Modulation of {beta}-Catenin and E-Cadherin Interaction by Vpu Increases Human Immunodeficiency Virus Type 1 Particle Release

Vpu (viral protein U) is a 17-kDa human immunodeficiency virustype 1 (HIV-1) accessory protein that enhances the release ofparticles from the surfaces of infected cells. Vpu recruitsβ-transducin repeat-containing protein (β-TrCP) andmediates proteasomal degradation of CD4. By sequestering β-TrCPaway from other cellular substrates, Vpu leads to the stabilizationof β-TrCP substrates such as β-catenin, I ${kappa}$ B ${alpha}$ , ATF4,and Cdc25A, but not of other substrates such as Emi1. This studyshows that in addition to stabilizing β-catenin, Vpu leadsto the depression of both total and β-catenin-associatedE-cadherin levels through β-TrCP-dependent stabilizationof the transcriptional repressor Snail. We showed that bothdownregulation of overall E-cadherin levels and dissociationof E-cadherin from β-catenin result in enhanced viral release.By contrast, the overexpression of E-cadherin or the preventionof the dissociation of E-cadherin from β-catenin resultsin depressed levels of virus release. Since E-cadherin is expressedonly in dendritic cells and macrophages, and not in T cells,our data suggest that the HIV-1 vpu gene may have evolved tocounteract different restrictions to assembly in different cells.

INTRODUCTION

Top

INTRODUCTION

Vpu (viral protein U) is an 81-amino-acid type I integral membraneprotein expressed by human deficiency virus type 1 (HIV-1) andseveral simian immunodeficiency viruses (SIVs), including SIVcpz,SIVgsn, SIVmus, and SIVmon, but not expressed by HIV-2 or otherSIVs. Vpu is not incorporated into HIV-1 virions (13-16, 22,55). Although dispensable for HIV replication in vitro, theVpu open reading frame is conserved in infected patients, andit is required for efficient replication of chimeric simian-humanimmunodeficiency viruses in macaques, indicating that it hasa critical role in pathogenesis (54, 66).

One of the best-defined functions of Vpu is the enhancementof viral particle release from the surfaces of infected cellsin a cell-type-dependent manner that occurs only in human T-lymphocyte,macrophage, HeLa, and 293T cells (17, 21, 24, 47). In Africangreen monkey kidney COS cells, viral budding occurs independentlyof Vpu (45, 59). The primary domain of Vpu associated with thisfunction is the N-terminal 27-amino-acid transmembrane regionthat inserts into intracellular and plasma membranes (4, 43,60). Randomization of the amino acid sequence of the N-terminaldomain of Vpu diminishes the ability of Vpu to enhance virusrelease (46). While the N-terminal domain is the principal regionresponsible for enhanced particle release, viruses expressingonly the N-terminal region of Vpu are unable to achieve wild-typelevels of particle release, indicating that other Vpu domainsmay also contribute to this activity (35, 47, 49). The mechanismof action through which Vpu mediates particle release remainsunclear and is suspected to be due to the ability of Vpu tocounteract a host cell factor specific to human cells (59).It may also depend on the ability of Vpu to interact with thepotassium ion channel protein TASK (30). In addition, recentstudies have shown that Vpu redirects nascent viral particlesfrom internal membrane vesicles to the plasma membrane (27,38).

The other well-defined function of Vpu is the degradation ofCD4 complexed with gp160 in the endoplasmic reticulum, whichis mediated by the C-terminal 54-amino-acid region of the protein(35, 49). This function is carried out by the binding and recruitmentof the β-transducin repeat-containing protein (β-TrCP),the receptor component of the multisubunit Skp1-cullin-F boxprotein complex E3 ubiquitin ligase (7, 36). β-TrCP bindsto Vpu after casein kinase II-mediated phosphorylation of Vpuresidues 52 and 56 (48). Vpu has been shown to stabilize severalβ-TrCP substrates, including the adherens junction proteinβ-catenin, NF- ${kappa}$ B inhibitor I ${kappa}$ B ${alpha}$ (nuclear factor ${kappa}$ B), ATF4 (activatingtranscription factor 4), and cdc25A (cell division cycle protein25A) (5, 20). However, other β-TrCP substrates, such asEmi-1 (early mitotic inhibitor-1), remain unaffected by Vpu(20). Of note, Vpu binding to β-TrCP does not promote itsown degradation, and in fact, Vpu is degraded through a β-TrCP-independentmechanism that depends on the phosphorylation of residue 61(20).

