Human Immunodeficiency Virus Nef Induces Rapid Internalization of the T-Cell Coreceptor CD8{alpha}{beta}

Human immunodeficiency virus (HIV) Nef is a membrane-associatedprotein decreasing surface expression of CD4, CD28, and majorhistocompatibility complex class I on infected cells. We reportthat Nef strongly down-modulates surface expression of the ß-chainof the CD8 ${alpha}$ ß receptor by accelerated endocytosis, whileCD8 ${alpha}$ -chain expression is less affected. By mutational analysisof the cytoplasmic tail of the CD8 ß-chain, an FMKamino acid motif was shown to be critical for Nef-induced endocytosis.Although independent of CD4, endocytosis of the CD8 ß-chainwas abrogated by the same mutations in Nef that affect CD4 down-regulation,suggesting common molecular interactions. The ability to down-regulatethe human CD8 ß-chain was conserved in HIV-1, HIV-2,and simian immunodeficiency virus SIVmac239 Nef and requiredan intact AP-2 complex. The Nef-mediated internalization ofreceptors, such as CD4, major histocompatibility complex classI, CD28, and CD8 ${alpha}$ ß, may contribute to the subversionof the host immune system and progression towards AIDS.

INTRODUCTION

Top

Introduction

The human immunodeficiency virus type 1 (HIV-1) Nef proteinis a 27-kDa protein that is abundantly produced during the earlyphase of viral gene expression (28, 54). Nef is posttranslationallymodified by phosphorylation and due to irreversible attachmentof myristic acid to its N terminus, it is targeted to the cellularmembrane. Nef has multiple distinct functions: it modulatescell surface molecules, such as CD4 (21), CD28 (63), major histocompatibilitycomplex (MHC) class I (59), MHC class II and MHC class II-associatedinvariant chain (Ii, CD74) (62), interferes with signal transductionpathways (reviewed by Tolstrup et al. [65]), T-cell generation(61, 67), and enhances virion infectivity and viral replication(reviewed by Fackler and Baur [17]). The molecular mechanismsof most of these effects and their contribution to pathogenesisare only partially understood.

To modulate cell surface receptor expression, Nef utilizes severalstrategies, linked to distinct regions within the Nef protein.Like many other pathogenic viruses, HIV-1 down-regulates thecell surface expression of MHC class I and circumvents in thisway the attack by cytotoxic T lymphocytes (59). Another profoundlyinvestigated Nef-mediated effect is down-regulation of the CD4receptor (21, 1), due to accelerated endocytosis via clathrin-coatedpits followed by lysosomal degradation (51). In addition, CD4down-regulation by HIV-1 and simian immunodeficiency virus (SIV)Nef proteins also involves intracellular retention mechanisms(55). As Nef has been shown to interact with the CD4 receptoras well as with the adaptor protein (AP) complex, either AP-1(8, 16, 33), AP-2 (16, 22, 26), or AP-3 (33), it may act asa connector between components of the cellular endocytic machineryand the cytoplasmic tail of CD4 (13, 41, 42). A Nef dileucinesequence was found to be required for accelerated internalizationof CD4 and CD28 from the cell surface to endosomes and lysosomes(8, 13, 25), rendering Nef the only nontransmembrane proteinknown to traffic via a dileucine-based motif (35).

The T-cell CD8 coreceptor exists as an ${alpha}$ ${alpha}$ homodimer, found onintestinal T cells, ${gamma}$ ${delta}$ T cells, thymic T-cell precursors, andNK cells, and an ${alpha}$ ß heterodimer, most commonly expressedon thymocytes and on peripheral T cells (19, 31). The surfaceexpression of the CD8 ß-chain is dependent on expressionof the CD8 ${alpha}$ -chain, to which it becomes covalently linked inthe endoplasmic reticulum (24). The cytoplasmic tail of CD8 ${alpha}$ comprises 30 amino acids and contains a motif of two vicinalcysteines for interaction with the Src kinase p56^lck by meansof a zinc chelate complex (68).

Although the tail of CD8ß consists of only 19 residuesand contains no known protein binding motifs, studies in miceindicated a role for CD8ß and its cytoplasmic tailin thymic development and in activation of CD8⁺ T cells (3,4, 5, 32). Pathological conditions in which CD8 ${alpha}$ ⁺ß^lowand CD8 ${alpha}$ ${alpha}$ T-cell receptor ${alpha}$ ß T cells increase in theperiphery include Wiskott-Aldrich syndrome, where peripheralblood CD8⁺ T-cell receptor ${alpha}$ ß T cells mostly expressCD8 ${alpha}$ ${alpha}$ homodimers (34), and HIV infection in which the appearanceof a major CD8 subpopulation with reduced CD8ß chainsmay occur (58).

