|Home | About | Journals | Submit | Contact Us | Français|
The p6 region of HIV-1 Gag contains two late (L) domains, PTAP and LYPXnL, that bind the cellular proteins Tsg101 and Alix, respectively. These interactions are thought to recruit members of the host fission machinery (ESCRT) to facilitate HIV-1 release. Here we report a new role for the p6-adjacent nucleocapsid (NC) domain in HIV-1 release. The mutation of basic residues in NC caused a pronounced decrease in virus release from 293T cells, although NC mutant Gag proteins retained the ability to interact with cellular membranes and RNAs. Remarkably, electron microscopy analyses of these mutants revealed arrested budding particles at the plasma membrane, analogous to those seen following the disruption of the PTAP motif. This result indicated that the basic residues in NC are important for virus budding. When analyzed in physiologically more relevant T-cell lines (Jurkat and CEM), NC mutant viruses remained tethered to the plasma membrane or to each other by a membranous stalk, suggesting membrane fission impairment. Remarkably, NC mutant release defects were alleviated by the coexpression of a Gag protein carrying a wild-type (WT) NC domain but devoid of all L domain motifs and by providing alternative access to the ESCRT pathway, through the in trans expression of the ubiquitin ligase Nedd4.2s. Since NC mutant Gag proteins retained the interaction with Tsg101, we concluded that NC mutant budding arrests might have resulted from the inability of Gag to recruit or utilize members of the host ESCRT machinery that act downstream of Tsg101. Together, these data support a model in which NC plays a critical role in HIV-1 budding.
The human immunodeficiency virus type 1 (HIV-1) Gag polyprotein Pr55Gag is the main structural component of the HIV-1 particle. It carries four domains: the N-terminal matrix (MA) followed by the capsid (CA), the nucleocapsid (NC), and the C-terminal p6 region. In MA, the N-terminal myristoylation signal and the cluster of basic residues are critical for Gag interactions with the plasma membrane (9, 27, 50, 76), whereas the CA domain, the main component of the HIV-1 core, carries regions responsible for Gag-Gag multimerization during assembly (21, 24, 28, 30, 39, 58). The NC region of Gag interacts with the viral genome as well as cellular RNAs that can serve as a scaffold for Gag-Gag assembly (2, 7, 10-12, 19, 30, 39). NC also interacts with cellular proteins believed to contribute to virus assembly (14, 77). The C-terminal p6 domain facilitates late steps in virus separation from the cell and is therefore essential for HIV-1 budding and release (26, 32).
Two short sequence motifs in the p6 region of HIV-1 Gag have been identified to act in late steps of budding and were therefore named late (L) domains; they carry the PTAP and LYPXnL motifs (32, 65) and bind the cellular proteins Tsg101 (18, 25, 43, 69) and Alix (65), respectively. These interactions lead to the recruitment of members of the host fission machinery, named the endosomal sorting complex required for transport (ESCRT) proteins, which promote membrane fission events that separate virus particles from the cell (42, 65, 70).
Tsg101 functions in HIV-1 release as part of ESCRT-I (44), which initiates the sorting of surface proteins into late endosomal compartments known as multivesicular bodies (MVBs) (4, 8). This process is topologically similar to HIV-1 budding and requires the recruitment of ESCRT-III proteins, also called the charged multivesicular body proteins (CHMPs), and the activity of the VPS4 AAA-type ATPase (3, 5, 6, 42, 70). The cellular protein Alix is involved in multiple cellular processes including endosomal metabolism (45, 48, 60). Alix binds the ESCRT-III member CHMP4 through its N-terminal Bro1 domain (22, 34, 36, 46, 65, 67), thereby directly linking HIV-1 Gag to members of the host fission machinery. Although the Alix/LYPXnL budding pathway's contribution to HIV-1 production from 293T cells is rather modest, its was shown previously to be fully functional in the absence of the PTAP L domain motif, provided that sufficient amounts of Alix were present in the cell (22, 67).
In the cell, CHMPs promote the sorting of cargo proteins and the generation of MVBs (3, 61) and are required for the fission events that separate HIV-1 particles from the plasma membrane (42, 70) as well as the two daughter cells at the completion of cytokinesis (13, 47). They are found soluble in the cytoplasm and are believed to be transiently recruited to membranes, where they assemble into tubes or filaments (29, 37) that facilitate membrane fission events (73). These structures are thought to be targets for the VPS4 ATPase, which disassembles ESCRT multimers from membranes (5, 6, 35, 37, 64). Both the PTAP/Tsg101 and LYPXnL/Alix HIV-1 budding pathways require an intact ESCRT-III pathway and the activity of VPS4. Any disruption of these sequential recruitments, such as a mutation or disruption of L domain motifs (18, 25, 32) or dominant negative interference with the function of ESCRT-III members (42, 70, 75) or the VPS4 ATPase (42, 70), adversely affects virus release. This indicated that Gag interactions with the ESCRT machinery are essential for virus separation from the cell. Whereas Alix has been found to directly bind CHMP4 isoforms, the protein-protein interactions that lead to recruitment of ESCRT-III factors by Tsg101 have yet to be identified.