β-catenin has two distinct functions. It binds the transcriptionfactors LEF (lymphoid enhancer factor) and TCF (T-cell factor)to enhance the expression of genes, including those involvedin T-cell activation (42). β-catenin also links the celladhesion molecule E-cadherin to the underlying microfilamentnetwork of the cytoskeleton, most prominently at cell confluence(18, 40). E-cadherin expression is regulated at the transcriptionallevel by the β-TrCP substrate Snail (9, 64). Since ourprevious work showed that Vpu-enhanced particle release dependson cell confluence (17), we examined whether Vpu-mediated virusrelease results from the modulation of β-catenin and E-cadherinlevels and their interaction.

MATERIALS AND METHODS

Top

MATERIALS AND METHODS

Cells, transfection, and antibodies. HeLa and 293T cells were maintained in Dulbecco's modified Eagle'smedium supplemented with 10% fetal calf serum, 2 mM L-glutamine,1 mM sodium pyruvate, and antibiotics (penicillin-streptomycin).Jurkat cells were cultured in RPMI supplemented with 10% fetalcalf serum, 2 mM L-glutamine, and antibiotics (penicillin-streptomycin).Plasmid transfections were carried out using TransIT-LT1 (Mirus),and silent interfering RNA (siRNA) transfections were carriedout using GeneEraser (Stratagene) according to the manufacturer'sinstructions. For the single TransIT transfection experiments,the cells were transfected with 2 µg of plasmid. For thecotransfection experiments, the cells were transfected with0.5 µg of a proviral clone and 1.5 µg of a secondplasmid. The antibodies used were β-catenin (BD Transduction),E-cadherin (BD Transduction), phospho-β-catenin Ser33/Ser37/Thr41(Cell Signaling Technology), phospho-β-catenin Thr41/Ser45(Cell Signaling Technology), actin (Santa Cruz), HA (Sigma),FLAG (Sigma), SNA-I (Santa Cruz), Vpu (34), and HIV-1 MA (52).Jurkat cells were electroporated at 250 V, 950 µF, R10,using a BTX 600 electroporator.

Plasmids and siRNA. Vphu, NLUdel, NLU35, and NLRd plasmids were provided by K. Strebeland S. Bour (39, 46). Wild-type β-catenin and S33Y plasmidswere donated by E. R. Fearon (8). An E-cadherin expression plasmidwas obtained from S. M. Troyanovsky (57). Vpu-HA was derivedby attaching a single HA epitope tag to the C-terminal end ofHIV-1-derived Vpu in pcDNA3.1, using PCR primers 5'-ATGCAACCTATAATAGTAGCAATAG-3'(forward) and 5'-GCGAGATCTCTAAGCGTAATCTGGTACGTCGTATGGGTAGATATCCAGATCATCAATATCCC-3'(reverse). The Vpsa plasmid was derived from Vphu by mutatingthe serines at 52 and 56 to alanine, using a QuikChange site-directedmutagenesis kit (Stratagene). β-catenin siRNA was designedas described previously (61). A QuikChange site-directed mutagenesiskit was used to generate plasmids Y654E and Y654F from the wild-typeβ-catenin plasmid, using primers 5'-GAAGGTGTGGCGACAGAAGCAGCTGCTGTTTTGTTCCG-3'(forward), 5'-CGGAACAAAACAGCAGCTGCTTCTGTCGCCACACCTTC-3' (reverse),5'-GAAGGTGTGGCGACATTTGCAGCTGCTGTTTTGTTCCG-3' (forward), and5'-CGGAACAAAACAGCAGCTGCAAATGTCGCCACACCTTC-3' (reverse), respectively.The siRNA-resistant plasmids bcatr, Y654Er, and Y654Fr weregenerated using primers 5'-CAAGAACAAGTAGCTGACATAGATGGACAGTATGCAATGACTC-3'(forward) and 5'-GAGTCATTGCATACTGTCCATCTATGTCAGCTACTTGTTCTTG-3'(reverse). E-cadherin siRNA was purchased from Dharmacon (On-Targetplus smart pool). All plasmids were confirmed by nucleotidesequence analyses.

