JVI Figure table search 04
Home Help [Feedback] [For Subscribers] [Archive] [Search] [Contents]
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Other Versions of this Article:
JVI.00284-07v1
81/12/6216    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Chatel-Chaix, L.
Right arrow Articles by DesGroseillers, L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Chatel-Chaix, L.
Right arrow Articles by DesGroseillers, L.

 Previous Article  |  Next Article 

Journal of Virology, June 2007, p. 6216-6230, Vol. 81, No. 12
0022-538X/07/$08.00+0     doi:10.1128/JVI.00284-07
Copyright © 2007, American Society for Microbiology. All Rights Reserved.

The Host Protein Staufen1 Participates in Human Immunodeficiency Virus Type 1 Assembly in Live Cells by Influencing pr55Gag Multimerization{triangledown}

Laurent Chatel-Chaix,1,2 Levon Abrahamyan,2 Céline Fréchina,1 Andrew J. Mouland,2,3,4* and Luc DesGroseillers1*

Département de biochimie, Université de Montréal,1 HIV-1 RNA Trafficking Laboratory, Lady Davis Institute for Medical Research-Sir Mortimer B. Davis Jewish General Hospital,2 Departments of Microbiology & Immunology,3 Medicine, McGill University, Montréal, Québec, Canada4

Received 9 February 2007/ Accepted 2 April 2007


   ABSTRACT
Top
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Human immunodeficiency virus type 1 (HIV-1) requires the sequential activities of virus-encoded proteins during replication. The activities of several host cell proteins and machineries are also critical to the completion of virus assembly and the release of infectious virus particles from cells. One of these proteins, the double-stranded RNA-binding protein Staufen1 (Stau1), selectively associates with the HIV-1 genomic RNA and the viral precursor Gag protein, pr55Gag. In this report, we tested whether Stau1 modulates pr55Gag assembly using a new and specific pr55Gag oligomerization assay based on bioluminescence resonance energy transfer (BRET) in both live cells and extracts after cell fractionation. Our results show that both the overexpression and knockdown of Stau1 increase the pr55Gag-pr55Gag BRET levels, suggesting a role for Stau1 in regulating pr55Gag oligomerization during assembly. This effect of Stau1 on pr55Gag oligomerization was observed only in membranes, a cellular compartment in which pr55Gag assembly primarily occurs. Consistently, expression of Stau1 harboring a vSrc myristylation signal led to a 6.5-fold enrichment of Stau1 in membranes and a corresponding enhancement in the Stau1-mediated effect on pr55Gag-pr55Gag BRET, demonstrating that Stau1 acts on assembly when targeted to membranes. A role for Stau1 in the formation of particles is further supported by the detection of membrane-associated detergent-resistant pr55Gag complexes and the increase of virus-like particle release when Stau1 expression levels are modulated. Our results indicate that Stau1 influences HIV-1 assembly by modulating pr55Gag-pr55Gag interactions, as shown in a live cell interaction assay. This likely occurs when Stau1 interacts with membrane-associated assembly intermediates.


   INTRODUCTION
Top
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Human immunodeficiency virus type 1 (HIV-1) assembly is a process that is defined as the formation of new infectious viral particles. Multimerization of approximately two to five thousand copies of the main structural protein pr55Gag in the cell is believed to be the driving force behind virus assembly, and pr55Gag expression alone is sufficient for the formation and release of virus-like particles (VLPs) from cells (18, 41). pr55Gag is a modular protein composed of distinct functional domains, which are matrix (MA), capsid (CA), nucleocapsid (NC), p6, and two spacer peptides (p2 and p1) (41). While many details surrounding the process of HIV-1 assembly are unclear, recently published work supports the idea that it is tightly regulated in time and space (18, 41, 53) and moreover depends not only on virus-encoded proteins but also on the activities of cellular cofactors (18, 22).

pr55Gag assembly is believed to be initiated in the cytosol by the formation of detergent-resistant assembly complexes (30). The molecular determinant responsible for some of these early pr55Gag assembly events likely resides in the NC domain. Indeed, NC has been shown to be the minimal pr55Gag subdomain required for self-association in yeast (51). Consistently, replacement of NC by a heterologous dimerization domain, such as a leucine zipper or another zinc finger motif, can reproduce the first steps of pr55Gag assembly (35, 52). Thus, NC dimerization likely serves to initiate the assembly of complete capsids. The carboxy-terminal domain of CA is also important during HIV-1 assembly. Indeed, mutations within the CA dimer interface severely compromise viral particle production (8, 25, 39, 49). In addition, CA influences HIV-1 assembly through the major homology region (6, 40) and the CA-p2 junction (1, 31). Therefore, pr55Gag multimerization is a multistep mechanism that involves several pr55Gag subdomains. Following the initial assembly events, pr55Gag complexes are directed to membranes, where most of the viral assembly presumably takes place (43, 47). However, very little is known about how the spatiotemporal regulation of pr55Gag multimerization is controlled by these determinants.

In an attempt to better understand viral replication and to discover new therapeutic targets, there is an increasing interest in dissecting the molecular pathways and identifying cellular proteins involved in HIV-1 assembly and/or release (22). For instance, the adaptor proteins AP-3 and AP-2 and the trans-Golgi network-associated proteins POSH-1 and Rab9 GTPase are known to be important for pr55Gag intracellular trafficking and virus production (3, 5, 14, 37). Moreover, the ATP-binding protein ABCE1/HP68 associates with specific viral assembly intermediates via its interaction with the NC domain of pr55Gag and is involved in the formation of complete capsids (53). Finally, virus budding and release involve the recruitment of the components of the endosomal sorting complexes, such as Tsg101 and AIP-1, by the p6 domain of pr55Gag (18).

The double-stranded RNA-binding protein Staufen1 (Stau1) (33, 50) is another host protein shown to interact with pr55Gag (9). The Stau1 gene encodes several isoforms by differential splicing, giving rise to 63,000-molecular-weight (63K) and 55K proteins that differ at their amino termini (50). At least a portion of Stau1 can be purified in ribonucleoprotein complexes that contain ribosomes (7, 32, 48). Recent studies indicate that Stau1 is involved in various cellular processes related to RNA, such as translation of a population of repressed mRNAs (17), degradation of specific mRNAs when bound in their 3' untranslated region (27), and dendritic transport of RNA granules in neurons (26). During HIV-1 expression, Stau1 coprecipitates in a complex that contains the 9-kb genomic RNA and associates with the NC domain of pr55Gag in an RNA-independent manner (9). Stau1 is also selectively encapsidated into HIV-1 particles (36). The biological relevance of these interactions is highlighted by the fact that both Stau1 overexpression and Stau1 depletion using RNA interference (RNAi) lead to impaired infectivity of neosynthesized viral particles (9, 36). Thus, Stau1 may be an important cofactor for viral replication and be involved in replication steps that depend on NC functions such as that in pr55Gag assembly.

In this report, we evaluated how Stau1 expression levels influence pr55Gag-pr55Gag interaction and therefore HIV-1 assembly. Our proximity-based interaction assay, which relies on bioluminescence resonance energy transfer (BRET), allowed us to study this both in live cells and in extracts after cell fractionation. Our results show that (i) Stau1 and pr55Gag interact in both membrane-associated and cytosolic complexes, (ii) Stau1 overexpression and depletion by RNAi modulate pr55Gag multimerization, and (iii) these treatments to modulate Stau1 expression levels promote VLP production and an accumulation of detergent-resistant pr55Gag complexes in membranes that are likely a consequence of enhanced pr55Gag multimerization in this system. These results demonstrate that Stau1 participates in HIV-1 assembly and may represent a potential target for intervention because of its impact on this process.


   MATERIALS AND METHODS
Top
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Cell culture and reagents. Human embryonic kidney fibroblasts (293T) were cultured in Dulbecco's modified Eagle medium (Invitrogen) supplemented with 10% cosmic calf serum (HyClone) and 1% penicillin-streptomycin antibiotics (Multicell). Transfections were carried out using either Lipofectamine 2000 reagent (Invitrogen) or the calcium phosphate precipitation method. For RNAi experiments, cells were transfected once with the silencing plasmid and again 16 h later with the pr55Gag expresser pCMV-GagM1-10 (44) and the silencing or nonsilencing (NS) control plasmid. For Western blots, mouse, rabbit, and goat horseradish peroxidase-coupled secondary antibodies were purchased from Dako Cytomation, and signals were detected using Western Lightning Chemiluminescence Reagent Plus (PerkinElmer Life Sciences). Signals were detected with a Fluor-S MultiImager apparatus (Bio-Rad) and quantitated with the Quantity One (version 4.5) software (Bio-Rad).