It has been established that the p6-located L domain motifs PTAP and LYPXnL bind Tsg101 and Alix, respectively, to gain access to members of the host ESCRT pathway, a key step in HIV-1 release from 293T, HeLa, and T cells (22, 23, 32, 65, 67). The HIV-1 PTAP motif retained activity when transplanted in heterologous retroviral Gag polyproteins, suggesting that it can act in virus release autonomously (33, 38, 54, 55, 66). Interestingly, however, the PTAP motif exhibited a clear functional context dependence (38, 54, 55, 66, 74), as its location in Gag greatly influenced its activity. This finding suggested the involvement of other regions of Gag in PTAP L domain function. Indeed, several observations are consistent with the notion that NC is involved in the L domain-driven release of HIV-1. In full-length HIV-1, a mutation or deletion of the NC domain was found to obliterate virus release, even though these mutant Gag proteins carried an intact p6 domain (20, 57) and retained the ability to assemble at the plasma membrane (30, 52, 71, 72). Additionally, the NC domain is also essential for Alix-mediated HIV-1 release (20, 57). We demonstrated that Alix's Bro1 domain interaction with NC led to the recruitment of the fission-inducing ESCRT-III components necessary for virus release through the LYPXnL L domain motif (20).
We have expanded on our recent observation that NC is involved in HIV-1 release driven through the PTAP motif and identified NC residues that are critical for virus release. In fact, the mutation of basic residues in the N- or C-terminal region of NC resulted in arrested HIV-1 budding at the plasma membrane of 293T cells as well as Jurkat and CEM T lymphocytes. NC mutant Gag proteins form spherical particles at the cell membrane, indicating their ability to undergo proper assembly. Remarkably, though, these particles remained tethered to the cell surface in structures similar to those seen following the disruption of the L domain PTAP motif. These results suggested that mutations in NC interfered with HIV-1 particle separation from the cell. Such defects were corrected when the NC mutants RKI and RKII were coexpressed with a release-defective HIV-1 construct carrying the wild-type (WT) NC domain but lacking all known L domains, indicating that NC mutant Gag proteins coassembled with WT NC and can act as an “L domain-like” domain in trans to trigger release. Importantly, the budding defects of the NC mutants were also alleviated by the ectopic overexpression of Nedd4.2s, a ubiquitin ligase known to link Gag to the host ESCRT pathway independently of NC and p6 domains (16, 68). As HIV-1 NC mutant Gag proteins retained the ability to bind Tsg101, we concluded that their virus budding defects resulted from the inability of Gag to recruit and/or utilize ESCRT members acting downstream of Tsg101. These data are the first to identify a new and critical role for the NC domain of Gag in virus budding, possibly as a participant in the recruitment and/or utilization of ESCRT members necessary for HIV-1 scission from the cell.
293T cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and transfected by using Lipofectamine 2000 (Invitrogen). The T-lymphocyte cell lines Jurkat (clone E6-1) and CCRF-CEM were obtained from the ATCC and maintained in RPMI 1640 medium supplemented with 10% FBS, 10 mM HEPES, and 1 mM sodium pyruvate. Jurkat and CEM cells were transfected by using Amaxa Nucleofector technology (Lonza).
The wild-type (WT) full-length HIV-1 molecular clone pNL4-3 (GenBank accession number AF324493) (1) and its derivatives, the NC and p6 mutants, were previously described (20, 32). The DelNC proviral construct was previously described (52). In the HIV-1 RKI mutant, NC residues K3, R7, R10, K11, K14, K20, and K26 were replaced with alanines. The RKII mutant carries alanine substitutions of residues K26, R29, R32, K33, K34, K38, K41, K47, and R52. The p6 L domain mutants PTAP− and YP− (LYPXnL−) contained the PTAP-to-LIRL and the LYPXnL-to-LLVXnL substitutions, respectively. All of the above-mentioned proviral constructs were also derived as protease-deficient (PR−) versions by introducing an R57G substitution into the protease sequence (52) using the overlap extension PCR method. The hemagglutinin (HA)-tagged Nedd4.2s expression vector was previously described (20). The Flag-Tsg101 expression vector was generated by PCR amplification using Tsg101 cDNA (GenBank accession number U52945) (40) and subcloning into p3XFLAG (Sigma) between the NotI and EcoRI sites.
