNA3-based MA and CA expression constructs (6 and data not shown). In addition to ubiquitinated MA and CA, several ubiquitinated proteins migrating less than 20 kDa were detected, and they could be largely assigned due to their apparent mass and their absence in the case of specific Gag variants (lanes 5 to 11). The product identified as NC-HA-Ub was strongly decreased in intensity in lanes where Gag proteins lacking lysine residues in NC were analyzed (lanes 5, 8, 9, and 11). A faint band migrating at the same position as NC-HA-Ub was still detected in these cases, but this product was also observed in preparations of protease-negative viruses and is therefore not Gag derived (data not shown). The product designated SP2-HA-Ub was completely lost in lanes 6, 8, 10, and 11, where pR constructs carrying mutations of the two Lys codons in SP2 were analyzed. Its size is consistent with the modification of SP2 by a single Ub moiety. The band designated SP2-(HA-Ub)2 was also lost when the lysine codons in SP2 were mutated, and its size is consistent with modification of SP2 by two Ub moieties. Since SP2 contains two lysine residues, either both residues could be monoubiquitinated or one residue could be modified by a diubiquitin chain. Ubiquitinated p6 was harder to detect in mature virions. Upon long exposure, a product migrating slightly faster than NC-HA-Ub was detected (data not shown). This product was lost upon mutation of p6 Lys residues and presumably represents ubiquitinated p6. In conclusion, all lysine-containing Gag domains including the spacer peptide SP2 are found to be ubiquitinated in mature HIV-1.
Levels of ubiquitination of wt and mutant Gag polyproteins. We next tested whether the overall level or the pattern of Gag ubiquitination was altered upon replacement of C-terminal Gag lysine residues. For this purpose, the set of mutations was introduced into pR/PR(–) which carries a mutation in the protease active site (6). Transfection of pR/PR(–) therefore yields immature virus-like particles containing only the uncleaved Gag polyprotein and no processed products. 293T cells were cotransfected with one of these constructs and a HA-Ub expression vector. Forty-eight hours after transfection, cell lysates were prepared under denaturing conditions, Gag proteins were immunoprecipitated, and the extent of wt and mutant Gag ubiquitination was examined by Western blotting with an anti-HA antibody (Fig. 2). In the case of the wt construct, a ladder of Gag-reactive products was observed as described previously (Fig. 2, lane 1) (6). The intensity of ubiquitination was markedly increased when lysine residues in NC were lacking (lanes 2, 5, and 6), while mutation of lysine codons in p6 reduced overall ubiquitination of Gag (lane 4). Several mutations led to an altered pattern of Ub-modified Gag polyproteins. This was particularly evident for the p6(KR) mutation, which caused the loss of one of the ubiquitinated Gag products (compare lanes 4 and 6 to 8 with the other lanes). This result indicated that the different Gag-reactive products correspond, at least in part, to different isomers of monoubiquitinated Gag, suggesting the use of different Ub attachment sites. Mutation of all lysine codons downstream of the CA coding region [NCSP2p6(KR)] resulted in diminished levels of Gag ubiquitination and a loss of specific ubiquitinated products as described above. This result suggests that the loss of C-terminal ubiquitination is not met by a compensatory increase in ubiquitination of other Gag domains.
Steady-state levels of release and infectivity. We next evaluated virus release at steady state as well as viral infectivity for all mutants. For this, 293T or HeLaP4 cells were transfected with wt pNL4-3 or pNL4-3 constructs carrying the mutations described above. The virus-containing supernatant was harvested 24 h after transfection, cleared by centrifugation, and analyzed by CA ELISA as well as by titration on TZM indicator cells (Fig. 3). Results are shown for 293T cell transfections but were similar when HeLaP4 cells were used (data not shown). Antigen release (numbers shown at the bottom of Fig. 3) as well as the pattern of cell- and virus-associated Gag-derived proteins (data not shown) were similar for all constructs, suggesting no pronounced effect of lysine mutation on the amount of virus that is ultimately produced from transfected cells. Up to twofold-reduced levels of antigen were often, but not always, observed in the case of NCSP2p6(KR) mutant virus (Fig. 3). Infectious titers were determined by counting infected cells and were normalized for the amount of input virus (measured by ELISA) (Fig. 3). The NC(KR) mutant was reproducibly 10-fold less infectious than wt virus. Mutation of the lysine codons in SP2 or p6 alone, as well as the combination of both mutations [SP2p6(KR)], had no effect on infectivity. Adding the p6 and NC mutations also did not lead to a further reduction of viral infectivity compared to the NC mutant alone. In contrast, adding the SP2 and NC mutations [NCSP2(KR)] caused a further twofold reduction in infectivity. Interestingly, the NCSP2p6(KR) mutant was 5- to 10-fold less infectious than the NC mutant (43- to 100-fold less infectious than the wt), indicating that combination of the different mutations had a more than additive effect.
Virus release kinetics. Since replacement of lysine residues did not change virus production from transfected cells at steady state, we performed pulse-chase experiments to test whether virus release kinetics were altered. This would be expected if lysine mutations cause a retardation of budding rather than a complete block to virus release. HeLaP4 cells were pulse labeled 24 h after transfection with wt or mutant pR, and cells and virus particles were harvested after a 2.5-h chase period. Levels of intracellular wt or mutant Gag-derived proteins were similar in all cases (Fig. 4A, upper panel). A difference in Gag processing was observed for all constructs harboring NC lysine mutations. Processing of SP1 from the C terminus of CA was reduced (leading to an increased ratio of CA-SP1 to CA), and a CA-related product with an apparent molecular mass of 30 kDa was greatly diminished (Fig. 4A, upper panel, lanes 2, 5, 6, and 8). The identity of this band is currently not known, but its size matches the predicted molecular mass of a CA-SP1-NC-processing intermediate. The levels of virus release were normal for all mutants except for NCSP2p6(KR), which exhibited a strong reduction in virus release (Fig. 4A, lower panel, compare lane 8 to lanes 1 to 7). Importantly, no such reduction was observed for the constructs harboring mutations in any one domain or combination of two domains. Thus, the effects of the NC(KR) mutation on Gag polyprotein processing and of the combined NCSP2p6(KR) mutation on release are genetically separable, and lysine residues in NC, SP2, and p6 appear to cooperatively contribute to virus release.
For a more detailed analysis, the release kinetics were quantified for wt and NCSP2p6(KR) over a 6-h chase period. The efficiencies of wt and mutant virus release were calculated as the virus-derived CA-related signal divided by the sum of cell- and virus-derived CA-related signal. This quantification revealed that the budding efficiency of NCSP2p6(KR) mutant virus is approximately threefold lower compared to that of the wt (Fig. 4B). Similar results were obtained when infectious pNL4-3-based wt and mutant constructs were used (data not shown). Analysis of the virus release kinetics of the other mutants confirmed that only NCSP2p6(KR) exhibits slower virus release (data not shown).
Ultrastructure of Gag lysine mutant virus release. In order to directly evaluate wt and mutant virus budding, transfected HeLaP4 cells were examined by thin-section electron microscopy. wt virus-expressing cells revealed a mixture of free, mature virus, identified by the condensed conical core, and virus-budding profiles at the plasma membrane (Fig. 5a). Only a few immature virions were detected in this case. In NC mutant samples, on the other hand, virions were also released but remained largely immature (Fig. 5b). It appears likely that this defect in maturation corresponds to the reduced processing at the CA-SP1 site, since removal of SP1 is essential for core maturation (29). Of those virions that did mature, a significant percentage contained aberrant core structures (Fig. 5d and data not shown), indicating that the substitution of arginine for all lysine residues in NC also causes a defect in core maturation. No ultrastructural defects were observed for viruses harboring mutations in SP2, p6, or a combination of the two (data not shown). This result correlates with their wt phenotype regarding virus release and infectivity.