Plasmid construction. The construction of pCDNA3-RSV-Stau155-HA, pCDNA3-RSV-Stau1F135A-HA, pCMV-pr55Gag-Rluc, and pCMV-NC-Rluc plasmids as well as silencing short hairpin 1 (sh1) and NS RNA-expressing plasmids has been reported before (9, 17, 36, 50). pCMV-Myc-TRBP2 plasmid was kindly provided by Anne Gatignol. The protease-defective HIV provirus HxBRU PR- and the Rev-independent pr55Gag expresser were previously described (36, 44).

To construct the pCMV-NC-p1-YFP plasmid, pr55Gag cDNA was PCR amplified from the Rev-independent pr55Gag expresser by use of the 5'-ACAGGTACCATGCAGAGAGGCAATTTTA-3' sense primer and the 5'-ACAGGATCCCCTCCAAAATTCCCTGGCCTTCC-3' antisense primer with Pfu Turbo (Stratagene). PCR products were purified, digested with BamHI and KpnI restriction enzymes (Fermentas), and cloned into BamHI/KpnI-digested pCMV-GFP-Topaz. The same strategy was used to construct pCMV-pr55Gag-YFP by use of the 5'-ACAGGTACCATGGGTGCGAGAGCGTCAG-3' sense primer and the 5'-ATCAGGATCCCCACCGCCACCTCCTCCTTGTGACGAGGGGTCGTT-3' antisense primer as well as pCMV-MA-CA-YFP by use of the 5'-ACAGGTACCATGGGTGCGAGAGCGTCAG-3' sense primer and 5'-ACAGGATCCCCTCCTCCCAAAACTCTTGCCTTATG-3'. The expressers of pr55Gag WM184-185AA-YFP, MA-CA WM184-185AA-YFP, and MA-CA WM184-185AA-Rluc were generated using the PCR all-around technique (34) with pCMV-pr55Gag-YFP, pCMV-MA-CA-YFP, and pCMV-MA-CA-Rluc (9) as respective templates and the primer pair 5'-AATGCTGCTACAGAAACCTTGTTG-3' and 5'-TGTAGCAGCATTTTTTACCTCCTG-3'. The PCRs were carried out with Phusion enzyme (New England Biolabs) and subjected to 18 PCR cycles (95°C for 50 s, 55°C for 60 s, and 72°C for 90 s). The resulting products were incubated with 10 units of DpnI enzyme (Fermentas) in order to digest methylated parental plasmid and then transformed into competent bacteria. Positive clones containing the mutation(s) were screened by restriction and sequencing analyses.

To construct the (vSrc)Myr-Stau155-HA expresser plasmid, the Stau155-HA coding sequence was PCR amplified using sense 5'-GATGCTCGAGACCATGGGAAGTTCAAAAAGCAAGCCTAAAGGAAAACTTGGAAAAAAACC-3' and antisense 5'-CACATCTAGATCATTTATTCAGCGGCCGCACTGAGCAGCGT-3' primers. The vSrc myristylation signal was included in the sense primer in frame with the 5' extremity of the Stau155 coding sequence. The resulting PCR fragment was purified, digested with XhoI and XbaI restriction enzymes (Fermentas), and cloned into the XhoI/XbaI cassette of pcDNA3.1(+)-Zeo (Invitrogen). For comparative analyses, Stau155-HA cDNA was also subcloned into the KpnI/XbaI cassette of pcDNA3.1(+)-Zeo.

BRET analyses. BRET experiments with live cells were performed as described before (9). Briefly, proteins of interest were separately fused to Renilla reniformis luciferase (Rluc) and yellow fluorescent protein (YFP) and expressed in 293T cells by use of transfection. Forty-eight hours posttransfection, cells were washed and diluted to 106 cells/ml. Coelenterazine H (NanoLight Technology) was added to the cells at a final concentration of 5 µM. Luminescence (440- to 500-nm) as well as total and transmitted fluorescence (510- to 590-nm) emissions was measured using a Fusion {alpha}-FP apparatus (Perkin-Elmer). The BRET ratio was defined as [(emission at 510 to 590 nm) – (emission at 440 to 500 nm) x Cf]/(emission at 440 to 500 nm), where Cf corresponds to (emission at 510 to 590 nm)/(emission at 440 to 500 nm) when Rluc-fused protein is expressed alone. The total YFP activity/Rluc activity ratio reflects the relative levels of activity of the two pr55Gag fusion proteins in the cells. The BRET ratio increases with the total YFP activity/Rluc activity ratio, since more pr55Gag-YFP molecules bind to pr55Gag-Rluc. A saturation curve can be drawn when pr55Gag-Rluc becomes saturated by pr55Gag-YFP molecules, and an optimal BRET ratio can be calculated at saturation for a specific protein-protein interaction. However, since the BRET ratio values at saturation may vary with the transfection protocol, each experiment was done with all the internal positive and negative controls. Using these controls, we first showed that the relative BRET ratios (increase [n-fold]) calculated for different proteins are similar regardless of the transfection protocol. Therefore, to compare results from independent sets of experiments, values of the BRET induction level were used instead of the BRET ratios per se.

Following cell fractionation, the BRET ratio was determined for 90 µl of each fraction by measuring luminescence after the addition of coelenterazine H as well as total and transmitted fluorescence. BRET ratios in fractions 1, 4, 5, and 6 from the membrane flotation assay were not considered because luciferase activity in these fractions was too low to provide a reliable BRET ratio.

VLP release analysis. 293T cells (confluence of 50%) were transfected as described above in order to express pr55Gag in a Rev-independent manner and to modulate Stau1 cellular levels. Twenty hours posttransfection, supernatants were collected and cleared with a 0.45-µm filter. VLPs were pelleted by ultracentrifugation at 220,000 x g through a sucrose cushion (20% in Tris-NaCl buffer). Pelleted VLPs were resuspended in Tris-NaCl buffer, and 2 to 5% of purified VLPs was analyzed by Western blotting using mouse monoclonal anti-CA antibodies. VLP-producing cells were washed, collected, and lysed using NP-40-containing lysis buffer. Cell lysates were analyzed for Stau1 and pr55Gag content by Western blotting using mouse monoclonal anti-Stau1 and anti-CA antibodies.

Membrane flotation assays and S100/P100 fractionation. Forty hours posttransfection, cells were washed twice with phosphate-buffered saline, collected, and homogenized in 400 µl of TE (10 mM Tris [pH 7.4], 1 mM EDTA [pH 8]) containing 10% sucrose and protease inhibitor (Roche) by performing 24 passages through a 23G1 needle or using a Dounce homogenizer. Nuclei were removed by centrifugation at 1,000 x g. Two hundred fifty microliters of lysate was mixed with 1.25 ml of TE 85.5% sucrose (adjusting the concentration of sucrose at 73%) and deposited at the bottom of a 5-ml centrifugation tube. TE 65% sucrose (2.5 ml) and then 1 ml of TE 10% sucrose were layered at the surface of the lysate. The samples were submitted to ultracentrifugation at 100,000 x g for at least 14 h at 4°C using either an AH650 rotor (Sorvall) or an SW55Ti rotor (Beckman Coulter). Nine fractions of 550 µl were collected from the top. An aliquot of each fraction was resolved on 10 to 12% sodium dodecyl sulfate-containing acrylamide gels and then analyzed by Western blotting using the mouse monoclonal anti-{alpha}1 subunit of sodium-potassium (Na-K) ATPase (generously provided by Michel Bouvier), mouse anti-ribosomal protein L7 (Novus Biologicals), mouse monoclonal anti-Stau1 (17), mouse monoclonal anti-hemagglutinin (anti-HA) (ascite #12CA5), and mouse monoclonal anti-HIV-1 CA (36).

For S100/P100 fractionation, 100 µl of TE 10% sucrose and 100 µl of fractions 2 and 3 (membranes) from the membrane flotation assay were pooled. Fifteen microliters of water or of Triton X-100 20% was added (final concentration of 0.5%). The samples were incubated at room temperature for 5 min and then ultracentrifuged at 100,000 x g for 1 h at 4°C. Supernatants (S100) were collected and pellets (P100) were resuspended in 300 µl of TE 28% sucrose (same sucrose concentration as in S100). The same amounts of S100 and P100 were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and analyzed by Western blotting using mouse anti-CA, mouse anti-Na-K ATPase {alpha}1 subunit, mouse anti-Stau1, and mouse anti-HA antibodies.