Two micrograms of proviral DNA was used to transfect 293T cells (3 × 106 cells), whereas 5 and 10 μg were used to electroporate Jurkat cells (3 × 106 cells) and CEM cells (5 × 106 cells), respectively. At 24 h posttransfection (48 h for T-cell lines), culture supernatants were filtered, and virions were pelleted at 151,000 × g for 1 h on a 20% (wt/vol in phosphate-buffered saline [PBS]) sucrose cushion. Cells were washed once with cold PBS and lysed in lysis buffer (1% [vol/vol] Igepal, 50 mM Tris [pH 8.0], 150 mM NaCl, and protease inhibitor cocktail [Complete; Roche]). Virion and cellular proteins were analyzed by SDS-PAGE and immunoblotting using either a mouse monoclonal anti-HIV-1 p24 antibody (NEA-9306; NEN) or an anti-HIV-1 p24 monoclonal antibody (clone 183-H12-5C; obtained through the NIH AIDS Research and Reference Reagent Program). The HIV-1 virus release efficiency was calculated as the ratio of the amount of virion-associated Gag to the total amount of cellular Gag as determined by densitometry analyses of Western blot films using ImageJ software (1997 to 2007; W. S. Rasband, NIH, Bethesda, MD [http://rsb.info.nih.gov/ij]). Nedd4.2s expression was detected by using a monoclonal anti-HA antibody (HA-7; Sigma). Anti-tubulin (DM 1A; Sigma) was used to probe for cellular tubulin. Virus release was also measured by a p24 enzyme-linked immunosorbent assay (ELISA). For this assay, the HIV-1 capsid protein was quantified in filtered cell culture medium and in pelleted virions (resuspended in cold PBS) using the Retrotek p24 ELISA kit (Zeptometrix).
We used an equilibrium flotation centrifugation protocol that was described previously by Ono and Freed (51). Briefly, 293T cells (3 × 106 cells) were transfected with 2 μg of the indicated plasmids. At 24 h posttransfection, cells were washed in cold PBS and then resuspended in 10% (wt/vol) sucrose in Tris-EDTA (TE) buffer (10 mM Tris-HCl [pH 7.5], 1 mM EDTA) in the presence of a protease inhibitor cocktail. Following sonication and centrifugation at low speed to pellet nuclei, 250 μl of the postnuclear fraction was mixed with 1.25 ml of 85.5% (wt/vol) sucrose in TE buffer and placed at the bottom of an ultracentrifuge tube. Next, 7 ml of 65% (wt/vol) sucrose in TE buffer and then 3.25 ml of 10% (wt/vol) sucrose in TE buffer were successively layered on top of the 71% sucrose-containing cellular extract fraction. The gradient was centrifuged at 100,000 × g for 18 h at 4°C in a Beckman SW41 rotor. Ten 1.2-ml fractions were collected from the top of the gradient and analyzed by SDS-PAGE and immunoblotting using an anti-HIV-1 p24 monoclonal antibody.
293T cells (3 × 106 cells) were transfected with 2 μg of the indicated pNL4-3 plasmids. At 24 h posttransfection, culture media were filtered and spun at 151,000 × g for 1 h on a 20% (wt/vol in PBS) sucrose cushion. Pelleted virus-like particles (VLPs) were resuspended in cold PBS, and their RNA was extracted by using the QIAmp viral RNA minikit (Qiagen). The amount of Gag protein present in pelleted VLPs was quantified by SDS-PAGE and Western blot analysis using an anti-HIV-1 p24 antibody. Extracted viral RNAs were treated with DNase I (Qiagen) and run through an RNeasy spin column (Qiagen). Five hundred nanograms of RNA was reverse transcribed with random primers using the SuperScriptIII kit (Invitrogen), and control experiments without the reverse transcriptase enzyme were performed in parallel. Quantitative PCR was conducted with 5 μl of diluted cDNA preparations (dilutions from 1/10 to 1/100) using IQ SYBR green Supermix (Bio-Rad) on a 7300 real-time PCR system (Applied Biosystems). Primers for HIV genomic RNA and 7SL RNA as well as the method of quantification were previously described by Houzet and colleagues (31). Relative quantification was based on a standard curve generated from 100, 10,000, and 1,000,000 copies of the pNL4-3 plasmid. RNA levels in VLPs were normalized to the amount of VLP-associated Gag measured by densitometry and Western blot analyses.