When the NC(KR) mutation was combined with mutation of either SP2 or p6 lysine residues [NCSP2(KR) and NCp6(KR), respectively], a budding defect became evident (Fig. 5c and d). Inspection of electron micrographs revealed more viral structures in late stages of budding in these cases than that observed for either wt, NC(KR), or SP2p6(KR). This increase in budding-arrested structures was observed in both cases, while an increase of "double buds" (late budding structures composed of two confluent budding sites) was only observed in NCSP2(KR)-transfected samples (Fig. 5c). A more pronounced late budding defect was seen for the NCSP2p6(KR) mutant, where all lysine residues downstream of the CA domain have been removed (Fig. 6). Many late budding structures often lining the plasma membrane (Fig. 6a to c) and connected to the cell by a thin membrane stalk (Fig. 6d) were observed. This phenotype is very reminiscent of observations made for mutants of the viral late domain.
Quantitative EM analysis of budding. To validate the impression from visual inspection of electron micrographs, we sought to quantify the budding properties of lysine mutation variants. For this, we separately counted released mature and immature virions as well as budding structures that were detected at or beyond the plane of the plasma membrane for the relevant subset of mutants (Fig. 7A). In each case, a virus-expressing cell was first located, and then all events were documented by digital imaging and counted without preselection. This procedure was repeated for different cells and different experiments as detailed in Materials and Methods and in the legend to Fig. 7. The relative percentages of released (both immature and mature) viruses and cell-attached virus buds is shown in Fig. 7A. This representation of the results was chosen because the number of total viral structures differed in each experiment due to variations in numbers of both particles and transfected cells, and calculating the average may thus give biased results. To avoid such bias, we calculated the Pearson product-moment correlation coefficient (measuring the degree of linear relationship between two variables) for the number of free viruses compared to total viruses. This correlation was linear in all cases, and we therefore depict the percentage of free viruses using a linear regression. Figure 7A plots the predicted number of free viruses per 100 viruses and the predicted number of budding structures (100 minus the percentage of free viruses). A similar result was obtained when actual numbers were averaged, with the deviation from the calculated result being <2% in each case (data not shown).
Quantification of free versus cell-attached viral structures confirmed that replacement of lysine residues in the C-terminal domains of Gag has additive effects on virus budding. In the case of wt-transfected cells, 82.5% of viral structures corresponded to cell-free virus, and 17.5% of viral structures corresponded to budding structures (Fig. 7A). The relative amount of free virus was slightly reduced for the NC(KR) (65%) and SP2p6(KR) (70%) variants, suggesting that they exhibit a subtle impairment in budding which is only detected by careful statistical analysis. The percentage of released virions was further decreased when all lysine residues were mutated in both NC and SP2 (60%) or in both NC and p6 (50%). When all lysine residues downstream of CA were removed [NCSP2p6(KR)], only 46% of virus structures corresponded to released virus, while 54% were tethered to the plasma membrane. In this quantification, we also distinguished immature and mature free virions, early and late budding structures, and single and chained buds. The results confirmed the pronounced maturation defect of the NC(KR) mutant, with 55% of the released virions being immature compared to only 8% in the case of the wt. Interestingly, SP2p6(KR) mutant free virions were also less mature than wt virions (30% immature free virions versus 8% in the wt). In the case of NCSP2p6(KR) mutant virions, 75% of released virions were immature. Analysis of the relative abundance of early and late budding showed no enrichment of early buds in any of the mutants. Statistical analysis confirmed that more "double" buds were present in the case of NCSP2(KR) mutant samples (Fig. 5c). While no chain buds or fused buds were detected in wt-transfected cells and 2.7% of the budding structures were in chains in the NC(KR) mutant samples, 21% of the budding structures detected in NCSP2(KR)-transfected cells were not separated from each other. Interestingly, this accumulation of double buds was only found for NCSP2(KR) and not in the other cases [including NCSP2p6(KR)].
A second method of quantification was applied in order to accurately determine the late budding defect of the NCSP2p6(KR) mutant. In this case, the number of budding profiles per micrometer of plasma membrane length was calculated for wt- and NCSP2p6(KR) mutant-transfected cells as described in Material and Methods (Fig. 7B). This analysis revealed that the density of budding structures was ca. fivefold higher in the case of the lysine mutation variant than that of the wt, thus confirming the observed late release block.
DISCUSSION
In this report, we describe a set of HIV-1 Gag C-terminal Lys-to-Arg mutants, which were analyzed in terms of infectivity, Gag ubiquitination, and virus release. Using these mutants, we extended our previous finding that MA, CA, and NC can be monoubiquitinated by showing that the spacer peptide SP2 is also mono- or diubiquitinated at low levels. Surprisingly, mutations of lysine residues in NC or SP2 increased ubiquitination of the Gag polyprotein in cell lysates (Fig. 2), while an overall increase in ubiquitination of any Gag-derived product was not observed in mature virions of these mutants (Fig. 1B). Although the reason for these observations is unknown, selective deubiquitination of Gag might take place during or after virus release, or "hyper-ubiquitinated" Gag proteins might be excluded from virions. Alteration of all Lys residues to Arg in any single domain downstream of CA had little or no effect on virus release. However, upon cumulative replacement of lysine residues in two or three of these domains, virus budding was reduced.
At this point, we cannot rule out that other lysine modifications, such as sumoylation, contribute to the observed phenotypes. However, no consensus sumoylation motif outside the p6 domain, which was previously reported to be sumoylated, is present within Gag (8, 11). While it seems very likely that NC(KR) mutant Gag has defects in addition to abolished Gag ubiquitination, this variant was released at essentially wt kinetics in biochemical experiments and with an efficiency similar to that of SP2p6 mutant virus as judged by EM. Therefore, the effect of lysine replacements in NC on polyprotein processing and maturation (and consequently on infectivity), which was not addressed in this report, can be clearly distinguished from the cumulative effects of lysine replacement in C-terminal Gag domains on virus budding. Taken together, our data strongly suggest that ubiquitination of Gag is indeed important for virus budding and is not a functionally irrelevant bystander effect.
The budding phenotype of NCSP2p6 mutant Gag was not readily observed under steady-state conditions at the high-expression level of transfected 293T cells, but it became obvious in pulse-chase experiments and electron micrographs. This suggests that the kinetics of membrane envelopment are altered, while budding is not prevented. Our unpublished results show that the expression level of Gag in transfected 293T cells is comparable to that in infected T-cell lines, suggesting that this is not an artifact of overexpression. One can speculate, however, that ubiquitination of lysine residues may be even more relevant in cells with lower levels of expression of Gag (e.g., primary T cells). The requirement for cumulative mutations to induce a budding defect suggests that no single specific ubiquitination site is necessary for this function and that ubiquitination of the different C-terminal domains may substitute for each other. Similarly, it had previously been shown that several normally available or newly introduced lysine residues, e.g., in the T-cell receptor -subunit, can be ubiquitinated when other sites are lost. Furthermore, the addition of extra lysine residues has been shown to rescue the budding defect of a lysine-mutated RSV Gag variant (10, 26). It is unclear at present whether the observed effect of the NCSP2p6(KR) mutation on budding would be even stronger if Ub acceptor sites in upstream domains were replaced in addition. This appears less likely, however, because Gag proteins lacking lysine residues downstream of CA did not show a compensatory induction of novel Gag-Ub conjugates while retaining ubiquitination at relatively low level. To test whether ubiquitination of NCSP2p6(KR) mutant Gag could be restored by reintroducing potential acceptor sites into the C-terminal part of Gag, we mutated three Arg codons to Lys in the NC domain of NCSP2p6(KR). These mutations led to an altered pattern of Gag ubiquitination, but overall ubiquitination of Gag was not increased (data not shown). This may be due to structural constraints preventing these sites from being used as ubiquitin acceptors or may suggest the existence of subtle position-dependent differences that were not detected in our general domain analysis.