Western blotting and immunoprecipitation. For immunoblotting experiments, cells were seeded at a densityof 0.5 x 10⁶ per well in 6-well plates and transfected withan expression plasmid, using TransIT-LT-1. After being transfectedand incubated, cells were dispersed in lysis buffer (50 mM HEPES[pH 7.5], 150 mM NaCl, 1% [vol/vol] Triton X-100, 10% glycerol,1.5 mM MgCl₂, 1 mM EDTA, 100 mM NaF, 1 µg/ml aprotinin,10 µg/ml pepstatin, 10 µg/ml leupeptin, 1 mM sodiumorthovanadate). The nuclei were pelleted, and the lysates wereresuspended in 2x sample buffer (125 mM Tris [pH 6.8], 10% sodiumdodecyl sulfate [SDS], 50% glycerol, 5% β-mercaptoethanol,and bromophenol blue), boiled for 5 min, and analyzed by SDS-polyacrylamidegel electrophoresis (PAGE) and immunoblotting. For immunoprecipitation,cell lysates prepared in lysis buffer were incubated on a rotatorwith mouse anti-β-catenin immunoglobulin G and proteinG-agarose beads (Sigma) for 1 h at 4°C. Immune complexeswere collected and washed three times by centrifugation withlysis buffer. The immunoprecipitated proteins were analyzedby SDS-PAGE and immunoblotting. Imaging and densitometry analyseswere performed using Alpha Innotech's AlphaImager and the AlphaEaseFCversion 4.0 program.

Virus release assays. The levels of virus released were determined by centrifugatingthe cell cultured, conditioned media through a 20% sucrose cushionat 45,000 rpm for 45 min in an SW55Ti rotor (Beckman), resuspendingthe pellets in sample buffer. The samples were analyzed by SDS-PAGEand immunoblotting, using anti-MA antibody.

RESULTS

Top

RESULTS

Vpu expression results in β-catenin stabilization and alters HIV-1 release. It is now well established that Vpu mediates stabilization ofβ-catenin in HeLa and 293T cells (3, 20). In this study,we examined the effects of Vpu on β-catenin in Jurkat 25T cells electroporated with an HA-tagged Vpu expression plasmid(Fig. 1A). The level of β-catenin, as measured by immunoblottingand densitometric band analysis, was elevated by 2.3-fold (standarderror, ±0.3; P < 0.01, as determined by 2-tailed ttesting of the paired samples) in both Jurkat and HeLa cellsin the presence of Vpu compared to that in the absence of Vpu.

View larger version (19K):
[in this window]
[in a new window]

FIG. 1. Vpu expression results in β-catenin stabilization. (A) Jurkat cells were electroporated with 10 µg of a Vpu-HA expression plasmid or an empty pcDNA vector (mock). At 24 h posttransfection, the cells were harvested, subjected to electrophoresis, and immunoblotted with anti-β-catenin (upper panel), anti-actin (middle panel), or anti-HA antibodies (lower panel). The experiment was repeated twice in Jurkat cells and once in 293T cells, with similar results (data not shown). (B) 293T cells were transfected with pcDNA or an optimized Vpu expression plasmid (vphu), and after 24 h, immunoblotting was performed with an anti-phospho-β-catenin ( ${alpha}$ -pβ-catenin) antibody that recognizes the phosphorylation of residues 41 and 45 (top panel). Vpu expression was measured by immunoblotting with an antibody against Vpu (lower panel). The experiment was performed twice with similar results. (C) 293T cells were transfected with NL4-3 or NLUdel, and immunoblotting was performed as for panel B. ${alpha}$ , anti.