293T cells were transfected with the indicated proviral plasmids and the Flag-Tsg101 expression vector. Forty-eight hours posttransfection, cells were harvested, washed twice in cold PBS, and lysed in radioimmunoprecipitation assay (RIPA) buffer (0.5% Nonidet P-40, 50 mM HEPES [pH 7.3], 150 mM NaCl, 2 mM EDTA, 20 mM β-glycerophosphate, 0.1 mM Na3VO4, 1 mM NaF, 1 mM phenylmethylsulfonyl fluoride [PMSF], 0.5 mM dithiothreitol [DTT], and protease inhibitor cocktail [Complete; Roche]). Cell lysates were cleared at 16,100 × g at 4°C for 10 min, and supernatants were incubated at 4°C with anti-Flag monoclonal antibody-conjugated EZview agarose beads (Sigma). Next, the beads were washed in RIPA buffer prior to elution with 3× Flag peptide (Sigma). Immunoprecipitates and cell lysates (input fractions) were analyzed by SDS-PAGE and immunoblotting using an anti-Flag M2 antibody (Sigma) and HIV-positive patient serum.
At 32 h (293T cells) and 48 h (T-cell lines) posttransfection, cells were fixed for 2 h at room temperature in 2% (vol/vol) glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4). Next, they were rinsed in cacodylate buffer and postfixed in 1% (vol/vol) osmium tetroxide in the same buffer. The samples were subsequently rinsed in 0.1 N sodium acetate buffer (pH 4.2), stained with 0.5% (vol/vol) uranyl acetate in the same buffer, dehydrated in graded ethanol, and then infiltrated overnight in pure epoxy resin. The samples were embedded in fresh resin the next day and cured at 55°C. Blocks were cut from the cured samples and mounted appropriately for ultramicrotomy. Thin sections were stained with uranyl acetate and lead citrate and stabilized by carbon evaporation. Images were obtained with a Hitachi H7600 electron microscope equipped with an AMT XL41M digital camera. A total of 100 to 200 virus particles were examined for each sample and categorized as immature budding, free immature particles, or free mature particles.
We recently reported that HIV-1 NC mutants failed to release particles from 293T cells even though they contained a functional PTAP L domain motif (20). To characterize the effect of mutations in NC on HIV-1 production from 293T cells, we replaced basic residues in either the N-terminal (RKI mutant) or C-terminal (RKII mutant) half of NC with alanines (Fig. (Fig.1A)1A) and compared the effect of these modifications to that caused by the disruption of the PTAP motif (PTAP−), which typically abrogates HIV-1 release from 293T cells. The NC mutants RKI and RKII displayed a severe reduction (~13-fold) in particle release (Fig. (Fig.1B,1B, lanes 2, 5, and 6). This profound decrease was comparable to that observed with the PTAP− mutant, as quantified by the respective release of WT HIV-1, NC mutants, and single and double L domain mutants from four independent experiments (Fig. (Fig.1C).1C). Gag and Gag-derived proteins in cells and virions produced with the NC and PTAP− mutants displayed remarkable similarities in their processing profiles, as illustrated by the presence of the p25-24CA doublet, a hallmark of budding-defective HIV-1 viruses (Fig. (Fig.1B,1B, lanes 2, 5, and 6). Taken together, our data demonstrate the critical role of NC's basic residues in HIV-1 release from 293T cells. Similar results were observed for HeLa cells (data not shown).
Substitutions of basic residues to alanines in NC were previously reported to inhibit virus release, a loss that was attributed to virion instability. In this context, it was proposed that unstable viral particles disintegrated and were released as “soluble” proteins rather than as intact virus particles (71, 72). We therefore sought to examine whether nascent particles from the RKI and RKII NC mutants exhibited any structural defects and reasoned that if they did, we should be able to capture “soluble” CA proteins released into the culture media. Using ELISA, we quantified the amount of viral proteins released in filtered culture media of cells expressing NC mutants and included the PTAP− mutant as a release-defective control. The RKI and RKII NC mutants released very little to no p24CA proteins in both culture media and pelleted virus particles and in amounts comparable to those observed with the PTAP− mutant (Fig. (Fig.1D).1D). This result suggested that the RKI and RKII mutants did not release “soluble” CA proteins, indicating that these mutations in NC had no detectable effect on virus particle integrity. Thus, we concluded that mutations of basic residues in NC interfered with virus release rather than the stability of nascent viral particles.