Immunoprecipitation. Following the membrane flotation assay, 200 µl of TE and 200 µl of membrane fractions 2 and 3 were mixed, whereas 200 µl of the last three collected fractions, fractions 7, 8, and 9, were pooled. Thirty microliters of Triton X-100 20% was added (1% final concentration), and samples were incubated for 15 min at room temperature to release pr55Gag complexes from membranes. Each pool was precleared for 1 hour with protein A-Sepharose (Amersham) and then subjected to immunoprecipitation using 60 µl of anti-HA affinity matrix (Roche) for 3 h at 4°C. The pellets were washed four times with TE and analyzed by Western blotting using anti-CA and anti-HA antibodies.

For immunoprecipitation of endogenous Stau1 proteins, purified mouse monoclonal anti-Stau1 antibodies were cross-linked to protein A-Sepharose (Amersham) by use of dimethylpimelimidate (Pierce Biotechnology). Pooled fractions were treated with 0.25% Triton X-100 for 3 min and then precleared with 45 µl of protein A-Sepharose for 1 h at 4°C. Stau1 proteins were immunoprecipitated using 45 µl of protein A-cross-linked anti-Stau1 antibodies for 2 h at 4°C. Immune complexes were washed three times with TE and eluted from the resin with 2 volumes of triethylamine 100 mM (pH 10.5). Extracts were neutralized by adding 1/10 volume of Tris 1 M (pH 6.8). Purified proteins were analyzed by Western blotting using mouse anti-CA, human anti-ribosomal protein P0 (kindly provided by M. Reichlin), and mouse anti-Stau1 antibodies.


   RESULTS
Top
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
BRET assay as a measure of pr55Gag-pr55Gag multimerization in live cells. To study pr55Gag-pr55Gag interactions in live cells, we used the BRET technology (see Materials and Methods). When YFP is in close proximity to Rluc (≤100 Å) as a consequence of protein interaction (in this case, pr55Gag-pr55Gag), resonance energy is transmitted from the donor Rluc to the acceptor YFP, which in turn emits a measurable fluorescence (9). For these BRET studies, pr55Gag was fused to either Rluc or YFP (Fig. 1A). Western blot analysis confirmed that the fusion proteins were well expressed in transfected 293T cells and that the proteins migrated at their predicted molecular weights (Fig. 1B). By transfecting constant amounts of plasmids expressing pr55Gag-Rluc and increasing amounts of plasmids coding for pr55Gag-YFP in an HIV-1 Rev-independent manner, a saturation curve can be generated and an optimal BRET ratio calculated for a specific interaction (see Materials and Methods). Using this assay, a specific BRET ratio of about 0.17 at saturation (Fig. 1C) was calculated, showing that the pr55Gag-Rluc-pr55Gag-YFP interactions can be studied in live cells.


Figure 1
View larger version (29K):
[in this window]
[in a new window]

  		FIG. 1. BRET assay as a sensor of pr55Gag multimerization. (A) Schematic representation of Rluc- and YFP-fused Gag proteins. The star shape represents the Rluc tag, whereas the cylinder corresponds to the YFP tag. The diamond indicates the WM184-185AA mutation within the carboxy-terminal domain of CA. (B) 293T cells were transfected with expressers of these fusion proteins by use of Lipofectamine 2000. Twenty-four hours posttransfection, cells were lysed and Gag proteins were analyzed by Western blotting using anti-CA antibodies. Numbers above the gel correspond to the proteins schematized in panel A. MW, molecular weights in thousands. (C) YFP-tagged Gag proteins were tested for their capacities to interact with the full-length pr55Gag-Rluc in live cells. BRET saturation curves were obtained by transfecting 293T cells with constant amounts of pCMV-pr55Gag-Rluc and increasing amounts of different YFP-tagged Gag protein expressers. Twenty hours posttransfection, Rluc activity as well as transmitted and total YFP activities was measured for live cells. BRET ratios were plotted as a function of their corresponding "total YFP/Rluc" ratios, which allows comparison of BRET ratios at the same relative levels of expression of Gag fusion proteins. (D) pr55Gag and MA-CA WM184-185AA proteins were tested for their capacities to homodimerize in live cells. BRET ratios were measured at saturation.

 
To confirm that the BRET assay reflects pr55Gag multimerization, the BRET ratio of pr55Gag-YFP was compared to those obtained with three Gag mutants known to be severely defective in pr55Gag assembly (2, 8, 25, 39). First, pr55Gag WM184-185AA-YFP was generated by changing for alanines two residues (Trp184 and Met185) in the dimer interface of the CA carboxy-terminal domain (49) (Fig. 1A). Although the mutant was severely defective in pr55Gag assembly, detectable particle release was observed. Second, MA-CA-YFP was generated by the deletion of the carboxy terminus of pr55Gag (including NC, p2, and p6 domains). Third, since complete inhibition of pr55Gag multimerization was seen when these two sets of mutations were combined in the same pr55Gag molecule (8), MA-CA WM184-185AA-YFP harboring both mutations was constructed. When transfected along with pr55Gag-Rluc, pr55Gag WM184-185AA-YFP and MA-CA-YFP both produced a reduced BRET ratio at saturation (Fig. 1C). Interestingly, the mutant with combined mutations generated the same BRET ratio as that obtained for our negative control condition (YFP alone). Similarly, homodimerization of MA-CA WM184-185AA-Rluc with MA-CA WM184-185AA-YFP resulted in a negligible BRET ratio (Fig. 1D). Western blot analysis confirmed that the fusion proteins were well expressed in transfected 293T cells and that the proteins migrated at their predicted molecular weights (Fig. 1B). Therefore, the BRET ratios obtained with the Gag assembly mutants indicated severe defects in Gag-Gag assembly as reported in earlier work and that this BRET assay faithfully reflects the ability of Gag to multimerize in live cells.

Stau1 overexpression modulates NC-mediated pr55Gag-pr55Gag interaction in live cells. We then tested a putative role of Stau1 during virus assembly, a step that is partly driven by the NC domain. We tested if HA-tagged Stau155 overexpression could alter the pr55Gag BRET ratio by modulating pr55Gag self-interaction. Therefore, Stau155-HA was coexpressed with pr55Gag fusion proteins and a saturation curve was generated to compare the BRET ratios at the same relative levels of expression of pr55Gag fusion proteins (same "total YFP/Rluc" ratio) (Fig. 2A). When Stau155-HA was overexpressed, the BRET ratio was increased compared to that obtained with the empty vector as a control. An average 2.57 (± 0.57 standard deviation [SD])-fold increase in the optimal pr55Gag-pr55Gag BRET ratio was observed (n = 17) (Fig. 2B), suggesting that Stau155-HA overexpression can modulate pr55Gag multimerization and then assembly. As controls for specificity, expression of chloramphenicol acetyltransferase, of a Stau1 mutant (Stau1F135A) that does not bind pr55Gag (9), and of TAR RNA-binding protein (TRBP2) (16), a member of the double-stranded RNA-binding protein family to which Stau1 belongs, did not significantly increase the pr55Gag-pr55Gag BRET ratio at saturation compared to mock conditions (empty vector) (Fig. 2A to D). These results show that the levels of expression of Stau1 influence pr55Gag oligomerization and that the effect on pr55Gag-pr55Gag interaction is specific to the overexpression of wild-type Stau1. They also suggest that a functional Stau1 double-stranded RBD3 known to be involved in pr55Gag association is required.


Figure 2
View larger version (20K):
[in this window]
[in a new window]

  		FIG. 2. Stau1 enhances pr55Gag assembly in live cells. (A) BRET saturation curves were performed by transfecting 293T cells with constant amounts of pCMV-pr55Gag-Rluc and increasing amounts of pCMV-pr55Gag-YFP. A third plasmid expressing Stau155-HA, StauF135A-HA, or chloramphenicol acetyltransferase (CAT) was included in the transfection procedure. Rluc activity as well as transmitted and total YFP activities was measured. BRET ratios were plotted as a function of their corresponding "total YFP/Rluc" ratios, which allows comparison of BRET ratios at the same relative levels of pr55Gag fusion proteins. (B) The value of the optimal BRET ratio was compared to the ratio corresponding to the pr55Gag fusions expressed alone. The BRET induction levels were then determined and are shown in the graph. Numbers of replicates (n) are shown above. (C) pr55Gag-pr55Gag BRET saturation experiment with 293T cells which express Stau155-HA or Myc-TRBP2. (D) Cellular expression of Stau155-HA and Myc-TRBP2 corresponding to the three last conditions of each curve from panel C was analyzed by Western blotting using anti-HA and anti-Myc antibodies, respectively. *, nonspecific signals; MW, molecular weights in thousands. (E) NC-Rluc, NC-p1-YFP, and YFP were detected by Western blotting using anti-NC (left panel) and anti-YFP (right panel) antibodies. (F) BRET saturation experiments in 293T cells expressing NC-Rluc and NC-p1-YFP or YFP. Increasing levels of Stau155-HA expresser were cotransfected. These results were representative of at least two experiments.