In the absence of Wnt signaling, β-catenin is constitutivelyphosphorylated at serines 33, 37, 45 and at threonine 41 byglycogen synthase kinase 3β, resulting in β-TrCP-mediatedubiquitination and subsequent proteasomal degradation (1, 41,44, 56). Immunoblot analysis was carried out on lysates of 293Tcells transfected with a codon-optimized Vpu construct (Vphu)with NL4-3, an infectious molecular HIV-1 clone which expressesfull-length Vpu, or with NLUdel, a proviral clone with a Vpudeletion in the NL4-3 background (39, 46). The levels of phospho-β-cateninwere measured with an antibody that specifically recognizesphosphorylation at the Ser45 and Thr41 sites. The levels ofphospho-β-catenin were elevated in the presence, but notin the absence, of Vpu (Fig. 1B and C). Similar results werefound in Jurkat 25 cells, with an antibody to β-cateninphosphorylated at sites 33, 37, and 41 (data not shown). Sinceβ-TrCP mediates specific degradation of phosphorylatedβ-catenin (28), this result is consistent with the decreasedturnover of β-catenin. Furthermore, this result is in linewith the previous demonstration that Vpu sequesters β-TrCPaway from its natural substrates (3, 5). Thus, similar effectsof Vpu on β-catenin stability were seen in three differentcell lines, with antibodies to all four N-terminal phosphorylationsites of β-catenin, with three different Vpu expressionplasmids, with or without epitope tags, and with or withoutthe expression of other HIV-1 proteins.

To examine the effects of β-catenin on virus release, NLUdelwas cotransfected into HeLa cells with expression plasmids forFLAG-tagged β-catenin or a stable β-catenin mutant,S33Y, which is resistant to degradation (Fig. 2) (44). As acontrol, Vphu was also cotransfected with NLUdel (39). At 48h posttransfection, viral protein expression was measured byimmunoblotting cell lysates with antibodies to Gag MA and Vpu.The levels of virus release were assessed by immunoblottingculture supernatants with the antibody to MA. Vpu expressionresulted in a 2.1-fold-enhanced level of virus release (standarderror, ±0.3; P < 0.05) (Fig. 2, lanes 3 and 4). Thisresult is in agreement with the previously reported effectsof Vpu on the levels of particle release, as several studieshave found Vpu effects to be variable depending on the celltype, with as low as a twofold-enhanced level of virus release(2, 47, 49). The expression of exogenous wild-type β-cateninhad no significant effect on virus release (1.06-fold ±0.3) (Fig. 2, lanes 2 and 4). However, the level of exogenouslyexpressed wild-type β-catenin was found not to be significantlyhigher than endogenous levels. In contrast, expression of thestable form of β-catenin, the S33Y β-catenin mutant,reduced virus release by 3.6-fold (±0.02; P < 0.01)(Fig. 2, lanes 1 and 4). Note that overall β-catenin levelswere found to be increased by 1.5-fold in S33Y-transfected cellscompared to that in mock-transfected cells. The levels of Gagin the cell lysates were not affected (0.97-fold ± 0.1).Similar results were seen when NL4-3 was cotransfected withwild-type β-catenin and with S33Y, indicating that theeffect of stabilized β-catenin is dominant over that ofVpu (data not shown). In addition, an enzyme-linked immunosorbentassay was used to determine p24 antigen levels and was foundto give similar results to that of the immunoblot analysis (datanot shown).

View larger version (16K):
[in this window]
[in a new window]

FIG. 2. HIV-1 budding is regulated by β-catenin. NLUdel was cotransfected into HeLa cells with empty pcDNA vector (mock), with optimized Vpu (vphu), or with a wild-type or S33Y FLAG-tagged β-catenin expression plasmid (wtbcat and S33Y, respectively). After 24 h, equivalent amounts of the supernatants (sups) and cell extracts (cells) were analyzed by immunoblotting, using an anti-MA polyclonal antibody ( ${alpha}$ -MA; top two panels) or an anti-Vpu polyclonal antibody ( ${alpha}$ -Vpu; panel 3) or anti-FLAG ( ${alpha}$ -FLAG; panel 4) monoclonal antibody. The experiment was repeated three times with similar results.

Thus, stabilized β-catenin leads to decreased virus particlerelease. However, Vpu stabilizes β-catenin and increasesvirus particle release. To explain this apparent conflict, theeffects of Vpu on β-catenin-interactive proteins were alsoexamined.