The basic residues in NC were previously described to be involved in targeting Gag to the plasma membrane and in binding to viral and cellular RNAs, both of which can serve as a scaffold for Gag-Gag assembly (17, 30, 56, 62). Therefore, the RKI and RKII mutants were examined for both properties to determine whether their release defect was due to an inability to assemble into particles following the loss of the interaction with the plasma membrane and/or RNA. Using flotation assays, we observed that the membrane-binding patterns of the RKI and RKII mutants were indistinguishable from that of WT Gag (Fig. (Fig.2A),2A), indicating that the mutation of basic residues in NC did not affect Gag interactions with cellular membranes.
Next, we assessed the binding of NC mutants to cellular RNAs by examining the ability of Gag virus-like particles (VLPs) to incorporate 7SL RNA (49) as well as PGK1 and PLEKHB2, two cellular mRNAs that are selectively packaged in HIV-1 particles (59). RKI and RKII immature Gag VLPs were generated by the inactivation of the viral protease, an approach that was previously described for NC and PTAP− HIV-1 mutants (32, 52, 72). As expected from previous reports (17, 19, 63, 71), RKI and RKII mutant Gag proteins displayed a reduced packaging of viral genomic RNA by 25- and 50-fold, respectively (data not shown). However, both mutants retained the ability to interact with 7SL RNA with ~75% and ~35% efficiencies, respectively (Fig. (Fig.2B).2B). Additionally, NC mutants incorporated PGK1 and PLEKHB2 mRNAs at levels comparable to or higher than those of their WT Gag counterpart (Fig. (Fig.2B).2B). These results indicated that the NC mutants RKI and RKII retained the ability to interact with cellular RNAs.
To further characterize the release defect observed with the NC mutant particles, we performed transmission electron microscopy (TEM) on 293T cells expressing WT HIV-1 or its NC mutant counterparts RKI and RKII. In the case of WT HIV-1, very few particles were found budding at the plasma membrane, and the majority of the particles seen were released around cells as mature particles (Fig. (Fig.3A).3A). In contrast, the RKI and RKII mutants displayed numerous spherical viral particles tethered to the plasma membrane or to each other (Fig. 3C and D). Examinations of the mutants RKI (Fig. 3Cb, c, and d) and RKII (Fig. 3Df, g, and h) at a high magnification showed that both mutants assembled particles that were similar in size and morphology to those observed with the PTAP− mutant (Fig. (Fig.3B),3B), demonstrating that mutations of basic residues in NC caused budding defects. Additionally, an electron-dense material was observed at the zone of contact between tethered RKI and RKII arrested budding particles (Fig. 3Cb and Df).
Next, we quantified budding defects for each NC mutant and compared them to the PTAP− mutant by counting the number of budding particles at the plasma membrane versus those released as immature and mature particles. Our enumerations indicated that the majority (~80%) of WT HIV-1 particles were released as mature particles (Fig. (Fig.3E).3E). In contrast, ~80% of particles from the PTAP− and NC mutants remained attached to the plasma membrane as arrested budding particles, and no released mature particles were observed with either NC or PTAP− mutants (Fig. (Fig.3E).3E). These results demonstrated that the mutation of basic residues in NC, or the p6-located PTAP motif, causes budding defects that led to a significant decrease in HIV-1 production from 293T cells.
Data obtained in this study demonstrated that mutations of basic residues in NC interfered with HIV-1 virion production from 293T cells. In fact, such mutations caused clear virus budding arrests at the plasma membrane. These findings suggest that NC plays a critical role in HIV-1 budding from 293T cells. To examine the role of NC in HIV-1 budding in a physiologically relevant setting, we assessed the effect of the RKI and RKII mutations on virus production from Jurkat and CEM T cells and compared it to that of HIV-1 L domain single mutants PTAP− and LYPXnL− (YP−) or the double mutant missing both L domain motifs (PTAP−/YP−) (Fig. 4A and B). Remarkably, the RKI and RKII mutations led to a complete obliteration of viral release from Jurkat and CEM cells (Fig. 4A and B). This release defect was comparable to that seen following the mutation of both L domains in the PTAP−/YP− mutant (Fig. 4C and D), confirming that the mutation of basic residues in NC caused a profound decrease in virus release from T cells.
Next, we used TEM to visualize the release defects of the HIV-1 RKI and RKII mutants in T cells. The majority (~80%) of WT HIV-1 particles seen around Jurkat cells were mature released particles, and only a few budding particles were observed (Fig. (Fig.5A).5A). In contrast, the NC mutants RKI and RKII displayed dramatic budding defects characterized by arrested particles decorating the entire plasma membrane (Fig. 5C and D). Examination of these structures at a higher magnification revealed spherical particles tethered to the plasma membrane and to each other with a continuous membranous stalk (Fig. 5Ca and b and Dc and d). Likewise, particles attached to the plasma membrane and to each other were seen with the PTAP−/YP− mutant (Fig. (Fig.5B5B).