The cumulative effect of Lys mutations is consistent with Ub acting as a transient protein modification that facilitates protein-protein interactions. Such a model would be in analogy to the role of Ub-based protein interactions during endocytosis. In this case, the sorting machinery is connected through a multitude of weak protein-protein interactions (9). Several different proteins of this network, cargo as well as endocytic machinery, are thought to become ubiquitinated in order to interact with proteins containing ubiquitin binding domains (UBDs). In several instances, UBDs carrying proteins themselves can be ubiquitinated. Since the interactions between UBDs and Ub are typically of low affinity, several such interactions are required to result in a transiently stable network. Our data are entirely consistent with a model in which Gag ubiquitination in the vicinity of the late domain would functionally contribute to such a network. Neither Gag ubiquitination nor any of the hypothetical Ub-based interactions of other components would be individually essential, while the overall number of interacting Ub moieties is. Of note, several proteins known to play important roles in Gag budding are ubiquitinated and/or contain UBDs (16). In vitro, the fusion of p6 to ubiquitin enhanced its interaction with TSG101, which contains a ubiquitin E2 variant-type ubiquitin binding domain (4, 16). Such a model would also be consistent with mutation of the Ub ligase recruiting late domain motifs in the case of PPXY-containing viruses having a much more pronounced effect on release than cumulative mutation of Gag lysine residues, and this needs to be experimentally addressed in the future.
ACKNOWLEDGMENTS
During this study, we made extensive use of the EM facility at the EMBL, Heidelberg, Germany. We thank Claude Antony and his team for hosting us as well as for their support throughout the project. We also thank Gareth Griffiths for ideas and comments concerning the quantification of EM data, Eric Freed and David Ott for expression constructs, and Piyush Kumar for statistical analysis of the data presented in Fig. 7A.
This work was supported in part by a grant from the Deutsche Forschungsgemeinschaft (SFB638, project A9). E.G. is the recipient of a stipend from the Boehringer Ingelheim Fonds.
Present address: Department of Microbiology and Genetics, Box 3025, Duke University Medical Center, Durham, NC 27710.
REFERENCES
Adachi, A., H. E. Gendelman, S. Koenig, T. Folks, R. Willey, A. Rabson, and M. A. Martin. 1986. Production of acquired immunodeficiency syndrome-associated retrovirus in human and nonhuman cells transfected with an infectious molecular clone. J. Virol. 59:284-291.
Freed, E. O. 1998. HIV-1 gag proteins: diverse functions in the virus life cycle. Virology 251:1-15.
Freed, E. O. 2002. Viral late domains. J. Virol. 76:4679-4687.
Garrus, J. E., U. K. von Schwedler, O. W. Pornillos, S. G. Morham, K. H. Zavitz, H. E. Wang, D. A. Wettstein, K. M. Stray, M. Cote, R. L. Rich, D. G. Myszka, and W. I. Sundquist. 2001. Tsg101 and the vacuolar protein sorting pathway are essential for HIV-1 budding. Cell 107:55-65.
Gottwein, E., J. Bodem, B. Muller, A. Schmechel, H. Zentgraf, and H. G. Krausslich. 2003. The Mason-Pfizer monkey virus PPPY and PSAP motifs both contribute to virus release. J. Virol. 77:9474-9485.
Gottwein, E., and H. G. Krausslich. 2005. Analysis of human immunodeficiency virus type 1 Gag ubiquitination. J. Virol. 79:9134-9144.
Griffiths, G. 1993. Fine structure immunocytochemistry. Springer-Verlag, Berlin, Germany.
Gurer, C., L. Berthoux, and J. Luban. 2005. Covalent modification of human immunodeficiency virus type 1 p6 by SUMO-1. J. Virol. 79:910-917.
Hicke, L., H. L. Schubert, and C. P. Hill. 2005. Ubiquitin-binding domains. Nat. Rev. Mol. Cell Biol. 6:610-621.
Hou, D., C. Cenciarelli, J. P. Jensen, H. B. Nguygen, and A. M. Weissman. 1994. Activation-dependent ubiquitination of a T cell antigen receptor subunit on multiple intracellular lysines. J. Biol. Chem. 269:14244-14247.
Johnson, E. S. 2004. Protein modification by SUMO. Annu. Rev. Biochem. 73:355-382.
Kikonyogo, A., F. Bouamr, M. L. Vana, Y. Xiang, A. Aiyar, C. Carter, and J. Leis. 2001. Proteins related to the Nedd4 family of ubiquitin protein ligases interact with the L domain of Rous sarcoma virus and are required for gag budding from cells. Proc. Natl. Acad. Sci. USA 98:11199-11204.
Konvalinka, J., M. A. Litterst, R. Welker, H. Kottler, F. Rippmann, A. M. Heuser, and H. G. Krausslich. 1995. An active-site mutation in the human immunodeficiency virus type 1 proteinase (PR) causes reduced PR activity and loss of PR-mediated cytotoxicity without apparent effect on virus maturation and infectivity. J. Virol. 69:7180-7186.
Martin-Serrano, J., S. W. Eastman, W. Chung, and P. D. Bieniasz. 2005. HECT ubiquitin ligases link viral and cellular PPXY motifs to the vacuolar protein-sorting pathway. J. Cell Biol. 168:89-101.
Martin-Serrano, J., D. Perez-Caballero, and P. D. Bieniasz. 2004. Context-dependent effects of L domains and ubiquitination on viral budding. J. Virol. 78:5554-5563.
Morita, E., and W. I. Sundquist. 2004. Retrovirus budding. Annu. Rev. Cell Dev. Biol. 20:395-425.
Muller, B., J. Daecke, O. T. Fackler, M. T. Dittmar, H. Zentgraf, and H. G. Krausslich. 2004. Construction and characterization of a fluorescently labeled infectious human immunodeficiency virus type 1 derivative. J. Virol. 78:10803-10813.
Ott, D. E., L. V. Coren, E. N. Chertova, T. D. Gagliardi, and U. Schubert. 2000. Ubiquitination of HIV-1 and MuLV Gag. Virology 278:111-121.
Ott, D. E., L. V. Coren, T. D. Copeland, B. P. Kane, D. G. Johnson, R. C. Sowder, Jr., Y. Yoshinaka, S. Oroszlan, L. O. Arthur, and L. E. Henderson. 1998. Ubiquitin is covalently attached to the p6Gag proteins of human immunodeficiency virus type 1 and simian immunodeficiency virus and to the p12Gag protein of Moloney murine leukemia virus. J. Virol. 72:2962-2968.
Ott, D. E., L. V. Coren, R. C. Sowder, Jr., J. Adams, and U. Schubert. 2003. Retroviruses have differing requirements for proteasome function in the budding process. J. Virol. 77:3384-3393.
Patnaik, A., V. Chau, F. Li, R. C. Montelaro, and J. W. Wills. 2002. Budding of equine infectious anemia virus is insensitive to proteasome inhibitors. J. Virol. 76:2641-2647.
Patnaik, A., V. Chau, and J. W. Wills. 2000. Ubiquitin is part of the retrovirus budding machinery. Proc. Natl. Acad. Sci. USA 97:13069-13074.
Putterman, D., R. B. Pepinsky, and V. M. Vogt. 1990. Ubiquitin in avian leukosis virus particles. Virology 176:633-637.
Schubert, U., D. E. Ott, E. N. Chertova, R. Welker, U. Tessmer, M. F.Princiotta, J. R. Bennink, H. G. Krausslich, and J. W. Yewdell. 2000. Proteasome inhibition interferes with gag polyprotein processing, release, and maturation of HIV-1 and HIV-2. Proc. Natl. Acad. Sci. USA 97:13057-13062.
Shehu-Xhilaga, M., S. Ablan, D. G. Demirov, C. Chen, R. C. Montelaro, and E. O. Freed. 2004. Late domain-dependent inhibition of equine infectious anemia virus budding. J. Virol. 78:724-732.
Spidel, J. L., R. C. Craven, C. B. Wilson, A. Patnaik, H. Wang, L. M. Mansky, and J. W. Wills. 2004. Lysines close to the Rous sarcoma virus late domain critical for budding. J. Virol. 78:10606-10616.