Vpu reduces the β-catenin interaction with E-cadherin. HeLa cells were transfected with NL4-3 and NLUdel (Fig. 3).At 24 h posttransfection, immunoblotting for β-catenin,E-cadherin, MA, and actin was performed on the total cell lysates(Fig. 3A, left panel). As expected, in the presence of NL4-3,we found increased levels of β-catenin. Furthermore, weobserved that in the presence of NL4-3, E-cadherin levels weredecreased by 3.7-fold (±0.7). Immunoprecipitation witha β-catenin antibody was carried out on cell lysates todetermine the specific effects of Vpu on the β-catenininteraction with E-cadherin (Fig. 3A, right panel). In the presenceof NL4-3, the levels of E-cadherin associated with β-cateninwere also decreased by 3.6-fold (±1.03). The levels of ${alpha}$ -catenin were not affected, indicating that there is a specificdissociation of β-catenin from E-cadherin and not an overalldisruption of the adherens complex.

View larger version (22K):
[in this window]
[in a new window]

FIG. 3. Expression of the Vpu C-terminal domain results in reduced association of E-cadherin with β-catenin. (A) HeLa cells were transfected with NL4-3, NLUdel, NLRd, or an empty vector (mock). The left panel shows immunoblots of the cell extracts, and the right panel shows cell extracts immunoprecipitated with an anti-β-catenin antibody prior to immunoblotting. The experiment was repeated four times. (B) HeLa cells were transfected with Vphu (vphu), mutant vpsa, with Ser residues 52 and 56 mutated to alanine, or an empty vector. The left and right panels show immunoblots of the cell extracts and anti-β-catenin antibody immunoprecipitates, respectively. The experiment was repeated three times. ${alpha}$ , anti; IP, immunoprecipitates.

Transfection of NLRd, which expresses a mutant Vpu with a randomizedtransmembrane domain, but which contains an intact C-terminaldomain (46), also resulted in a decrease in E-cadherin levelsby 5.1-fold (±1.5) (Fig. 3A). In contrast, cells transfectedwith NLU35, which expresses a mutant Vpu truncated after residue35, exhibited no effects on E-cadherin or β-catenin-associatedE-cadherin levels (data not shown) (46).

We further confirmed these results by transfecting cells withthe Vphu expression plasmid and found that both total and β-catenin-associatedE-cadherin levels were reduced, by 2.7-fold (±0.8) and4.0-fold (±0.8), respectively (Fig. 3B). The cytoplasmicDSGXXS phosphorylation motif of Vpu has been shown to mediateinteractions with β-TrCP (36, 46). We found that cellstransfected with Vpsa, a plasmid expressing codon-optimizedVpu with the serines of the phosphorylation motif mutated toalanines, demonstrated no alteration of total or β-catenin-associatedE-cadherin levels. This indicates that the C-terminal domainof Vpu, and specifically the β-TrCP-interacting motif ofVpu, is responsible for reduced E-cadherin levels.

Snail is a zinc finger transcriptional factor known to be regulatedby β-TrCP and to repress E-cadherin expression (9, 64).We examined Snail levels in Vpu-transfected HeLa cells (Fig.4A, 4B). Snail levels were found to be upregulated in both Vphu-and NL4-3-transfected cells. NLUdel-transfected cells did notshow an upregulation of Snail levels, indicating that otherHIV-1 proteins are not involved in this process. We also performedreal time reverse transcription-PCR analysis and found a 5.7-folddecrease in E-cadherin mRNA levels in Vphu-transfected HeLacells compared to that in mock-transfected cells, confirmingthat the depression of E-cadherin levels was due to Snail-mediatedtranscriptional repression.

View larger version (14K):
[in this window]
[in a new window]

FIG. 4. Vpu expression results in increased Snail levels. (A) HeLa cells were transfected with either an empty vector (mock) or Vphu (vphu). At 24 h posttransfection, the cell lysates were analyzed by immunoblotting. (B) HeLa cells were transfected with either an empty vector, NLUdel, or NL4-3. At 24 h posttransfection, the cell lysates were analyzed by immunoblotting. ${alpha}$ , anti.