 
We previously showed that Stau155 binds the NC domain of pr55Gag (9). Therefore, to determine if the Stau1 effect on pr55Gag-pr55Gag interaction is mediated by the NC domain, we tested NC-NC self-interaction in the presence or absence of Stau155-HA by use of the BRET assay. Following transfection of the plasmids, NC-Rluc and NC-p1-YFP fusion proteins were well expressed and migrated at the expected molecular weights on a Western blot (Fig. 2E). In the absence of Stau155-HA, the BRET saturation curve in live cells showed an interaction between NC-Rluc and NC-p1-YFP fusion proteins (optimal BRET = 0.03) (Fig. 2F). Interestingly, dose-dependent increases in the BRET ratio were observed when different concentrations of cDNA coding for Stau155-HA were cotransfected with plasmids coding for the fusion proteins (Fig. 2F). These results show that the levels of expression of Stau155-HA modulate NC self-interaction and suggest that the effect of Stau1 on pr55Gag-pr55Gag interaction occurs during NC-dependent processes.

Stau1 depletion increases pr55Gag-pr55Gag interaction in live cells. Our results suggest that Stau1 modulates pr55Gag assembly. To determine if Stau1 is necessary for this function, pr55Gag-pr55Gag BRET analysis was performed for cells in which the amount of Stau155 was knocked down by the expression of a silencing sh1 RNA (17). An 88% reduction in the amounts of both the 55K and 63K Stau1 isoforms was observed (Fig. 3A). Surprisingly, under these conditions, the pr55Gag-pr55Gag BRET profile was increased by approximately twofold (Fig. 3B) suggesting that pr55Gag assembly is modified when low levels of Stau1 are present. An increase in pr55Gag self-interaction was also observed with a short interfering RNA that targeted a different part of Stau1 mRNA (not shown), eliminating the possibility of an off-target effect. As controls, transfection of an NS RNA had no major effect on pr55Gag assembly, whereas overexpression of Stau155-HA markedly increased pr55Gag-pr55Gag BRET ratios (Fig. 3B). Thus, Stau1 overexpression or depletion leads to similar effects on pr55Gag multimerization.


Figure 3
View larger version (29K):
[in this window]
[in a new window]

  		FIG. 3. Stau1 depletion by RNAi promotes pr55Gag multimerization. pr55Gag-pr55Gag BRET saturation curves were obtained for live 293T cells following down-regulation of endogenous Stau1 by RNAi. As controls, the empty vector and an NS RNA were also transfected. (A) Levels of endogenous Stau1 and of pr55Gag fusion proteins were determined by Western blotting using anti-Stau1 and anti-CA antibodies, respectively. Anti-calnexin antibody was used as a loading control. MW, molecular weights in thousands. (B) pr55Gag-pr55Gag BRET ratios are plotted as a function of the "total YFP/Rluc" ratio.

 
Stau1 overexpression increases VLP release. Alteration of pr55Gag multimerization is likely to affect a downstream step(s) in virus assembly. We previously showed that Stau1 overexpression and Stau1 down-regulation impaired virus infectivity (9, 36), genomic RNA encapsidation (36), and virion morphogenesis (L. Abrahamyan and A. J. Mouland, unpublished data). Virus or VLP release is an intermediate step considered as an indirect readout of pr55Gag multimerization and particle formation (2, 25). Therefore, we hypothesized that the Stau1-mediated changes in pr55Gag multimerization seen with the BRET assay should be translated into modification of particle formation as measured by the levels of extracellular VLP release. To test that, VLP release from cells that overexpressed Stau155-HA was first studied. 293T cells were transfected with plasmids that expressed pr55Gag with or without Stau155-HA. Cell extracts and VLPs were prepared 20 h posttransfection and analyzed by Western blotting (see Materials and Methods). In the cell extract, the cellular levels of pr55Gag did not significantly change with the expression of Stau155-HA (Fig. 4A). In contrast, the amounts of pr55Gag in VLP preparation increased by 3.31 (± 0.74 SD)-fold (n = 3) when Stau1 was overexpressed (Fig. 4A and D). This effect of Stau155-HA on VLP production is dose dependent (Fig. 4B). These results further support the idea that Stau1 influences the multimerization of pr55Gag and VLP production.


Figure 4
View larger version (31K):
[in this window]
[in a new window]

  		FIG. 4. Levels of expression of Stau1 affect VLP production. (A) 293T cells were transfected with a Rev-independent pr55Gag expresser, an empty vector, and/or a Stau155-HA expresser. Twenty hours posttransfection, cells were lysed and VLPs purified by ultracentrifugation. Cell extracts and VLPs were analyzed by Western blotting using anti-Stau1, anti-CA, and anti-calnexin antibodies as indicated. Calnexin signals were used to normalize pr55Gag levels in the cell extracts. MW, molecular weights in thousands. (B) Overexpression of Stau155-HA increases VLP production in a dose-dependent manner. Same experiment as shown in panel A except that 293T cells were cotransfected with constant amounts of pr55Gag expresser and increasing amounts of the Stau155-HA expresser. (C) Same experiment as shown in panel A except that 293T cells were cotransfected with the pr55Gag expresser and NS or sh1 RNA. (D) The Stau1-mediated effect on pr55Gag levels in cell lysates and VLPs from three independent experiments was calculated and is shown in the graph.

 
VLP release was then studied when Stau1 was down-regulated. Expression of the sh1 RNA reduced Stau1 expression by 91% (Fig. 4C). Consistent with the BRET assay, Stau1 down-regulation also caused an increase in VLP release by 2.67 (± 0.6 SD)-fold (n = 3) (Fig. 4C and D). However, this increase was often partly paralleled by an increase in the intracellular levels of pr55Gag (Fig. 4D). Therefore, both overexpression and down-regulation of Stau1 increase the levels of pr55Gag in the supernatant, although the mechanisms by which this occurs may be different. These phenotypes are likely a downstream consequence of Stau1's influence in pr55Gag multimerization.

Stau1 does not influence HIV-1 pr55Gag membrane association. To study the mechanism by which Stau1 modulates pr55Gag multimerization and particle formation, we first tested if Stau1 is involved in pr55Gag subcellular localization, a step that may affect assembly. Shortly after synthesis, pr55Gag assembles on membranes (13, 43, 47). Therefore, at steady state, most of the pr55Gag is observed within membranes (23, 24, 38). We first tested whether pr55Gag membrane association is modulated in the presence or absence of Stau1. 293T cells were cotransfected with either Stau155-HA or the silencing sh1 RNA expresser and the protease-defective HxBRU provirus. This provirus generates only the unprocessed species of pr55Gag and therefore facilitates the interpretation of results (23, 38, 47). Compared to endogenous Stau1 in control cells (mock, empty vector, or NS RNA), Stau155-HA was overexpressed fivefold, whereas the sh1 RNA expresser knocked down the expression of Stau1 by 65% (Fig. 5A). Cell extracts were then fractionated by the membrane flotation assay in order to separate membrane-free from membrane-associated complexes (23, 47). After centrifugation, nine fractions were collected and their protein contents analyzed by Western blotting (Fig. 5B). Antibodies against the {alpha}1 subunit of sodium-potassium (Na-K) ATPase and the ribosomal protein L7 were used as markers for membrane (fractions 2 and 3) and cytosol (fractions 6 to 9), respectively. In mock conditions (when an empty vector or the NS RNA plasmid was transfected), about 85% of the total pr55Gag was found in the membrane fractions (Fig. 5B and C), as reported earlier (23, 47). Overexpression of Stau155-HA or down-regulation of endogenous Stau1 expression by sh1 RNA did not significantly modulate pr55Gag membrane association (Fig. 5B and C). Similar conclusions were generated following cytoplasm/membrane separation on iodixanol continuous gradients (data not shown) (47). Moreover, the distribution of pr55Gag was primarily at the plasma membrane in protease-defective HxBRU provirus-expressing 293T cells, as determined by confocal laser scanning microscopy, and there was no marked or detectable change in this when Stau1 was depleted or overexpressed in these conditions (data not shown). Therefore, these results indicate that the effect of Stau1 on pr55Gag multimerization is not attributable to any marked change in the cellular distribution of pr55Gag or in the impairment of its association with membranes.