Strack, B., A. Calistri, M. A. Accola, G. Palu, and H. G. Gottlinger. 2000. A role for ubiquitin ligase recruitment in retrovirus release. Proc. Natl. Acad. Sci. USA 97:13063-13068.
Strack, B., A. Calistri, and H. G. Gottlinger. 2002. Late assembly domain function can exhibit context dependence and involves ubiquitin residues implicated in endocytosis. J. Virol. 76:5472-5479.
Wiegers, K., G. Rutter, H. Kottler, U. Tessmer, H. Hohenberg, and H. G. Krausslich. 1998. Sequential steps in human immunodeficiency virus particle maturation revealed by alterations of individual Gag polyprotein cleavage sites. J. Virol. 72:2846-2854.(Eva Gottwein, Stefanie Jg)
ABSTRACT
The p6 domain of human immunodeficiency virus type 1 (HIV-1) Gag has long been known to be monoubiquitinated. We have previously shown that the MA, CA, and NC domains are also monoubiquitinated at low levels (E. Gottwein and H. G. Krausslich, J. Virol. 79:9134-9144, 2005). While several lines of evidence support a role for ubiquitin in virus release, the relevance of Gag ubiquitination is unclear. To directly address the function of Gag ubiquitination, we constructed Gag variants in which lysine residues in the NC, SP2, and p6 domains were mutated to arginine either in individual domains or in combination. Using these mutants, we showed that in addition to MA, CA, NC, and p6, SP2 is also mono- or diubiquitinated at levels comparable to those of the other domains. Replacement of all lysine residues in only one of the domains had minor effects on virus release, while cumulative mutations in NC and SP2 or in NC and p6 resulted in an accumulation of late budding structures, as observed by electron microscopy analysis. Strikingly, replacement of all lysine residues downstream of CA led to a significant reduction in virus release kinetics and a fivefold accumulation of late viral budding structures compared to wild-type levels. These results indicate that ubiquitination of lysine residues in Gag in the vicinity of the viral late domain is important for HIV-1 budding, while no specific lysine residue may be needed and individual domains can functionally substitute. This is consistent with Gag ubiquitination being functionally involved in a transient protein interaction network at the virus budding site.
INTRODUCTION
Retroviral Gag proteins drive the essential stages of virus assembly, including Gag membrane targeting, RNA encapsidation, virus bud formation, and release of the newly assembled virion by a membrane fission event (2, 16). Cleavage of Gag by the viral protease is closely coupled to virus release and, in the case of human immunodeficiency virus type 1 (HIV-1), results in the release of the Gag domains MA, CA, NC, and p6 as well as of two spacer peptides, SP1 and SP2. The late stages of virus assembly involve the action of so-called late (L) domains within Gag (3, 16). Late domains interact with host cell components of the cellular multivesicular protein-sorting pathway which ultimately mediate the release of the virus from the cell. Well-characterized late domain motifs include the PPXY motif of e.g., Rous sarcoma virus (RSV) and murine leukemia virus, which recruits members of the Nedd4-like ubiquitin ligase family, and the P(T/S)AP motif of, e.g., HIV-1, which functions by interaction with TSG101. Late domain mutation results in the accumulation of virus buds that fail to detach from the host membrane.
Ubiquitin (Ub) has been considered to play a role in retrovirus release, mainly because (i) the late motif PPXY recruits an E3 Ub ligase (12, 14), (ii) the Gag proteins of several different retroviruses are monoubiquitinated at low levels (18-20), (iii) the presence of different late domain motifs within Gag alters the level of Gag ubiquitination (6, 15, 27, 28), (iv) unconjugated Ub is present in the virions of many retroviruses (19, 23), and (v) expression of Ub mutants interferes with virus release (28). Furthermore, the release of many but not all retroviruses is sensitive to proteasome inhibitors which may be caused by disturbing Ub turnover (20-22, 24, 25). The relevance and function of Gag ubiquitination is currently unknown, however. It may be functionally important for budding, but it could also be a bystander effect of recruiting a Ub ligase required for ubiquitination of cellular proteins or may even be nonspecific.
Lysine residues in the proximity of the late domains of HIV-1 (p6) and murine leukemia virus (p12) have been shown to be monoubiquitinated, but mutation of these Ub acceptor residues had no effect on virus release (18). In other systems, however, alternative lysine residues were ubiquitinated when the primary acceptor site had been mutated (10), and this could also apply for retroviruses. In the case of RSV, combined mutation of five lysine residues close to the viral late domain caused a release defect which could be overcome by reintroducing one or more lysine residues into the same or a downstream region (26). If these residues are indeed subject to transient ubiquitin modification, this result would suggest that Gag ubiquitination in the vicinity of the viral late domain is functionally relevant. However, RSV contains a PPXY late domain recruiting a Nedd4 ubiquitin ligase, and it is not clear whether the same would also hold true for other viruses, like HIV, which do not contain a PPXY motif. We recently observed that HIV-1 Gag is monoubiquitinated in several domains outside p6 which may substitute in the case of p6 mutation (6). Here, we extend this analysis and show that combined mutation of all Ub acceptor residues downstream of the CA domain of HIV-1 Gag causes a late budding defect. These observations strongly suggest that ubiquitination of the C-terminal region of Gag contributes to HIV-1 budding, while individual ubiquitination sites are not important and ubiquitination of different domains may functionally substitute for each other.
MATERIALS AND METHODS
Expression constructs, cell culture, and transfections. Hemagglutinin (HA)-Ub was expressed from the previously described plasmid pHA-Ub (6). Lysine mutants are based on the previously described plasmids pNL4-3, pR, and pR/PR(–) (1, 6). All changes were introduced by PCR using mutagenic primers. For mutation of the 5' NC coding region (upstream of the ApaI site), unique SpeI and ApaI or SphI and ApaI sites were used in the case of pNL4-3 or pR, respectively. In all other cases, unique ApaI and SdaI sites were used for introduction of mutated inserts. Detailed cloning procedures and primer sequences are available on request. 293T, HeLaP4, and TZM cells were cultivated as previously described (5, 17). For transfection of HeLaP4 cells, FuGENE6 (ROCHE) was used as instructed. 293T cells were transfected using the calcium phosphate precipitation method. In each case, the culture medium was changed 3 to 6 h after transfection.
Infectivity assay. For infectivity assays, 3 x 105 293T cells were transfected with 2 μg plasmid. One day after transfection, media were harvested and cleared by centrifugation (400 x g, 5 min). Aliquots were adjusted to 0.25% Triton X-100 and subjected to CA antigen enzyme-linked immunosorbent assay (ELISA) as described previously (13). Virus titration on TZM cells was carried out as described previously (17).
Virion preparation and analysis of ubiquitination. Culture media were harvested, cleared by low-speed centrifugation, and filtered through 450-nm filters. Virus was recovered by ultracentrifugation through sucrose cushions (20% [wt/vol] in phosphate-buffered saline [PBS]) at 4°C (Beckman SW28 with 8 ml sucrose, 103,900 x g, 60 min [for detection of Ub conjugates], or Beckman SW60 with 1 ml sucrose, 260,840 x g, 30 min [in the case of pulse-chase experiments]). For external digestion of viral particles with trypsin, sucrose pellets were resuspended in 0.1 mg/ml trypsin (ICN) in PBS and incubated at 30°C for 30 min. The reaction was stopped by addition of protease inhibitors. When virus was lysed with 0.1% Triton X-100 before trypsinization, a complete loss of the viral CA band was observed by Western blotting. For detection of cell-associated Gag-Ub conjugates, cells were lysed in a buffer containing 1% sodium dodecyl sulfate (SDS) at 95°C for 10 min and adjusted to radioimmunoprecipitation assay (RIPA) buffer as described previously (6). Cell lysates were sonicated, cleared by high-speed centrifugation, and subjected to immunoprecipitation with anti-CA antiserum. Immunoprecipitates were normalized for their Gag content (determined by ELISA), and equal amounts were subjected to Western blotting with anti-HA and anti-CA antibodies. CA antigen ELISA of immature Gag was carried out as previously described (6, 13).