The dissociation of β-catenin and E-cadherin complexes results in enhanced particle release. In order to examine the functional consequences of the downregulationof E-cadherin, we first determined the effects of downregulatingthe overall levels of E-cadherin on virus particle release.HeLa cells were transfected with siRNA against E-cadherin orwith luciferase siRNA as a control (Fig. 5A). At 24 h posttransfectionwith siRNA, the cells were transfected with NLUdel, and afteran additional 24 h, the cell lysates and supernatants were harvestedand subjected to immunoblot analysis. In the presence of siRNAagainst E-cadherin, the level of E-cadherin was found to bedownregulated by 4-fold and virus release upregulated by 2.1-fold,as determined by densitometric analysis of the bands (Fig. 5A).The levels of β-catenin and actin were not affected. Thus,a specific reduction in the total levels of E-cadherin leadsto an increase in virus particle release.

View larger version (13K):
[in this window]
[in a new window]

FIG. 5. Effect of E-cadherin on HIV-1 particle release. (A) HeLa cells were transfected with 20 nM of siRNA against E-cadherin or with luciferase siRNA as a control. At 24 h posttransfection, the cells were transfected with NLUdel. After an additional 24 h, the supernatants (sups) and cells were collected and analyzed by immunoblotting with an antibody to MA (first and fifth panels), E-cadherin (second panel), β-catenin (third panel), or actin (fourth panel). (B) 293T cells were cotransfected with a proviral clone, either NL4-3 or NLUdel, and with an empty vector (mock) or E-cadherin expression plasmid. At 48 h posttransfection, the cell lysates and supernatants were harvested and analyzed by immunoblotting. ${alpha}$ , anti.

When NL4-3 was transfected into cells expressing reduced levelsof E-cadherin, the effect on particle release was less markedthan that seen with NLUdel (data not shown). Thus, Vpu-mediateddownregulation of E-cadherin is sufficient to increase particlerelease, and further downregulation of E-cadherin levels withsiRNA did not result in additive effects on virus release.

We also determined the effect of the overexpression of E-cadherinon virus release. Vpu was found to have effects on E-cadherindownregulation in 293T cells similar to those described abovefor HeLa cells (Fig. 5B). 293T cells were cotransfected witha proviral vector, either NL4-3 or NLUdel, and either an emptycytomegalovirus promoter expression vector or a cytomegalovirus-E-cadherinexpression vector (53). At 48 h posttransfection, the cellsand supernatants were analyzed for Gag expression. In the presenceof overexpressed E-cadherin, reduced levels of the virus werereleased into the supernatant from both NL4-3- and NLUdel-transfectedcells (Fig. 5B).

Vpu has been shown to overcome a dominant restriction to assemblyin human cells (59). COS cells do not have this restrictionto assembly. Varthakavi and colleagues showed that while human-Africangreen monkey cell heterokaryons show reduced levels of Gag release,the expression of Vpu in these heterokaryons can overcome thedefect (59). We found that in COS cells, E-cadherin could notbe detected with the human ${alpha}$ -E-cadherin antibody. Therefore,we tested whether E-cadherin could be the dominant restrictionfactor to virus assembly by overexpressing human E-cadherinwith or without Vpu in COS cells. We found that overexpressionof human E-cadherin in COS cells did not have an effect on eitherNL4-3 or NLUdel release, indicating that E-cadherin is not therestriction factor that is overcome by Vpu (data not shown).