Figure 5
View larger version (29K):
[in this window]
[in a new window]

  		FIG. 5. Stau1 overexpression or depletion by RNAi does not change pr55Gag association with membranes. 293T cells were transfected with plasmids expressing either the empty vector, Stau155-HA, sh1 RNA, or the NS RNA. The next day, cells were transfected with the protease-defective HxBRU PR- provirus. (A) An aliquot of each cell extract was analyzed by Western blotting using anti-Stau1 antibodies. As a loading control, anti-calnexin antibodies were used. (B) Cytoplasmic extracts were fractionated by the membrane flotation assay. Nine fractions were collected and analyzed by Western blotting using anti-Na-K ATPase, anti-ribosomal protein L7, and anti-CA antibodies. M, membranes; Cy, cytosol. (C) The pr55Gag signal in each fraction was quantified and plotted as a percentage of the total pr55Gag signal in the gradient.

 
Stau1 interacts with pr55Gag within membrane and cytosolic complexes. pr55Gag assembly appears to occur mostly within membranes (13, 43). On the other hand, Stau1 is known to be primarily cytosolic. Consequently, in order to determine whether Stau1 stimulates pr55Gag-pr55Gag interactions during assembly in membranes or whether it modulates an earlier assembly step in the cytoplasm, we first tested if endogenous Stau1 is at least partly present in membranes. Extracts of nontransfected 293T cells were separated by the membrane flotation assay (23, 47) and each fraction was blotted with anti-Stau1 antibodies (Fig. 6A, top panel). As reported before (32), Stau1 was detected mainly in cytosolic fractions. However, a proportion of Stau1 representing 2 to 5% of the total Stau1 protein was found in the membrane compartment. Its association with membranes or membrane-associated complexes was confirmed by treating the cell homogenates with detergent prior to fractionation. In these conditions, membranes are solubilized and membrane-associated proteins redistribute in the cytosolic fractions. Stau1 and the membrane-associated control protein, sodium-potassium ATPase, were no longer seen in fractions 2 and 3 (Fig. 6A, bottom panel), demonstrating that the fraction of Stau1 that redistributed to the cytosolic compartment was indeed membrane associated. Similarly, a fraction of overexpressed Stau155-HA was observed in the membrane fraction (Fig. 6B, top panel). As expected, the signal disappeared following treatment of the cell extracts with detergent prior to centrifugation (Fig. 6B, bottom panel, and C). The proportion of Stau155-HA in membranes versus cytosol was found to be increased severalfold compared to the distribution of endogenous Stau1 (15% versus 2 to 5%, respectively). These results show that cellular Stau1 is present in membrane-associated complexes as well as in the cytosol.


Figure 6
View larger version (23K):
[in this window]
[in a new window]

  		FIG. 6. A subset of Stau1 is membrane associated. Extracts from untransfected cells (A) or cells expressing Stau155-HA (B) were treated (bottom panels) or not (top panels) with 1% NP-40 to disrupt membranes and then submitted to the membrane flotation assay. Each collected fraction was analyzed by Western blotting using anti-Stau1 antibodies. Distribution of the membrane marker Na-K ATPase is shown to ensure the efficiency of NP-40 treatment. MW, molecular weights in thousands. (C) Quantification of the percentage of Stau155-HA signal in each fraction of the gradient shown in Fig. 4B. M, membranes; Cy, cytosolic.

 
To identify the subcellular compartment in which Stau1 interacts with pr55Gag, 293T cells were transfected with the Rev-independent pr55Gag expresser in the presence or absence of a plasmid coding for Stau155-HA. Membrane and cytosolic fractions were prepared in step gradients as described above. Stau155-HA expression did not change the pr55Gag subcellular distribution in the context of VLP production (Fig. 7A). Then, membrane and cytosolic fractions were pooled and subjected to immunoprecipitation using monoclonal anti-HA antibodies in analysis for Stau155-HA-containing complexes in each pool. Identification of proteins in the immunoprecipitated complexes was performed by Western blotting. As expected, when the anti-HA antibody was used, Stau155-HA was immunoprecipitated from both the cytosol and membrane fractions (Fig. 7B). pr55Gag was also successfully coprecipitated with Stau155-HA in each pool (Fig. 7B). As a control, pr55Gag was not detected in the precipitate when Stau155-HA was not expressed. This interaction between Staufen and pr55Gag is likely to be direct based on our BRET data and to be RNA independent, since the interaction was maintained when we treated cell extracts with RNAses (9). We next tested whether endogenous Stau1 is also associated with pr55Gag in membranes and cytosol. 293T cells were transfected with pr55Gag or, as a control, with a ß-galactosidase (ß-Gal) expresser. Cytosolic and membrane fractions were prepared as before. Anti-Stau1 antibody that was cross-linked to the protein A-Sepharose was added to the pooled fractions to immunoprecipitate endogenous Stau1 (Fig. 7C). To control for immunoprecipitation efficiency, anti-Stau1 and anti-ribosomal protein P0 antibodies were used for Western blotting, since Stau1 is a ribosome-associated protein (32). As seen in Fig. 7C, Stau1 and P0 were found in the precipitates from cytosol and membranes. A band of 55K corresponding to pr55Gag was also detected as a coimmunoprecipitating protein in both the membrane and cytosolic fractions when pr55Gag but not ß-Gal was expressed (Fig. 7C). Similarly, no specific signal was observed when anti-Stau1 was omitted (Fig. 7C). These results demonstrate that Stau1 and pr55Gag interact within both membrane-associated and cytosolic complexes.


Figure 7
View larger version (34K):
[in this window]
[in a new window]

  		FIG. 7. Stau1 associates with pr55Gag in the cytosol and in membranes. 293T cells were transfected with pr55Gag-expressing plasmid and either pcDNA3-RSV (empty vector) or pcDNA3-RSV-Stau155-HA. Cytoplasmic extracts were prepared and submitted to the membrane flotation assay. (A) An aliquot of each collected fraction was analyzed by Western blotting using anti-Na-K ATPase, anti-ribosomal protein L7, anti-CA, and anti-HA antibodies. (B) Fractions 2 and 3 (membranes [M]) and fractions 7 to 9 (cytosolic [Cy]) were pooled. Triton X-100 was added to release pr55Gag from membranes. Each pool was subjected to immunoprecipitation using anti-HA antibodies. Coimmunoprecipitated proteins were analyzed by Western blotting using anti-CA and anti-HA antibodies. MW, molecular weights in thousands. (C) 293T cells were transfected with pr55Gag or ß-Gal expressers. Following the membrane flotation assay, fractions were pooled as for panel B and treated with Triton X-100. Proteins in each pool were immunoprecipitated (IP) using protein A-Sepharose alone (A) or anti-Stau1 antibodies which were cross-linked to protein A-Sepharose (St). Pelleted material was analyzed by Western blotting using anti-CA, anti-ribosomal P0, and anti-Stau1 antibodies. *, nonspecific signals.

 
Stau1 modulates pr55Gag multimerization within membrane-associated complexes. At steady state, most of the pr55Gag is observed within membranes, where HIV-1 primarily assembles (13, 24, 43). Because the Stau1-pr55Gag interaction occurs in both membrane and cytosolic compartments, we wanted to determine in which of these compartments Stau1 was exerting its action on the pr55Gag-pr55Gag interaction that we observed earlier (Fig. 2 and 3). Therefore, 293T cells were transfected with pr55Gag-Rluc and pr55Gag-YFP expressers in the presence or absence of the Stau155-HA expresser. Wild-type pr55Gag was cotransfected with these plasmids since it is known to rescue the phenotype of tagged versions of pr55Gag that are mislocalized (29). Cell extracts were tested by Western blotting for the expression of the proteins (Fig. 8A) and then separated by membrane flotation assay. Following centrifugation, Western blot analyses showed that pr55Gag fusion proteins were present mostly in fractions with membranes, as expected (data not shown). Then, fractions containing membranous or cytosolic materials were used to measure pr55Gag-pr55Gag association by BRET (Fig. 8B). pr55Gag-pr55Gag interaction was detected mainly in fractions containing membranes (BRET ratio in fraction 2 = 0.37 ± 0.05 SD; n = 5), consistent with the fact that pr55Gag assembles on membranes (13, 43, 47). In the cytosolic fractions, the pr55Gag-pr55Gag BRET ratios were negligible, although Rluc activity was readily detectable. The absence of BRET in cytosolic factions was not due to the high percentage of sucrose in these samples, since Stau1 dimerization was readily detectable by BRET in these fractions (data not shown). As a control, BRET was not detected when YFP was expressed with pr55Gag-Rluc (Fig. 8B).