Pulse-chase experiments. HeLaP4 cells were transfected with the desired plasmid. Twenty-four hours after transfection, cells were detached using PBS containing 15 mM EDTA (10 min, 37°C), collected by centrifugation, washed with PBS, and starved for 20 min in 1 ml methionine-free medium (ICN) containing 2% fetal calf serum. Cells were collected by centrifugation and pulse labeled for 25 min with [35S]methionine (0.8 mCi; SJ-5050; Amersham) in a small volume. The pulse period was ended by adjusting the medium to 1 mM unlabeled methionine (Sigma), followed by removal of the labeling medium. The cells were washed once with Dulbecco's modified Eagle medium (DMEM) containing 0.4 mM methionine and divided into aliquots, which were chased in DMEM containing 0.4 mM methionine for the indicated period of time (in 6-well dishes). After the chase period, medium was removed and cells that had not reattached were recovered by centrifugation (400 x g, 5 min). The supernatant virion fraction was diluted with PBS and filtered through 450-nm-pore-size filters. Cells that had reattached during the chase period (the large majority) were treated with PBS/15 mM EDTA for 5 min and combined with the pellet of unattached cells. Cells were recovered by centrifugation as described above and lysed in RIPA buffer. Cell lysates were cleared by centrifugation (16,000 x g, 15 min, 4°C). Virions were pelleted through sucrose cushions and lysed in RIPA buffer. The cleared cell lysates and virion lysates were immunoprecipitated with sheep anti-CA antiserum, and immunoprecipitates were resolved by SDS-polyacrylamide gel electrophoresis. Gels were fixed, incubated with Amplify (Amersham), and dried. Labeled viral proteins were quantified using a Bio-Rad Personal FX Phosphorimager and Bio-Rad Quantitiy One software.
Electron microscopy (EM) and statistical analysis. For electron microscopy, 6 x 105 HeLaP4 cells were transfected with 4 μg plasmid DNA. Forty-eight hours after transfection, the medium was aspirated and cells were fixed with 1% glutaraldehyde in 60 mM piperazine-N,N'-bis(2-ethanesulfonic acid), 25 mM HEPES, 2 mM MgCl2, 19 mM EGTA (pH 6.9), which was replaced with fresh fixative after 5 min. After incubation for 30 min at room temperature, samples were stored at 4°C until further processing. The cells were washed 6 times for 5 min each with 100 mM sodium cacodylate (pH 7.2), transferred to ice, scraped into 1 ml 100 mM sodium cacodylate (pH 7.2), and collected by low-speed centrifugation. Cells were resuspended in 2% low-melting-point agarose (37°C; Serva), centrifuged at low speed, and placed on ice. The cell-containing pellet was cut into cubes of less than 1 mm2, which were transferred into 100 mM sodium cacodylate. Embedded cells were incubated with 1% osmium tetroxide and 1.5% potassium hexacyanoferrate in 100 mM cacodylate buffer for 1 h on ice. Cells were washed with water and processed through a graded series of ethanol. After two brief incubations with propylene oxide, cells were stepwise infiltrated with epon, which was polymerized for 2 days at 60°C. We obtained 50-nm sections using a Leica Ultracut S. Sections were stained with 3% uranyl acetate in methanol, followed by Reynold's lead citrate staining. Digital micrographs were taken using a Philips CM120 microscope and Gatan DigitalMicrograph software.
To characterize and quantify the budding phenotypes, transfected cells were identified by the presence of viral profiles at or close to the plasma membrane. For each transfected cell, all virus-related structures were photographed, irrespective of the number of viral particles present (except in cases where the number of viruses exceeded approximately 50, where only a few pictures were taken). To cope with potential experimental variation, cells from three different transfections were analyzed. For each of these, two epon blocks were cut and analyzed. For each block, images from at least 20 individual cells were collected. A second method was chosen to quantify the amount of budding profiles per micrometer of plasma membrane in the case of wild-type (wt) and NCSP2p6(KR) mutant-expressing cells. For this purpose, transfected cells were identified and photographed in a systematic random way. Markers for the negative size imprinted on the microscope screen were used to identify the size of one field of view. The top edge of the cell was oriented into that field. Moving clockwise around the cell, every third field of view was photographed independent of the presence or absence of viral profiles in the field. The plasma membrane length in those pictures was determined by intersection counting as described previously (7). The number of budding profiles per picture was counted, and the total number of budding profiles was related to the length of plasma membrane analyzed.
RESULTS
To address a potential role of HIV-1 Gag ubiquitination in virus release, we mutated a subset of lysine codons in gag. HIV-1 Gag (NL4-3 strain) contains 38 lysine residues (for an overview, see Fig. 1A). Of these, 13 are located in MA (6 in the domain important for membrane binding and targeting of Gag and 7 in the C-terminal part of MA between residues 95 and 114), 11 lie in CA, 10 in NC, and 2 each in SP2 and p6. For this study, the lysine residues in the NC, SP2, and p6 domains were targeted due to their proximity to the described HIV-1 Gag late domain motifs (16).
A series of plasmids carrying Lys-to-Arg mutations was constructed in the context of the previously described HIV-1 expression plasmids pR and pNL4-3 (1, 6). The resulting plasmids were termed according to the mutated region and the type of mutation. In the NC(KR) mutant, all 10 Lys codons in NC were mutated to Arg; in the SP2(KR) mutant, the 2 Lys codons in SP2 were mutated to Arg; and in p6(KR) the 2 Lys codons in p6 were mutated to Arg. Lys mutations in two domains were combined to give the constructs NCSP2(KR), NCp6(KR), and SP2p6(KR). Finally, a mutant that lacks all lysine codons downstream of CA [NCSP2p6(KR)] was constructed.
Detection of ubiquitinated viral proteins in mature HIV-1. We first examined the pattern of mature ubiquitinated Gag proteins within wt and mutant virions produced from pR-derived plasmids. pR is a full-length HIV-1 construct that results in the production of noninfectious virus due to mutations in the reverse transcriptase active site and in the region encompassing the primer binding site. 293T cells were transfected with either pHA-Ub alone or pR alone or were cotransfected with pHA-Ub and wt or lysine mutant pR constructs. Forty-eight hours after transfection, virus was harvested and pelleted through sucrose cushions. To remove soluble contaminants, the virion preparations were subjected to trypsin digestion. Equal amounts of virus (as determined by ELISA) were resolved by SDS-polyacrylamide gel electrophoresis and analyzed by Western blotting using anti-HA antibodies (Fig. 1B, upper panel).
When pHA-Ub or pR was expressed alone, no HA-reactive bands were detected by Western blotting (Fig. 1B, lanes 1 and 2). Upon cotransfection of pHA-Ub and pR, bands for several ubiquitinated proteins were observed (lanes 3 and 4). One of these bands was lost upon trypsin treatment (compare lane 4 to lane 3). This contaminating band was sometimes also found in the supernatant of cells expressing HA-Ub alone. Ubiquitinated MA and CA proteins were assigned based on their comigration with immunoprecipitated HA-Ub-conjugated MA and CA expressed from pcDNA3-based MA and CA expression constructs (6 and data not shown). In addition to ubiquitinated MA and CA, several ubiquitinated proteins migrating less than 20 kDa were detected, and they could be largely assigned due to their apparent mass and their absence in the case of specific Gag variants (lanes 5 to 11). The product identified as NC-HA-Ub was strongly decreased in intensity in lanes where Gag proteins lacking lysine residues in NC were analyzed (lanes 5, 8, 9, and 11). A faint band migrating at the same position as NC-HA-Ub was still detected in these cases, but this product was also observed in preparations of protease-negative viruses and is therefore not Gag derived (data not shown). The product designated SP2-HA-Ub was completely lost in lanes 6, 8, 10, and 11, where pR constructs carrying mutations of the two Lys codons in SP2 were analyzed. Its size is consistent with the modification of SP2 by a single Ub moiety. The band designated SP2-(HA-Ub)2 was also lost when the lysine codons in SP2 were mutated, and its size is consistent with modification of SP2 by two Ub moieties. Since SP2 contains two lysine residues, either both residues could be monoubiquitinated or one residue could be modified by a diubiquitin chain. Ubiquitinated p6 was harder to detect in mature virions. Upon long exposure, a product migrating slightly faster than NC-HA-Ub was detected (data not shown). This product was lost upon mutation of p6 Lys residues and presumably represents ubiquitinated p6. In conclusion, all lysine-containing Gag domains including the spacer peptide SP2 are found to be ubiquitinated in mature HIV-1.