We quantified NC mutant budding defects in Jurkat cells by counting the number of budding particles arrested at the plasma membrane and those released as immature and mature particles and compared them to those observed with the double mutant PTAP−/YP−. In stark contrast to WT HIV-1 and like the PTAP−/YP− mutant, the majority of budding particles (~70% to 90%) produced by the NC mutants RKI and RKII remained attached to the cell (Fig. (Fig.5E).5E). Similar defects were also observed for CEM cells (data not shown). Taken together, the biochemical and TEM analyses of the RKI and RKII mutants indicated that the disruption of basic residues in NC affected the budding and release of HIV-1 virions from T cells as a consequence of membrane scission impairment.
NC mutant viruses formed assembled spherical particles that remained attached to the cell membrane, indicating that they failed to complete the membrane fission steps necessary for virus release. This finding prompted us to hypothesize that mutations of basic residues in NC affected the recruitment and/or the function of the host's ESCRT proteins responsible for inducing virus separation/scission from the cell. In this case, we reasoned that such release defects could be remedied provided Gag gains access to the cell's ESCRT machinery via an alternative pathway. To this end, we utilized Nedd4.2s, a Nedd4-like ubiquitin ligase that can stimulate the release of HIV-1 independently of the NC and p6 domains (16, 68), and examined the effect of its in trans expression on the NC mutants RKI and RKII. Remarkably, Nedd4.2s rescued the release defects of both the RKI and RKII mutants to levels comparable to those obtained following the stimulation of the PTAP−/YP− double mutant (Fig. 6A and B, lanes 3, 4, 6, and 7). Importantly, Nedd4.2s failed to rescue the release of the assembly-defective HIV-1 lacking the entire NC domain (DelNC) (Fig. (Fig.6C,6C, lanes 10 to 12), indicating that Nedd4.2s exerts its stimulatory effect during postassembly events. These results indicate that restoring access to the host ESCRT pathway through the overexpression of Nedd4.2s alleviates budding defects caused by the mutation of basic residues in NC, suggesting the inability of these viruses to access or utilize the host ESCRT function.
The rescue of the budding-defective NC mutants by the Nedd4.2s ligase (Fig. (Fig.6)6) clearly demonstrated that these budding defects can be corrected by providing NC mutant Gag proteins with access to the ESCRT machinery, an indication that mutations of NC interfered with Gag's ability to access or utilize members of the ESCRT pathway necessary to release virus, despite the presence of an intact L domains in its p6 domain. This result further supported the notion that NC is involved in recruiting ESCRT function to HIV-1 Gag and led us to speculate that if this hypothesis is correct, the coexpression of a WT NC domain in trans should provide a new access to ESCRT function(s) and thus correct NC mutant release defects. In fact, we found that the coexpression of as little as 12.5% of a Gag protein carrying a WT NC but devoid of all known L domains rescued the release of the NC mutants RKI and RKII (Fig. 7A and B, compare lanes 3 and 4), whereas the coexpression of 25% of this protein corrected their release to near WT levels (lane 5). We concluded that the coexpression of an intact NC can correct budding defects efficiently and in a dose-dependent manner, suggesting that it can act in trans as an “L domain-like” domain and participate in linking Gag to the cell's ESCRT pathway.
Electron microscopy examination of NC mutants showed tethered virus particles at the plasma membrane, demonstrating that virus release is arrested at late stages of budding. These phenotypes are reminiscent of those seen following a disruption the L domain PTAP in the p6 domain of HIV-1 Gag. As the PTAP motif is the sequence that recruits the cellular protein Tsg101 (25, 69), an interaction that is critical for HIV-1 scission from 293T cells (25), we examined whether mutations introduced into NC affected Tsg101 interactions. As expected, we found that WT Gag proteins coimmunoprecipitated with Flag-Tsg101 (Fig. (Fig.8),8), in contrast to the Gag mutant lacking both L domain motifs in p6 (PTAP−/YP−) (Fig. (Fig.8,8, lanes 2 and 3). However, Gag proteins expressing either the RKI or RKII NC mutant retained the ability to bind to Flag-Tsg101 (Fig. (Fig.8,8, lanes 4 and 5). These data indicated that the mutation of basic residues in NC did not interfere with Gag's ability to interact with Tsg101, implying that NC mutant budding defects are not due to Gag's inability to recruit the ESCRT-I member Tsg101.