Figure 8
View larger version (49K):
[in this window]
[in a new window]

  		FIG. 8. Stau1 overexpression and depletion by RNAi both increase pr55Gag assembly in membranes but not in the cytosol. (A) 293T cells were cotransfected with pr55Gag-Rluc, pr55Gag-YFP, pr55Gag, and Stau155-HA expressers. An aliquot of each cell extract was analyzed by Western blotting for Stau1 and pr55Gag content. As a loading control, anti-ribosomal protein L7 antibodies were used. (B) Cytoplasmic extracts were submitted to the membrane flotation assay and pr55Gag-pr55Gag BRET ratio was determined for each collected fraction. (C) 293T cells were transfected with the empty vector or the NS or sh1 RNA-expressing plasmids. The next day, cells were cotransfected with the same plasmids and pr55Gag-Rluc, pr55Gag-YFP, and pr55Gag expressers. Cytoplasmic extracts were prepared and analyzed as for panel A. As a loading control, anti-calnexin antibodies were used. (D) Cytoplasmic extracts were submitted to the membrane flotation assay, and the pr55Gag-pr55Gag BRET ratio was determined for each collected fraction. M, membrane; Cy, cytosol.

 
When Stau155-HA was overexpressed, an increase of the pr55Gag-pr55Gag BRET ratio was observed (1.8-fold ± 0.37 SD; n = 2) in membrane fractions only (Fig. 8B), and there was no detectable enhancement in the cytosolic fractions. In membrane extracts, there was no change in the Rluc activity/total YFP activity ratio, showing that the variations in the BRET ratio were not due to changes in the saturation levels of the energy donor by the acceptor (data not shown). Similar conclusions for a Stau1-mediated increase of the pr55Gag-pr55Gag BRET ratio in membranes were reached after separation of cell extracts on a 0 to 21% continuous iodixanol density gradient (data not shown) (47). These results demonstrate that despite Stau1-pr55Gag interaction in both cytosolic and membrane compartments, the effects of Stau1 on pr55Gag-pr55Gag interaction is confined to the membrane, where pr55Gag assembly primarily takes place. This is consistent with the possibility that Stau1 participates in pr55Gag multimerization.

pr55Gag-pr55Gag BRET ratios were also obtained when endogenous Stau1 was knocked down by RNAi. Compared to what was seen for control cells, expression of the sh1 silencing RNA resulted in a 79% decrease in endogenous Stau1 expression in this experiment (Fig. 8C). Following membrane flotation assay (Fig. 8D) or separation on continuous iodixanol density gradients (data not shown), pr55Gag-pr55Gag BRET ratios were calculated for fractions containing membranes or cytosolic materials. When Stau1 was depleted by RNAi, a reproducible 1.47 (± 0.1 SD)-fold (n = 4) increase in the pr55Gag-pr55Gag BRET ratio was observed for fractions corresponding to membranes but not for fractions containing cytosolic materials (Fig. 8D). In contrast, no significant effect on the BRET ratio was observed under NS conditions compared to what was seen for mock-transfected cells. Altogether, these results show that Stau1 affects pr55Gag assembly in the membrane compartments.

Targeting Stau1 to membranes further increases pr55Gag multimerization. Altogether, our results indicate that overexpression or depletion of Stau1 modulates pr55Gag-pr55Gag interaction in membrane fractions. To test if membrane-associated Stau1 is responsible for the increase in pr55Gag multimerization, we fused the vSrc myristylation signal to the amino terminus of Stau155-HA to enhance the targeting to membranes [(vSrc)Myr-Stau155-HA; Fig. 9A ]. (vSrc)Myr-Stau155-HA was expressed in 293T cells, and its subcellular distribution was analyzed by the membrane flotation assay. As seen in Fig. 9B, the fraction of (vSrc)Myr-Stau155-HA in the membrane fractions was increased severalfold compared to that of Stau155-HA. This represented about 50% of the total (vSrc)Myr-Stau155-HA protein (Fig. 9C) and an enrichment of 6.5 (± 0.47 SD)-fold (n = 2) in membrane compared to that of Stau155-HA. As a control, treatment of the postnuclear extract with detergent prior to the membrane flotation assay prevented the presence of (vSrc)Myr-Stau155-HA in membrane fractions, confirming its association with membranes (data not shown). Next, (vSrc)Myr-Stau155-HA was expressed in 293T cells to test its capacity to improve pr55Gag multimerization as measured by the pr55Gag-pr55Gag BRET assay in live cells. Compared to Stau155-HA (1.98 [± 0.21 SD]-fold BRET induction; n = 4), (vSrc)Myr-Stau155-HA further enhanced the pr55Gag-pr55Gag BRET ratio up to a 3.29 (± 0.54 SD)-fold induction (n = 4) (Fig. 9D and E). As controls, Western blot experiments showed that Stau155-HA and (vSrc)Myr-Stau155-HA were expressed to the same levels (not shown). Therefore, this result indicates that the enrichment of Stau1 in membrane compartments substantially enhanced its effect on pr55Gag multimerization, strongly suggesting that Stau1 acts on assembly when it is membrane associated.


Figure 9
View larger version (32K):
[in this window]
[in a new window]

  		FIG. 9. Enrichment of Stau1 on membranes increases pr55Gag multimerization. (A) Schematic representation of (vSrc)Myr-Stau155-HA harboring the 9-amino-acid vSrc myristylation signal. (B) (vSrc)Myr-Stau155-HA-expressing plasmid was transfected in 293T cells. The next day, a membrane flotation assay was performed. Each fraction was analyzed by Western blotting using mouse anti-HA, anti-L7, or anti-Na-K ATPase antibodies. MW, molecular weights in thousands. M, membranes; Cy, cytosol. (C) Quantification of the percentage of (vSrc)Myr-Stau155-HA signal in each fraction. The result was compared to that obtained with Stau155-HA (Fig. 6B and C). These two gradients were done at the same time. (D) pr55Gag-pr55Gag BRET saturation experiments were performed by transfecting 293T cells with empty vector, Stau155-HA, or (vSrc)Myr-Stau155-HA expressers. Western blot analysis revealed that the expression levels of Stau155-HA and (vSrc)Myr-Stau155-HA were similar in this experiment (data not shown). (E) The induction level of the optimal BRET ratio at the same (total YFP/Rluc) ratio for each condition is shown (n = 4).

 
Stau1 increases the density of pr55Gag-containing complexes in membranes. Altogether, our BRET results show that Stau1 modulates pr55Gag multimerization and hence assembly within membranes. As a consequence, Stau1 may contribute to the formation of more-stable pr55Gag-containing complexes and/or complexes of higher densities. To test this, we transfected cells with the HxBRU PR- provirus in the presence or absence of the Stau155-HA expresser and performed membrane flotation assays as described above. As a control, the Stau1 mutant (Stau1F135A) that does not associate with pr55Gag and does not modulate pr55Gag-pr55Gag interaction (Fig. 2A and B) was also expressed. Following centrifugation, fractions 2 and 3 were pooled and membrane-associated materials were subjected to high-speed centrifugation at 100,000 x g for 1 h. The resulting pellets and supernatants were analyzed by Western blotting with anti-CA, anti-Na-K ATPase (as a marker of membranes), and anti-HA antibodies. Proteins detected in the pellet (P100) are insoluble or components of heavy complexes, whereas those in the supernatant (S100) are soluble or part of smaller complexes. In mock conditions (empty vector or NS RNA), pr55Gag was found mainly in the P100 fraction (95%) (Fig. 10A). Overexpression of Stau155-HA or Stau1F135A-HA did not change the pattern of sedimentation of pr55Gag-containing complexes. In these conditions, overexpressed Stau155-HA and Stau1F135A-HA were also found mostly in the pellets at levels corresponding to 88% and 93%, respectively. When endogenous Stau1 was depleted with the sh1 RNA (60% reduction; data not shown), no change in the S100-P100 distribution of pr55Gag was observed, 91% of the protein having been found in the P100 pellet (Fig. 10A). The membrane marker Na-K ATPase was found almost completely in the P100 fraction (88%) (Fig. 10A), meaning that membranes and membrane-associated proteins were pelleted in the P100 fractions in these conditions.


Figure 10
View larger version (33K):
[in this window]
[in a new window]

  		FIG. 10. Both Stau1 overexpression and depletion by RNAi induce the formation of detergent-resistant pr55Gag complexes. 293T cells were transfected with the Stau155-HA-, Stau1F135A-HA-, NS RNA-, or sh1 RNA-expressing plasmids. The next day, cells were cotransfected with the same plasmids and the protease-defective HxBRU PR- provirus. Membrane-associated proteins were isolated by the membrane flotation assay (fractions 2 and 3) and incubated in the absence (A) or presence (B) of Triton X-100. Samples were centrifuged at 100,000 x g for 1 h at 4°C. The resulting supernatants (S) and pellets (P) were analyzed by Western blotting using anti-CA, anti-Stau1, anti-HA, and anti-Na-K ATPase antibodies.