Levels of ubiquitination of wt and mutant Gag polyproteins. We next tested whether the overall level or the pattern of Gag ubiquitination was altered upon replacement of C-terminal Gag lysine residues. For this purpose, the set of mutations was introduced into pR/PR(–) which carries a mutation in the protease active site (6). Transfection of pR/PR(–) therefore yields immature virus-like particles containing only the uncleaved Gag polyprotein and no processed products. 293T cells were cotransfected with one of these constructs and a HA-Ub expression vector. Forty-eight hours after transfection, cell lysates were prepared under denaturing conditions, Gag proteins were immunoprecipitated, and the extent of wt and mutant Gag ubiquitination was examined by Western blotting with an anti-HA antibody (Fig. 2). In the case of the wt construct, a ladder of Gag-reactive products was observed as described previously (Fig. 2, lane 1) (6). The intensity of ubiquitination was markedly increased when lysine residues in NC were lacking (lanes 2, 5, and 6), while mutation of lysine codons in p6 reduced overall ubiquitination of Gag (lane 4). Several mutations led to an altered pattern of Ub-modified Gag polyproteins. This was particularly evident for the p6(KR) mutation, which caused the loss of one of the ubiquitinated Gag products (compare lanes 4 and 6 to 8 with the other lanes). This result indicated that the different Gag-reactive products correspond, at least in part, to different isomers of monoubiquitinated Gag, suggesting the use of different Ub attachment sites. Mutation of all lysine codons downstream of the CA coding region [NCSP2p6(KR)] resulted in diminished levels of Gag ubiquitination and a loss of specific ubiquitinated products as described above. This result suggests that the loss of C-terminal ubiquitination is not met by a compensatory increase in ubiquitination of other Gag domains.
Steady-state levels of release and infectivity. We next evaluated virus release at steady state as well as viral infectivity for all mutants. For this, 293T or HeLaP4 cells were transfected with wt pNL4-3 or pNL4-3 constructs carrying the mutations described above. The virus-containing supernatant was harvested 24 h after transfection, cleared by centrifugation, and analyzed by CA ELISA as well as by titration on TZM indicator cells (Fig. 3). Results are shown for 293T cell transfections but were similar when HeLaP4 cells were used (data not shown). Antigen release (numbers shown at the bottom of Fig. 3) as well as the pattern of cell- and virus-associated Gag-derived proteins (data not shown) were similar for all constructs, suggesting no pronounced effect of lysine mutation on the amount of virus that is ultimately produced from transfected cells. Up to twofold-reduced levels of antigen were often, but not always, observed in the case of NCSP2p6(KR) mutant virus (Fig. 3). Infectious titers were determined by counting infected cells and were normalized for the amount of input virus (measured by ELISA) (Fig. 3). The NC(KR) mutant was reproducibly 10-fold less infectious than wt virus. Mutation of the lysine codons in SP2 or p6 alone, as well as the combination of both mutations [SP2p6(KR)], had no effect on infectivity. Adding the p6 and NC mutations also did not lead to a further reduction of viral infectivity compared to the NC mutant alone. In contrast, adding the SP2 and NC mutations [NCSP2(KR)] caused a further twofold reduction in infectivity. Interestingly, the NCSP2p6(KR) mutant was 5- to 10-fold less infectious than the NC mutant (43- to 100-fold less infectious than the wt), indicating that combination of the different mutations had a more than additive effect.
Virus release kinetics. Since replacement of lysine residues did not change virus production from transfected cells at steady state, we performed pulse-chase experiments to test whether virus release kinetics were altered. This would be expected if lysine mutations cause a retardation of budding rather than a complete block to virus release. HeLaP4 cells were pulse labeled 24 h after transfection with wt or mutant pR, and cells and virus particles were harvested after a 2.5-h chase period. Levels of intracellular wt or mutant Gag-derived proteins were similar in all cases (Fig. 4A, upper panel). A difference in Gag processing was observed for all constructs harboring NC lysine mutations. Processing of SP1 from the C terminus of CA was reduced (leading to an increased ratio of CA-SP1 to CA), and a CA-related product with an apparent molecular mass of 30 kDa was greatly diminished (Fig. 4A, upper panel, lanes 2, 5, 6, and 8). The identity of this band is currently not known, but its size matches the predicted molecular mass of a CA-SP1-NC-processing intermediate. The levels of virus release were normal for all mutants except for NCSP2p6(KR), which exhibited a strong reduction in virus release (Fig. 4A, lower panel, compare lane 8 to lanes 1 to 7). Importantly, no such reduction was observed for the constructs harboring mutations in any one domain or combination of two domains. Thus, the effects of the NC(KR) mutation on Gag polyprotein processing and of the combined NCSP2p6(KR) mutation on release are genetically separable, and lysine residues in NC, SP2, and p6 appear to cooperatively contribute to virus release.
For a more detailed analysis, the release kinetics were quantified for wt and NCSP2p6(KR) over a 6-h chase period. The efficiencies of wt and mutant virus release were calculated as the virus-derived CA-related signal divided by the sum of cell- and virus-derived CA-related signal. This quantification revealed that the budding efficiency of NCSP2p6(KR) mutant virus is approximately threefold lower compared to that of the wt (Fig. 4B). Similar results were obtained when infectious pNL4-3-based wt and mutant constructs were used (data not shown). Analysis of the virus release kinetics of the other mutants confirmed that only NCSP2p6(KR) exhibits slower virus release (data not shown).
Ultrastructure of Gag lysine mutant virus release. In order to directly evaluate wt and mutant virus budding, transfected HeLaP4 cells were examined by thin-section electron microscopy. wt virus-expressing cells revealed a mixture of free, mature virus, identified by the condensed conical core, and virus-budding profiles at the plasma membrane (Fig. 5a). Only a few immature virions were detected in this case. In NC mutant samples, on the other hand, virions were also released but remained largely immature (Fig. 5b). It appears likely that this defect in maturation corresponds to the reduced processing at the CA-SP1 site, since removal of SP1 is essential for core maturation (29). Of those virions that did mature, a significant percentage contained aberrant core structures (Fig. 5d and data not shown), indicating that the substitution of arginine for all lysine residues in NC also causes a defect in core maturation. No ultrastructural defects were observed for viruses harboring mutations in SP2, p6, or a combination of the two (data not shown). This result correlates with their wt phenotype regarding virus release and infectivity.
When the NC(KR) mutation was combined with mutation of either SP2 or p6 lysine residues [NCSP2(KR) and NCp6(KR), respectively], a budding defect became evident (Fig. 5c and d). Inspection of electron micrographs revealed more viral structures in late stages of budding in these cases than that observed for either wt, NC(KR), or SP2p6(KR). This increase in budding-arrested structures was observed in both cases, while an increase of "double buds" (late budding structures composed of two confluent budding sites) was only observed in NCSP2(KR)-transfected samples (Fig. 5c). A more pronounced late budding defect was seen for the NCSP2p6(KR) mutant, where all lysine residues downstream of the CA domain have been removed (Fig. 6). Many late budding structures often lining the plasma membrane (Fig. 6a to c) and connected to the cell by a thin membrane stalk (Fig. 6d) were observed. This phenotype is very reminiscent of observations made for mutants of the viral late domain.