The current model of HIV-1 morphogenesis and release can be divided into multiple sequential and well-orchestrated steps. The MA domain mediates Gag interactions with the plasma membrane, the site of particle assembly. This step is followed by Gag-Gag multimerization into a spherical shell that requires the CA domain but also involves the NC domain by virtue of its ability to interact with RNA. The viral egress function was believed to be confined to the p6 region and carried out by the PTAP and LYPXnL L domain motifs, the two sequences endowed with the function of recruiting components of the ESCRT pathway that promote virus release. Although initially considered autonomous, L domain motifs are greatly weakened when moved away from their native location (38, 41, 54, 55, 74) and cease to function when the NC domain of Gag is mutated or deleted (17, 52, 72). These observations indicated that p6 does not function autonomously and suggested a functional interdependence between the p6 and NC domains in promoting virus release. Here we present evidence supporting a role for NC in HIV-1 release. Following mutations of basic residues in NC, fully assembled virus particles remained tethered to the plasma membrane, demonstrating a clear defect in virus abscission from the cell. Such phenotypes were seen for both 293T and T cells and were similar to budding defects observed for the HIV-1 PTAP− mutant in 293T cells. As these NC mutant viruses retained all viral morphogenetic properties, including binding to the cell's membrane and RNA and the assembly of spherical particles, we concluded that NC is involved in virus budding, possibly by participating in the recruitment and/or the function of components of the ESCRT pathway, the host cell fission machinery members that promote HIV-1 release.
Previous studies reported that mutations of basic residues in NC result in the inhibition of HIV-1 production. Such a phenotype was attributed to various defects, including Gag's inability to interact with RNA (56), to multimerize and form particles (17), or to assemble stable virions (52, 72). These seemingly conflicting explanations reflected the difficulty in capturing the multiple and interconnected functions of the NC domain's basic residues in virus morphogenesis and release. In fact, mutation of basic residues in NC seems to have inhibited an aspect of HIV-1 morphogenesis other than assembly, since NC mutant Gag proteins were previously shown to retain the ability to multimerize (17, 30, 39). Consistent with these observations, we found that Gag proteins carrying mutations of basic residues in either the N- or C-terminal regions of NC (RKI and RKII mutants) formed fully assembled spherical HIV-1 particles at the plasma membrane (Fig. (Fig.22 and and4),4), where they remained arrested at late stages of particle budding, suggesting a failure to release. NC mutant budding defects were ascertained by a careful examination of dozens of distinct fields and morphogenetic characterizations of 100 to 200 particles. These enumerations showed that NC mutants exhibited tethered budding particles in numbers comparable to those seen with the PTAP− mutant in 293T cells and to those seen with the PTAP−/YP− double mutant in T cells (Fig. (Fig.33 and and5),5), further supporting a budding defect phenotype.
The NC mutant Gag polyproteins (RKI and RKII) formed spherical particles at the membrane and coassembled with WT Gag proteins (Fig. (Fig.7),7), further attesting to their ability to assemble virus. In fact, the proper particle assembly of these NC mutants was not surprising and is most likely the result of a combination of factors: (i) mutations introduced into NC were confined to either the N- or the C-terminal region, leaving half of the basic residues in NC available to engage in interactions with cellular RNAs that were reported to be specifically incorporated into HIV-1 particles (59) (Fig. (Fig.2)2) and can substitute for genomic RNA by serving as a scaffold for Gag-Gag multimerization (11, 17, 59); (ii) NC mutant proteins retained an intact MA domain, in which basic residues can also bind RNA (53) and indirectly favor Gag-Gag multimerization at the plasma membrane (15); and (iii) NC mutants bound the plasma membrane with efficiencies comparable to that of the wild-type protein (Fig. (Fig.2),2), an HIV-1 morphogenetic property that is known to compensate for defects caused by a disruption in the NC domain, by serving as a scaffold for Gag-Gag assembly (30). By retaining these properties, NC mutants were able to efficiently multimerize, further excluding an assembly defect as a cause for their inability to release. NC mutants formed stable virus particles that tethered to the plasma membrane, thus exhibiting a clear release defect. Together, our findings present the first evidence that the NC domain of Gag plays a role in HIV-1 budding and release.