 
Therefore, the presence of pr55Gag in the pellets may be due to the density of the pr55Gag assembly complexes (23) or to their association with membranes. To discriminate between these possibilities, we repeated the same experiment in the presence of Triton X-100 to solubilize membranes before the S100-P100 centrifugation. These experiments were done at room temperature to also solubilize lipid rafts and other membrane compartments that are detergent resistant at 4°C. Proteins in the pellets and supernatants were analyzed by Western blotting as described above (Fig. 10B). In mock (empty vector or NS RNA) conditions, pr55Gag complexes shifted into the S100 fraction, showing that pr55Gag complexes cannot be pelleted under the conditions used here. This might mean that the complexes are not stable or heavy enough to pellet after membrane solubilization. In contrast, when Stau155-HA was overexpressed, about 33% of the pr55Gag complexes was still found in the P100 fractions (Fig. 10B), consistent with only a partial solubilization of pr55Gag complexes by detergent due to enhanced multimerization. Similar results were obtained following Stau1 depletion by RNAi, where about 24% of pr55Gag was found in the pellet (Fig. 10B). Overexpression of Stau1F135A-HA had no effect on pr55Gag sedimentation, which was found completely in the S100 fraction following detergent treatment. Similarly, overexpressed Stau155-HA but not Stau1F135A-HA was partly resistant to solubilization, consistent with the fact that Stau155-HA but not Stau1F135A-HA interacts with pr55Gag. As a control for membrane solubilization, the Na-K ATPase was always shifted to the S100 fraction. Altogether, these results show that Stau1 overexpression or depletion leads to the formation of detergent-resistant pr55Gag complexes in membranes. It parallels the observed increase in the BRET ratios in the same conditions, supporting the notion that the BRET assay reflects pr55Gag multimerization. These results also suggest that pr55Gag multimerization or pr55Gag association with heavy membrane complexes are regulated by Stau1 expression levels in the cell.


   DISCUSSION
Top
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
Stau1's interaction with pr55Gag via the NC domain (9) suggests that its role is related to the functions of this domain of pr55Gag during HIV-1 replication. Multiple functions for NC have been described, including pr55Gag dimerization, cellular RNA recruitment, and selection of the 9-kb genomic RNA for encapsidation as well as functions in reverse transcription and integration (18, 46). In this report, we studied the potential role of Stau1 in pr55Gag-mediated HIV-1 assembly. We developed a new pr55Gag homodimerization assay based on BRET that provides for an exquisitely sensitive measure of protein-protein proximity, allowing us to follow virus assembly in live cells. The BRET measures of pr55Gag multimerization were validated with the use of pr55Gag mutants shown to be defective in multimerization by alternative techniques (2, 8, 25, 39). The reported severity of these mutations for the multimerization of Gag is reflected by a corresponding decrease of the BRET ratios. Also, a Gag mutant with a combined disruption of the CA dimer interface and the NC domain was completely impaired for pr55Gag multimerization (8) and did not generate a positive BRET ratio even when expressed with the wild-type pr55Gag (Fig. 1C and D). On the other hand, the study of these mutants shows that a decrease of about 60% in the BRET ratio at saturation results in a severe impairment of particle release (2, 8, 25, 39). Altogether, these experiments demonstrate that the pr55Gag-pr55Gag BRET assay is a reliable and sensitive tool for the direct study of pr55Gag multimerization and the cellular factors that control HIV-1 assembly in live cells.

Stau1 levels influence pr55Gag assembly. Our results indicate that the levels of expression of Stau1 influence HIV-1 assembly by modulating pr55Gag-pr55Gag interaction, a key event during virus assembly. We demonstrate that Stau1 overexpression and Stau1 depletion both increase the pr55Gag/pr55Gag BRET ratio in live cells. The BRET ratio indicates a change in the packing, the proximity, and/or the orientation of the pr55Gag molecules in the complexes as a result of Stau1 expression. This Stau1-mediated enhancement of pr55Gag multimerization is supported by an increase in VLP production (Fig. 4) and the presence of detergent-resistant pr55Gag-containing complexes in membranes (Fig. 10). This concurs with the observed morphogenesis defects of budding virus particles (Abrahamyan and Mouland, unpublished data) and the loss of virus infectivity when Stau1 is over- or underexpressed during virus assembly (9, 36). Our results cannot be explained by changes in pr55Gag subcellular trafficking that could have influenced its multimerization state. Indeed, both cytoplasmic fractionation assays (Fig. 5) and confocal microscopy analysis (data not shown) indicate that Stau1 expression, while increasing the pr55Gag BRET ratio by about twofold in membranes (Fig. 8), did not change the subcellular distribution of unprocessed pr55Gag following protease-minus proviral expression in 293T cells.

Immunoprecipitation analyses of Stau1 indicate that it interacts with pr55Gag within membranous and nonmembranous cytosolic compartments. Stau1/pr55Gag association may thus occur before pr55Gag association to membranes. Because Stau1 is associated to ribosomes in the cytosol (17, 32), it is possible that the Stau1-pr55Gag interaction occurs cotranslationally or shortly after pr55Gag synthesis, as we have proposed before (9). Nevertheless, despite the evidence for a bona fide interaction between Stau1 and pr55Gag in the cytosolic compartment (this study), the pr55Gag/pr55Gag interaction as determined by BRET analysis was not detectable in this compartment (Fig. 8). Accordingly, the effect of Stau1 on pr55Gag assembly was observed only in membranes (Fig. 8). The low levels of pr55Gag-pr55Gag BRET signal in the cytosol fractions may indicate that Stau1 interacts with pr55Gag monomers in this compartment before being targeted to membranes, where assembly occurs. Alternatively, other viral or cellular cofactors that are present only in membranes and/or are recruited during later steps of virus assembly may be required for Stau1 function in assembly. Finally, the very short transit period of pr55Gag in the cytoplasm following synthesis (43, 47) might not allow us to detect the effect of Stau1 on pr55Gag-pr55Gag interaction in the cytoplasm before pr55Gag's incorporation into membranes. Therefore, we do not exclude the possibility that the primary interaction between Stau1 and pr55Gag occurs in the cytosol. However, the effect of Stau1 on pr55Gag assembly is likely to occur in membranes through a Stau1-pr55Gag complex. This conclusion is strengthened by several experimental data: (i) overexpression and depletion of Stau1 both increase the pr55Gag-pr55Gag BRET ratio in the membrane fractions only and (ii) the pr55Gag-pr55Gag BRET ratio increases proportionally to the amounts of overexpressed Stau155-HA in membranes. Compared to endogenous Stau1, with 2 to 5% of the total Stau1 protein being found in membranes, Stau155-HA shows 15% of the proteins associated with membranes (Fig. 6) and the myristylated form of Stau1 a further enrichment (Fig. 9). As a consequence, a 2-fold or a 3.5-fold increase in the pr55Gag-pr55Gag BRET ratio is observed following overexpression of Stau155-HA or of (vSrc)Myr-Stau155-HA, respectively (Fig. 9).

Although detergent-resistant pr55Gag complexes putatively correspond to primary assembly complex intermediates that have been described for the cytosol (30), recent studies show that pr55Gag associates with membranes shortly after its synthesis (43, 47) and that pr55Gag-pr55Gag interaction occurs in the membrane compartments (13). This is in agreement with our BRET assay results, which demonstrate that the pr55Gag-pr55Gag interaction occurs only in membrane extracts (Fig. 8B and D). The positive effect of Stau1 overexpression on pr55Gag assembly that also occurs within membranes strengthens this conclusion. Since the membrane flotation assay cannot discriminate among plasma membrane, late endosomes, and other vesicular compartments (data not shown), the membrane subcompartment in which Stau1 contributes to pr55Gag multimerization remains to be defined. However, considering that Stau1 is packaged into HIV-1 particles and influences virion assembly, the interaction between Stau1 and pr55Gag likely occurs in a membrane compartment that supports assembly, such as plasma membrane or late endosomes/multivesicular bodies (41).