Quantitative EM analysis of budding. To validate the impression from visual inspection of electron micrographs, we sought to quantify the budding properties of lysine mutation variants. For this, we separately counted released mature and immature virions as well as budding structures that were detected at or beyond the plane of the plasma membrane for the relevant subset of mutants (Fig. 7A). In each case, a virus-expressing cell was first located, and then all events were documented by digital imaging and counted without preselection. This procedure was repeated for different cells and different experiments as detailed in Materials and Methods and in the legend to Fig. 7. The relative percentages of released (both immature and mature) viruses and cell-attached virus buds is shown in Fig. 7A. This representation of the results was chosen because the number of total viral structures differed in each experiment due to variations in numbers of both particles and transfected cells, and calculating the average may thus give biased results. To avoid such bias, we calculated the Pearson product-moment correlation coefficient (measuring the degree of linear relationship between two variables) for the number of free viruses compared to total viruses. This correlation was linear in all cases, and we therefore depict the percentage of free viruses using a linear regression. Figure 7A plots the predicted number of free viruses per 100 viruses and the predicted number of budding structures (100 minus the percentage of free viruses). A similar result was obtained when actual numbers were averaged, with the deviation from the calculated result being <2% in each case (data not shown).
Quantification of free versus cell-attached viral structures confirmed that replacement of lysine residues in the C-terminal domains of Gag has additive effects on virus budding. In the case of wt-transfected cells, 82.5% of viral structures corresponded to cell-free virus, and 17.5% of viral structures corresponded to budding structures (Fig. 7A). The relative amount of free virus was slightly reduced for the NC(KR) (65%) and SP2p6(KR) (70%) variants, suggesting that they exhibit a subtle impairment in budding which is only detected by careful statistical analysis. The percentage of released virions was further decreased when all lysine residues were mutated in both NC and SP2 (60%) or in both NC and p6 (50%). When all lysine residues downstream of CA were removed [NCSP2p6(KR)], only 46% of virus structures corresponded to released virus, while 54% were tethered to the plasma membrane. In this quantification, we also distinguished immature and mature free virions, early and late budding structures, and single and chained buds. The results confirmed the pronounced maturation defect of the NC(KR) mutant, with 55% of the released virions being immature compared to only 8% in the case of the wt. Interestingly, SP2p6(KR) mutant free virions were also less mature than wt virions (30% immature free virions versus 8% in the wt). In the case of NCSP2p6(KR) mutant virions, 75% of released virions were immature. Analysis of the relative abundance of early and late budding showed no enrichment of early buds in any of the mutants. Statistical analysis confirmed that more "double" buds were present in the case of NCSP2(KR) mutant samples (Fig. 5c). While no chain buds or fused buds were detected in wt-transfected cells and 2.7% of the budding structures were in chains in the NC(KR) mutant samples, 21% of the budding structures detected in NCSP2(KR)-transfected cells were not separated from each other. Interestingly, this accumulation of double buds was only found for NCSP2(KR) and not in the other cases [including NCSP2p6(KR)].
A second method of quantification was applied in order to accurately determine the late budding defect of the NCSP2p6(KR) mutant. In this case, the number of budding profiles per micrometer of plasma membrane length was calculated for wt- and NCSP2p6(KR) mutant-transfected cells as described in Material and Methods (Fig. 7B). This analysis revealed that the density of budding structures was ca. fivefold higher in the case of the lysine mutation variant than that of the wt, thus confirming the observed late release block.
DISCUSSION
In this report, we describe a set of HIV-1 Gag C-terminal Lys-to-Arg mutants, which were analyzed in terms of infectivity, Gag ubiquitination, and virus release. Using these mutants, we extended our previous finding that MA, CA, and NC can be monoubiquitinated by showing that the spacer peptide SP2 is also mono- or diubiquitinated at low levels. Surprisingly, mutations of lysine residues in NC or SP2 increased ubiquitination of the Gag polyprotein in cell lysates (Fig. 2), while an overall increase in ubiquitination of any Gag-derived product was not observed in mature virions of these mutants (Fig. 1B). Although the reason for these observations is unknown, selective deubiquitination of Gag might take place during or after virus release, or "hyper-ubiquitinated" Gag proteins might be excluded from virions. Alteration of all Lys residues to Arg in any single domain downstream of CA had little or no effect on virus release. However, upon cumulative replacement of lysine residues in two or three of these domains, virus budding was reduced.
At this point, we cannot rule out that other lysine modifications, such as sumoylation, contribute to the observed phenotypes. However, no consensus sumoylation motif outside the p6 domain, which was previously reported to be sumoylated, is present within Gag (8, 11). While it seems very likely that NC(KR) mutant Gag has defects in addition to abolished Gag ubiquitination, this variant was released at essentially wt kinetics in biochemical experiments and with an efficiency similar to that of SP2p6 mutant virus as judged by EM. Therefore, the effect of lysine replacements in NC on polyprotein processing and maturation (and consequently on infectivity), which was not addressed in this report, can be clearly distinguished from the cumulative effects of lysine replacement in C-terminal Gag domains on virus budding. Taken together, our data strongly suggest that ubiquitination of Gag is indeed important for virus budding and is not a functionally irrelevant bystander effect.
The budding phenotype of NCSP2p6 mutant Gag was not readily observed under steady-state conditions at the high-expression level of transfected 293T cells, but it became obvious in pulse-chase experiments and electron micrographs. This suggests that the kinetics of membrane envelopment are altered, while budding is not prevented. Our unpublished results show that the expression level of Gag in transfected 293T cells is comparable to that in infected T-cell lines, suggesting that this is not an artifact of overexpression. One can speculate, however, that ubiquitination of lysine residues may be even more relevant in cells with lower levels of expression of Gag (e.g., primary T cells). The requirement for cumulative mutations to induce a budding defect suggests that no single specific ubiquitination site is necessary for this function and that ubiquitination of the different C-terminal domains may substitute for each other. Similarly, it had previously been shown that several normally available or newly introduced lysine residues, e.g., in the T-cell receptor -subunit, can be ubiquitinated when other sites are lost. Furthermore, the addition of extra lysine residues has been shown to rescue the budding defect of a lysine-mutated RSV Gag variant (10, 26). It is unclear at present whether the observed effect of the NCSP2p6(KR) mutation on budding would be even stronger if Ub acceptor sites in upstream domains were replaced in addition. This appears less likely, however, because Gag proteins lacking lysine residues downstream of CA did not show a compensatory induction of novel Gag-Ub conjugates while retaining ubiquitination at relatively low level. To test whether ubiquitination of NCSP2p6(KR) mutant Gag could be restored by reintroducing potential acceptor sites into the C-terminal part of Gag, we mutated three Arg codons to Lys in the NC domain of NCSP2p6(KR). These mutations led to an altered pattern of Gag ubiquitination, but overall ubiquitination of Gag was not increased (data not shown). This may be due to structural constraints preventing these sites from being used as ubiquitin acceptors or may suggest the existence of subtle position-dependent differences that were not detected in our general domain analysis.
The cumulative effect of Lys mutations is consistent with Ub acting as a transient protein modification that facilitates protein-protein interactions. Such a model would be in analogy to the role of Ub-based protein interactions during endocytosis. In this case, the sorting machinery is connected through a multitude of weak protein-protein interactions (9). Several different proteins of this network, cargo as well as endocytic machinery, are thought to become ubiquitinated in order to interact with proteins containing ubiquitin binding domains (UBDs). In several instances, UBDs carrying proteins themselves can be ubiquitinated. Since the interactions between UBDs and Ub are typically of low affinity, several such interactions are required to result in a transiently stable network. Our data are entirely consistent with a model in which Gag ubiquitination in the vicinity of the late domain would functionally contribute to such a network. Neither Gag ubiquitination nor any of the hypothetical Ub-based interactions of other components would be individually essential, while the overall number of interacting Ub moieties is. Of note, several proteins known to play important roles in Gag budding are ubiquitinated and/or contain UBDs (16). In vitro, the fusion of p6 to ubiquitin enhanced its interaction with TSG101, which contains a ubiquitin E2 variant-type ubiquitin binding domain (4, 16). Such a model would also be consistent with mutation of the Ub ligase recruiting late domain motifs in the case of PPXY-containing viruses having a much more pronounced effect on release than cumulative mutation of Gag lysine residues, and this needs to be experimentally addressed in the future.