In order to specifically examine the functional effect on particlerelease of the β-catenin interaction with E-cadherin, wefirst reduced endogenous β-catenin levels in HeLa cellsby using siRNA. The level of β-catenin was found to bereduced by 2.8-fold with siRNA against β-catenin comparedto the control siRNA against luciferase (±0.5) (Fig.6, lanes 4 and 5). At 24 h posttransfection with siRNA, thecells were transfected with NLUdel and an empty vector. Thisresulted in an 1.8-fold (±0.4; P < 0.03)-increasedlevel of virus release, as measured by densitometry after anadditional 24 h, compared to that for cells transfected withcontrol siRNA against luciferase (Fig. 6, lanes 4 and 5). Plasmidswere cotransfected with NLUdel into siRNA-expressing HeLa cellsto express siRNA-resistant forms of β-catenin or β-cateninmutated at residue Y654. Y654Er is an siRNA-resistant β-cateninmutant unable to associate with E-cadherin (Fig. 6, lane 2),whereas Y654Fr is an siRNA-resistant β-catenin that remainsconstitutively associated with E-cadherin (Fig. 6, lane 1) (31).Y654Er had little effect on virus release compared to that inpcDNA-transfected cells (1.2-fold) (Fig. 6, lanes 2 and 4),whereas Y654Fr resulted in a 3.8-fold decrease in virus release(±0.7; P < 0.03) (Fig. 6, lanes 1 and 4). Thus, whenβ-catenin could not dissociate from E-cadherin, we founda decrease in the levels of virus release. However, in the presenceof a β-catenin mutant that cannot associate with E-cadherin,the levels of virus release were comparable to that seen withsiRNA against β-catenin; i.e., we found an increase invirus release compared to that for cells expressing controlsiRNA.

View larger version (15K):
[in this window]
[in a new window]

FIG. 6. Dissociation of E-cadherin and β-catenin leads to increased HIV-1 particle release. HeLa cells were transfected with 50 nM of siRNA against β-catenin or luciferase. At 24 h posttransfection, the cells were transfected with NLUdel and either pcDNA or an expression plasmid for an siRNA-resistant form of wild-type β-catenin (bcatr) or mutant β-catenin (Y654Er or Y654Fr). After an additional 24 h, the supernatants (sups) and cells were collected and analyzed by immunoblotting with an anti-MA (top two panels) or anti-β-catenin antibody (bottom panel). The experiment was repeated three times. ${alpha}$ , anti.

DISCUSSION

Top

DISCUSSION

There appear to be two distinct effects of Vpu on β-cateninand E-cadherin. While the C-terminal region of Vpu stabilizesβ-catenin, this domain also results in downregulation ofE-cadherin. Our current work on β-catenin stabilizationextends previous work, demonstrating similar effects in T-cells(3). In addition, our work highlights the functional consequencesof β-catenin stabilization by showing that in the presenceof stabilized β-catenin, there is a reduced level of virusparticle release.

We observed that HeLa cells transfected with NL4-3 show a sevenfoldincrease in the level of virus release compared to that forcells transfected with NLUdel as measured by enzyme-linked immunosorbentassay (data not shown). NLRd-transfected cells have levels ofvirus release comparable to those of NLUdel. However, NLU35-transfectedcells show a 30% reduction in the levels of virus release comparedto that for the wild type, indicating that the C-terminal domainalso plays a role in virus release, albeit a smaller one. Thisfinding is consistent with previous reports that show similarresults with Vpu C-terminal mutants (46, 47).

One point of interest is that we found increased levels of N-terminallyphosphorylated β-catenin in the presence of Vpu, whichis consistent with a reduced turnover of β-catenin. Previousstudies have shown that β-catenin-mediated coactivationof LEF/TCF transcription factors cannot be mediated by β-cateninphosphorylated at the N terminus (44, 53). However, the Besnard-Guerinstudy showed that Vpu-mediated stabilization of β-cateninresults in enhanced TCF activity (3). Therefore, further investigationis required to determine how phosphorylated β-catenin generatedby Vpu can mediate this function. It may be explained by differencesin the cell types used in each study. In addition, our studyshows that cotransfection of a proviral clone with the stabilizedS33Y mutant of β-catenin leads to depressed levels of particlerelease. One hypothesis to explain this result is that theremay be greater numbers of complexes of E-cadherin associatedwith β-catenin in the presence of S33Y. This is consistentwith the finding of Sadot and coworkers, who showed that phosphorylatedβ-catenin is present in newly formed β-catenin/E-cadherincomplexes (44).