Both Stau1 depletion by RNAi and overexpression increase pr55Gag multimerization. One intriguing question that has no obvious answer so far is why Stau1 depletion by RNAi and Stau1 overexpression cause nearly identical virus morphogenesis (Abrahamyan and Mouland, unpublished data) and infectivity (9, 36) defects. Here we also show that both Stau1 depletion and Stau1 expression increase the pr55Gag BRET ratio in membrane preparations, although the effect of overexpression is more pronounced (Fig. 8D). The molecular mechanism by which Stau1 influences pr55Gag assembly is not known. The interaction between Stau1 and pr55Gag may be one of the determinants, since a Stau1 mutant that does not interact with pr55Gag has no effect on pr55Gag multimerization and VLP release. In addition, considering that very few molecules of Stau1 are encapsidated into virus particles, the stoichiometry of the Stau1-multimerizing pr55Gag complex may be crucial for normal HIV-1 assembly. In this context, pr55Gag would recruit the optimal number of Stau1 molecules that is needed for assembly. Depletion or overexpression of Stau1 would change this equilibrium and modulate pr55Gag multimerization. Although the resulting phenotypes following Stau1 depletion and overexpression are similar, it is possible that they nevertheless impair different steps during virus assembly. The fact that Stau1 down-regulation slightly increases the levels of expression of pr55Gag in the cells (Fig. 4C) in contrast to Stau1 overexpression is consistent with this hypothesis.

The observation that depletion and overexpression of a protein lead to similar defects during HIV-1 replication was also reported for the host protein, TSG101. Modifying the levels of full-length TSG101 leads to two phenotypes. First, it disrupts the cellular endosomal sorting machinery and induces the formation of aberrant endosomal structures. As a result, lysosomal sorting and degradation of the cell surface epidermal growth factor receptor are impaired (4, 15, 21). Second, in the context of HIV-1 particle production, TSG101 depletion or overexpression blocks viral budding and leads to virus-virus tethering (19-21). The relationship between these two phenotypes remains to be fully elucidated but could rely on the importance of TSG101's precise cellular levels for its functions during the HIV-1 replication cycle.

Role of Stau1 during HIV-1 assembly. This study raises the question of the exact role of Stau1 in pr55Gag multimerization. Stau1 likely associates with pr55Gag in the cytosol and is transported to the membrane, where assembly occurs. Then, since RNA is an important cofactor for assembly (8), Stau1 may recruit in the complexes scaffolding RNA by its capacity to bind double-stranded RNA molecules. Moreover, Stau1 is likely to play a role in genomic RNA trafficking and packaging, since it coprecipitates in a complex that contains the 9-kb genomic RNA but not the spliced RNA species and is incorporated into HIV-1 particles along with genomic RNA (9, 36). Furthermore, Stau1 knockdown leads to enhanced encapsidation (Abrahamyan and Mouland, unpublished data). Its association with unprocessed pr55Gag may coordinate this function, since in vitro studies suggested that pr55Gag multimerization serves as a signal for specific HIV-1 genomic RNA recognition and packaging (42). Thus, the proposed few Stau1 molecules necessary for pr55Gag assembly may help to sort only two genomic RNA molecules per virion through its action on pr55Gag multimerization.

Moreover, if Stau1 is recruited as a member of ribonucleoprotein complexes instead of as a monomer, other Stau1-associated host proteins will also be recruited and will participate in the regulation of HIV-1 assembly (11). It remains unclear if the impairment of virus infectivity and morphogenesis observed following depletion and overexpression of Stau1 (9) (Abrahamyan and Mouland, unpublished data) is a consequence of Stau1-dependent modulation of pr55Gag multimerization or defects at other steps of virus assembly. Through its multiple functions in gene expression, Staufen may regulate pr55Gag cellular levels in the proviral context or the ratio between genomic RNA and pr55Gag and hence the pr55Gag multimerization process. It is conceivable that Stau1-mediated changes in pr55Gag multimerization state lead to an overpacking of Gag molecules within the neosynthesized virus particles. This defect in intraviral pr55Gag stoichiometry would likely compromise viral particle size and morphogenesis, which are critical factors for optimal HIV-1 infectivity. Characterization of a Stau1 mutant that interacts with the NC domain of pr55Gag but is unable to stimulate pr55Gag multimerization may help decipher the mechanism of action and the role of Stau1 during virus assembly.

Recently, two studies reported that the deoxycytidine deaminase APOBEC3G (A3G) (member of the apolipoprotein B mRNA-editing catalytic polypeptide 3 family) and Stau1 are components of the same ribonucleoprotein complexes (10, 28). In the absence of HIV-1 viral infectivity factor (Vif), A3G interacts with the NC domain of pr55Gag, is encapsidated into nascent viral particles, and hypermutates reverse transcripts following viral entry (12, 45). This antiviral activity is neutralized by the presence of Vif, which causes A3G proteasomal degradation and exclusion from viral particles (12). Considering that Stau1 and A3G are found in the same complexes (10, 28) and that both interact with pr55Gag (9, 12), it will be interesting to investigate if A3G modulates Stau1 action during pr55Gag multimerization in the presence or the absence of Vif. Moreover, since Stau1 is associated with HIV-1 genomic RNA in the viral particle (36) and hence probably during the reverse transcription events, it is possible that Stau1 influences not only A3G encapsidation into HIV-1 but also its antiviral activity during the early steps of the HIV-1 life cycle.

This study highlights the use of BRET as a powerful, fast, and technically convenient tool to detect subtle changes in pr55Gag multimerization in live cells. Furthermore, the results shown here unveil a new role for Stau1 in HIV assembly, and to our knowledge, Stau1 is the first identified host factor shown to modulate pr55Gag multimerization in live cells. Stau1 could potentially represent a suitable target for disrupting assembly processes.


   ACKNOWLEDGMENTS
 
We thank Louise Cournoyer, Kimberly Hu, and Alexandre Desjardins for technical assistance; Karine Boulay for preparing cross-linked antibodies; Anne Gatignol, George Pavlakis, and Éric Cohen for constructs; Lou Henderson, Robert Gorelick, Michel Bouvier, Gerardo Ferbeyre, and M. Reichlin for antibodies; Eric Freed for comments on an early version of the manuscript; and Karine Boulay and Frédérick-Antoine Mallette for critical reading of the manuscript.

A.J.M. is supported by a Canadian Institutes of Health Research (CIHR) New Investigator award. This work was supported by a grant from the Natural Sciences and Engineering Research Council of Canada to L.D. and grants from the CIHR to L.D. and A.J.M.


   FOOTNOTES
 
* Corresponding author. Mailing address for Luc DesGroseillers: Département de biochimie, Université de Montréal, P.O. Box 6128, Centre Ville, Montréal, QC, Canada H3C 3J7. Phone: (514) 343-5802. Fax: (514) 343-2210. E-mail: luc.desgroseillers{at}umontreal.ca. Mailing address for Andrew J. Mouland: Lady Davis Institute for Medical Research, 3755 Côte-Ste-Catherine Road, Montréal, QC, Canada H3T 1E2. Phone: (514) 340-8260. Fax: (514) 340-7502. E-mail: andrew.mouland{at}mcgill.ca Back

{triangledown} Published ahead of print on 11 April 2007. Back


   REFERENCES
Top
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Accola, M. A., S. Hoglund, and H. G. Gottlinger. 1998. A putative alpha-helical structure which overlaps the capsid-p2 boundary in the human immunodeficiency virus type 1 Gag precursor is crucial for viral particle assembly. J. Virol. 72:2072-2078.[Abstract/Free Full Text]
  2. Accola, M. A., B. Strack, and H. G. Gottlinger. 2000. Efficient particle production by minimal Gag constructs which retain the carboxy-terminal domain of human immunodeficiency virus type 1 capsid-p2 and a late assembly domain. J. Virol. 74:5395-5402.[Abstract/Free Full Text]
  3. Alroy, I., S. Tuvia, T. Greener, D. Gordon, H. M. Barr, D. Taglicht, R. Mandil-Levin, D. Ben-Avraham, D. Konforty, A. Nir, O. Levius, V. Bicoviski, M. Dori, S. Cohen, L. Yaar, O. Erez, O. Propheta-Meiran, M. Koskas, E. Caspi-Bachar, I. Alchanati, A. Sela-Brown, H. Moskowitz, U. Tessmer, U. Schubert, and Y. Reiss. 2005. The trans-Golgi network-associated human ubiquitin-protein ligase POSH is essential for HIV type 1 production. Proc. Natl. Acad. Sci. USA 102:1478-1483.[Abstract/Free Full Text]
  4. Babst, M., G. Odorizzi, E. J. Estepa, and S. D. Emr. 2000. Mammalian tumor susceptibility gene 101 (TSG101) and the yeast homologue, Vps23p, both function in late endosomal trafficking. Traffic 1:248-258.[CrossRef][Medline]
  5. Batonick, M., M. Favre, M. Boge, P. Spearman, S. Honing, and M. Thali. 2005. Interaction of HIV-1 Gag