Both of the NC mutants RKI and RKII failed to release virus particles from 293T cells despite having an intact PTAP L domain motif. Because the majority of HIV-1 particle release from 293T cells is mediated through the PTAP motif, we concluded that an intact NC domain is critical for HIV-1 release through the PTAP/Tsg101 pathway. Remarkably, NC mutant viruses remained tethered to the plasma membrane with long membranous stalks that are morphologically similar to those seen following the disruption of the PTAP L domain motif (Fig. (Fig.33 and and5).5). In addition, the mutation of basic residues in NC or the PTAP L domain motif in p6 led to an intracellular Gag processing defect, as illustrated by the accumulation of the p25/24CA doublet, a hallmark of HIV-1 budding defects. These data indicate that the PTAP L domain motif in p6 requires the basic residues in its adjacent NC domain to facilitate virus release, thus establishing the first functional link between these two adjacent domains of Gag in virus budding.
NC mutants were arrested in late stages of virus budding and remained attached to the plasma membrane of 293T and T cells with long membranous stalks, indicating a failure to abscise their particles away from the cell. To achieve the latter, HIV-1 requires the recruitment of ESCRT-I through the Tsg101-PTAP motif interaction, members of ESCRT-III that promote membrane fission, and the activity of the VPS4 ATPase. Both NC mutant Gag proteins retained binding to Tsg101 (Fig. (Fig.7),7), thereby excluding the disruption of Tsg101 recruitment as a cause for the failure of NC mutants to release virus. These findings suggested that mutations in NC impeded the ability of Gag to interact with a stable ESCRT-I complex and/or with the fission machinery factors that act downstream of ESCRT-I (i.e., ESCRT-III and VPS4). Alternatively, a mutation in NC could have prevented Gag from utilizing ESCRT members that were bound to its adjacent p6 domains. In either case, NC mutants remained arrested in late stages of virus budding at the cell surface, indicating improper release most likely due to the lack of ESCRT function. This notion was supported by the critical finding that the NC mutants’ release defects were remedied by providing Gag with an alternative access to the host's ESCRT pathway, through the overexpression of Nedd4.2s (Fig. (Fig.6),6), an E3 ubiquitin ligase that links Gag to the cell's abscission complexes independently of both NC and p6 but in an ESCRT-dependent manner (16, 68). Such a rescue was specific for viruses that completed assembly, since Nedd4.2s failed to rescue an assembly-defective virus (DelNC), indicating that its activity can remedy only postassembly events and strongly suggesting that Nedd4.2s triggered the fission of the NC mutant particles from the cell's membrane. Since Nedd4.2s rescues virus release in a cellular ESCRT-dependent manner (16, 68) and is considered to act as an alternative link between Gag and the cellular ESCRT pathway to promote membrane fission (16, 68), we concluded that the NC domain is involved in postassembly events necessary for HIV-1 budding.
Other evidence in this study also supported a role for NC in virus budding and release. In fact, the NC mutants’ release defects were relieved when the mutants coexpressed with release-defective HIV-1 carrying a WT NC but lacking all L domain functions. This result not only indicated that NC mutants retained the ability to coassemble with WT NC-Gag proteins but also indicated that NC can act in trans as an “L domain-like” domain to promote virus budding and release. Interestingly, only a small amount of WT NC was needed to trigger virus release (12.5%) (Fig. (Fig.7),7), suggesting that only a subset of NC proteins is involved in these late events of virus budding. One would thus envision that NC domains within Gag proteins localized at the budding neck would be engaged in a dynamic process with its adjacent p6 domain to promote late stages of virus abscission from the cell.
We previously reported that an interaction between the Bro1 domain of the ESCRT-associated Alix protein and NC is involved in the recruitment of the ESCRT-III member CHMP4 that is critical for virus release (20), implicating NC in HIV-1 release mediated through the LYPXnL/Alix budding pathway. Here we showed evidence of a role for NC in HIV-1 release through the PTAP/Tsg101 pathway. The nature of the involvement of NC in this pathway is not clear. Evidence presented in this report, however, suggested that NC participates in the recruitment and/or utilization of ESCRT proteins and bears a cis-acting “L domain-like” region that functions in concert with the PTAP L domain in p6. Further studies are warranted to elucidate the nature of the involvement of NC in these processes.
We are grateful to Adam Harned (SAIC, NCI-Frederick) for assistance with electron microscopy processing and imaging, Laurent Houzet (Laboratory of Molecular Microbiology [LMM]) for valuable help with RNA incorporation experiments, Rajat Varma (LCMI) for advice on T-lymphocyte electroporation, Que Dang (Laboratory of Cellular and Molecular Immunology [LMM]) for helpful discussions and critical reading of the manuscript, and Alan Rein (DRP, NCI-Frederick) for insight as this work evolved.
This work was supported by the Intramural Research Program of the NIAID, NIH.
Published ahead of print on 15 December 2010.