ACKNOWLEDGMENTS
During this study, we made extensive use of the EM facility at the EMBL, Heidelberg, Germany. We thank Claude Antony and his team for hosting us as well as for their support throughout the project. We also thank Gareth Griffiths for ideas and comments concerning the quantification of EM data, Eric Freed and David Ott for expression constructs, and Piyush Kumar for statistical analysis of the data presented in Fig. 7A.
This work was supported in part by a grant from the Deutsche Forschungsgemeinschaft (SFB638, project A9). E.G. is the recipient of a stipend from the Boehringer Ingelheim Fonds.
Present address: Department of Microbiology and Genetics, Box 3025, Duke University Medical Center, Durham, NC 27710.
REFERENCES
Adachi, A., H. E. Gendelman, S. Koenig, T. Folks, R. Willey, A. Rabson, and M. A. Martin. 1986. Production of acquired immunodeficiency syndrome-associated retrovirus in human and nonhuman cells transfected with an infectious molecular clone. J. Virol. 59:284-291.
Freed, E. O. 1998. HIV-1 gag proteins: diverse functions in the virus life cycle. Virology 251:1-15.
Freed, E. O. 2002. Viral late domains. J. Virol. 76:4679-4687.
Garrus, J. E., U. K. von Schwedler, O. W. Pornillos, S. G. Morham, K. H. Zavitz, H. E. Wang, D. A. Wettstein, K. M. Stray, M. Cote, R. L. Rich, D. G. Myszka, and W. I. Sundquist. 2001. Tsg101 and the vacuolar protein sorting pathway are essential for HIV-1 budding. Cell 107:55-65.
Gottwein, E., J. Bodem, B. Muller, A. Schmechel, H. Zentgraf, and H. G. Krausslich. 2003. The Mason-Pfizer monkey virus PPPY and PSAP motifs both contribute to virus release. J. Virol. 77:9474-9485.
Gottwein, E., and H. G. Krausslich. 2005. Analysis of human immunodeficiency virus type 1 Gag ubiquitination. J. Virol. 79:9134-9144.
Griffiths, G. 1993. Fine structure immunocytochemistry. Springer-Verlag, Berlin, Germany.
Gurer, C., L. Berthoux, and J. Luban. 2005. Covalent modification of human immunodeficiency virus type 1 p6 by SUMO-1. J. Virol. 79:910-917.
Hicke, L., H. L. Schubert, and C. P. Hill. 2005. Ubiquitin-binding domains. Nat. Rev. Mol. Cell Biol. 6:610-621.
Hou, D., C. Cenciarelli, J. P. Jensen, H. B. Nguygen, and A. M. Weissman. 1994. Activation-dependent ubiquitination of a T cell antigen receptor subunit on multiple intracellular lysines. J. Biol. Chem. 269:14244-14247.
Johnson, E. S. 2004. Protein modification by SUMO. Annu. Rev. Biochem. 73:355-382.
Kikonyogo, A., F. Bouamr, M. L. Vana, Y. Xiang, A. Aiyar, C. Carter, and J. Leis. 2001. Proteins related to the Nedd4 family of ubiquitin protein ligases interact with the L domain of Rous sarcoma virus and are required for gag budding from cells. Proc. Natl. Acad. Sci. USA 98:11199-11204.
Konvalinka, J., M. A. Litterst, R. Welker, H. Kottler, F. Rippmann, A. M. Heuser, and H. G. Krausslich. 1995. An active-site mutation in the human immunodeficiency virus type 1 proteinase (PR) causes reduced PR activity and loss of PR-mediated cytotoxicity without apparent effect on virus maturation and infectivity. J. Virol. 69:7180-7186.
Martin-Serrano, J., S. W. Eastman, W. Chung, and P. D. Bieniasz. 2005. HECT ubiquitin ligases link viral and cellular PPXY motifs to the vacuolar protein-sorting pathway. J. Cell Biol. 168:89-101.
Martin-Serrano, J., D. Perez-Caballero, and P. D. Bieniasz. 2004. Context-dependent effects of L domains and ubiquitination on viral budding. J. Virol. 78:5554-5563.
Morita, E., and W. I. Sundquist. 2004. Retrovirus budding. Annu. Rev. Cell Dev. Biol. 20:395-425.
Muller, B., J. Daecke, O. T. Fackler, M. T. Dittmar, H. Zentgraf, and H. G. Krausslich. 2004. Construction and characterization of a fluorescently labeled infectious human immunodeficiency virus type 1 derivative. J. Virol. 78:10803-10813.
Ott, D. E., L. V. Coren, E. N. Chertova, T. D. Gagliardi, and U. Schubert. 2000. Ubiquitination of HIV-1 and MuLV Gag. Virology 278:111-121.
Ott, D. E., L. V. Coren, T. D. Copeland, B. P. Kane, D. G. Johnson, R. C. Sowder, Jr., Y. Yoshinaka, S. Oroszlan, L. O. Arthur, and L. E. Henderson. 1998. Ubiquitin is covalently attached to the p6Gag proteins of human immunodeficiency virus type 1 and simian immunodeficiency virus and to the p12Gag protein of Moloney murine leukemia virus. J. Virol. 72:2962-2968.
Ott, D. E., L. V. Coren, R. C. Sowder, Jr., J. Adams, and U. Schubert. 2003. Retroviruses have differing requirements for proteasome function in the budding process. J. Virol. 77:3384-3393.
Patnaik, A., V. Chau, F. Li, R. C. Montelaro, and J. W. Wills. 2002. Budding of equine infectious anemia virus is insensitive to proteasome inhibitors. J. Virol. 76:2641-2647.
Patnaik, A., V. Chau, and J. W. Wills. 2000. Ubiquitin is part of the retrovirus budding machinery. Proc. Natl. Acad. Sci. USA 97:13069-13074.
Putterman, D., R. B. Pepinsky, and V. M. Vogt. 1990. Ubiquitin in avian leukosis virus particles. Virology 176:633-637.
Schubert, U., D. E. Ott, E. N. Chertova, R. Welker, U. Tessmer, M. F.Princiotta, J. R. Bennink, H. G. Krausslich, and J. W. Yewdell. 2000. Proteasome inhibition interferes with gag polyprotein processing, release, and maturation of HIV-1 and HIV-2. Proc. Natl. Acad. Sci. USA 97:13057-13062.
Shehu-Xhilaga, M., S. Ablan, D. G. Demirov, C. Chen, R. C. Montelaro, and E. O. Freed. 2004. Late domain-dependent inhibition of equine infectious anemia virus budding. J. Virol. 78:724-732.
Spidel, J. L., R. C. Craven, C. B. Wilson, A. Patnaik, H. Wang, L. M. Mansky, and J. W. Wills. 2004. Lysines close to the Rous sarcoma virus late domain critical for budding. J. Virol. 78:10606-10616.
Strack, B., A. Calistri, M. A. Accola, G. Palu, and H. G. Gottlinger. 2000. A role for ubiquitin ligase recruitment in retrovirus release. Proc. Natl. Acad. Sci. USA 97:13063-13068.
Strack, B., A. Calistri, and H. G. Gottlinger. 2002. Late assembly domain function can exhibit context dependence and involves ubiquitin residues implicated in endocytosis. J. Virol. 76:5472-5479.
Wiegers, K., G. Rutter, H. Kottler, U. Tessmer, H. Hohenberg, and H. G. Krausslich. 1998. Sequential steps in human immunodeficiency virus particle maturation revealed by alterations of individual Gag polyprotein cleavage sites. J. Virol. 72:2846-2854.(Eva Gottwein, Stefanie Jg)