We also showed that Vpu mediates the downregulation of boththe total levels of E-cadherin and the levels of E-cadherinassociated with β-catenin. This effect is caused by Vpu-mediatedsequestration of β-TrCP, resulting in upregulation of thetranscriptional repressor Snail. We have previously shown thatthe effects of Vpu on particle release depend on cell confluence(17). Since complexes of β-catenin associated with E-cadherinare more apparent at higher levels of cell confluence, we determinedthe effect of Vpu on the β-catenin interaction with E-cadherinat different cell densities. We found that while Vpu-mediateddissociation of E-cadherin from β-catenin was not apparentat a 25% confluence, there was dissociation at a 75% confluencein Vpu-expressing 293T cells (data not shown). Thus, the effectof Vpu appears to be larger at higher levels of cell confluence,where there are a larger number of mature E-cadherin/β-catenincomplexes.

Cell-to-cell transmission of HIV requires a functional hostcell cytoskeleton and the polarization of both host and viralcomponents with the cytoskeletal machinery (32). The interactionof viral proteins with cell junction components has been shownto facilitate the transmission of herpes simplex virus betweenadjacent cells, and disruption of the adherens junctions hasbeen shown to liberate the receptor for various viruses, includingherpes simplex virus and pseudorabies virus, and contributeto Epstein-Barr virus (EBV) and hepatitis B virus tumorigenesis(10, 19, 29, 65). E-cadherin is anchored to the actin cytoskeletonvia the complex it forms with β-catenin (63). Since thedisruption of β-catenin and E-cadherin complexes resultsin enhanced particle release, we hypothesize that disruptionof the underlying cytoskeletal ${alpha}$ -catenin-actin complex is responsiblefor the β-TrCP-dependent effects of Vpu in enhancing virusrelease (12). Furthermore, both the gp41 transmembrane envelopedomain and HIV-1 coreceptors CCR5 and CXCR4 interact with ${alpha}$ -catenin,though their relative contributions to virus budding requirefurther investigation (11, 33, 51).

It is interesting that β-catenin is also stabilized byother viral proteins, including EBV latent membrane protein1 (LMP1), LMP2A, and hepatitis B virus X protein (10, 29). Severalviral proteins also downregulate E-cadherin expression, includingEBV LMP1 (58).

Since T cells have little E-cadherin, Vpu can promote assemblyindependently of the β-catenin interaction with E-cadherin(26). However, there are high levels of E-cadherin in macrophagesand dendritic cells, in which these effects of Vpu are likelyto be important for pathogenesis (6, 23, 25, 37, 62). It hasbeen reported that Vpu can mediate up to a 1,000-fold enhancementof particle release levels in macrophages (2, 17). Langerhanscells (LC) express high levels of E-cadherin, which is importantfor mediating the adhesion of LC to keratinocytes. It is hypothesizedthat the downregulation of E-cadherin in LC is essential fortheir mobilization from the epidermis to lymph nodes, wherethey can then interact with T cells (50). Thus, Vpu-mediateddownregulation of E-cadherin may have pathological consequencesvia the mobilization of infected LC in the periphery. Therefore,Vpu may have evolved to overcome different restrictions andcarry out different functions in different cells. It is necessaryto investigate the contributions mediated by the N- and C-terminaldomains of Vpu on E-cadherin levels and virus release in eachof these cell types.

ADDENDUM IN PROOF

Top

ADDENDUM IN PROOF

After the present paper was accepted, Kumar et al. reportedthat β-catenin signaling inhibits HIV replication in peripheralblood mononuclear cells (A. Kumar, A. Zloza, R. T. Moon, J.Watts, A. R. Tenorio, and L. Al-Harthi, J. Virol. 82:2813-2820,2008), and Neil et al. reported that tetherin inhibits retrovirusrelease and is antagonized by Vpu (S. J. Neil, T. Zang, andP. D. Bieniasz, Nature 451:425-430, 2008).

ACKNOWLEDGMENTS

We thank E. R. Fearon, S. M. Troyanovsky, K. Strebel, and S.Bour for generous provisions of plasmids and N. Campbell, D.Rauch, H. W. Virgin, A. Pekosz, W. A. Swat, and G. D. Longmorefor critical reviews of the manuscript.

This study was supported by PHS grant AI24745.

FOOTNOTES

* Corresponding author. Mailing address: Box 8069, 660 S. Euclid, St. Louis, MO 63110. Phone: (314) 362-8836. Fax: (314) 747-2120. E-mail: lratner{at}im.wustl.edu

Published ahead of print on 6 February 2008.

REFERENCES

Top