Here, we report that HIV-1 as well as HIV-2 and SIVmac239 Nefdown-regulate cell surface expression of the human CD8 ${alpha}$ ßreceptor. The CD8 ß-chain cytoplasmic tail containsan FMK amino acid sequence that allows Nef-mediated modulation.Based on our results we suggest Nef is using clathrin-mediatedendocytosis, requiring AP-2, for accelerated down-regulationof CD8 ${alpha}$ ß and CD4. As a subset of CD8⁺ T cells havebeen shown to be infected by HIV (39), down-regulation of CD8 ${alpha}$ ßmight harm CD8 lymphocyte function and contribute in this wayto HIV-mediated subversion of the immune system.

MATERIALS AND METHODS

Top

Materials and Methods

Plasmid constructions. All retroviral nef constructs were made as previously described(61). The deletion mutant Del 3 was provided by J Zack (2),and SIVmac239 and HIV-2 Rod were obtained from J. Skowronskiand R. Benarous, respectively. The CD8 ${alpha}$ (GenBank accession numberNM_001768) gene, kindly provided by S. Bonatti, was subclonedin the LZRS vector, in which the internal ribosome entry site(IRES)-enhanced green fluorescent protein (EGFP) cassette wasdeleted by NotI digestion. The CD8ß (GenBank accessionnumber NM_004931) gene, also provided by S. Bonatti, was taggedwith hemagglutinin (HA) for confocal imaging and was subclonedin the BamHI and EcoRI sites in LZRS-IRES-EGFP or BamHI andNotI sites in LZRS-IRES- ${Delta}$ NGFR. In all experiments, control transductionswere done with the parental LZRS vectors expressing only themarker gene.

The CD8 ${alpha}$ (EC-TM)-CD8ß(IC) chimera was constructed usingthe primer pair CD8 ${alpha}$ S (BamHI): 5'-TATTGGATCCATGGCCTTACC-3';CD8 ${alpha}$ EC+TM AS (SalI): 5'-CTTACGTCGACGTGATAACCAGTGACAG-3' andCD8ß IC S (SalI): 5'-CTTACGTCGACACACCTGTGCTGC-3';CD8ß AS (NotI): 5'-ACTATAGCGGCCGCTTATTTGTAAAATTG-3'.The CD8 ${alpha}$ (EC-TM)-CD8 ${alpha}$ (IC) was constructed as a control, using theprimers CD8 ${alpha}$ IC S (SalI): 5'-CTTACGTCGACACTTTACTGCAAC-3'; CD8 ${alpha}$ AS (NotI): 5'-TTTATAGCGGCCGCTTAGACGTATC-3' for amplificationof the CD8 ${alpha}$ cytoplasmic tail. For both constructs, the two PCRfragments were SalI digested and ligated. In the next step theCD8 ${alpha}$ ß and CD8 ${alpha}$ ${alpha}$ chimers were amplified by PCR, BamHI-NotIdigested, and ligated in the LZRS-IRES- ${Delta}$ NGFR vector.

For confocal microscopy, LZRS-Nef.EGFP-N2 fusion protein constructswere designed. The EGFP-N2 gene was isolated from the pEGFP-N2vector (Clontech, Palo Alto, CA) by BamHI-NotI digestion andcloned into the LZRS vector, in which the IRES-EGFP cassettewas excised by NotI digestion. In a next step the nef genesNA-7 wild-type, PPAA, and LLAA were amplified by PCR, and BamHI-BamHIinserted 5' of the EGFP sequence. The RNAi probes were BglII-HindIIIcloned into the pSUPER vector (9) containing an extra EcoRIrestriction site 3' of the cloning site, further referred toas pSUPER(EcoRI).

To construct the 19-nucleotide hairpin short interfering RNAcassettes, two cDNA oligonucleotides were chemically synthesized(Invitrogen, Merelbeke, Belgium), annealed, and inserted immediatelydownstream of the H1 promoter: 5'-GATCCCC-19-TTCAAGAGA-19-TTTTTGGAAA-3'and 5'-AGCTTTTCCAAAAA-19-TCTCTTGAA-19-GGG-3'. The target sequencesfor each of the genes were as follows: AP-2 µ2 subunit(GenBank accession number NM_004068), 5'-GTGGATGCCTTTCGGGTCA-3'(47), clathrin heavy chain (GenBank accession number NM_004859),5'-TATCTCGCTTGCTCAGCGT-3', and dynamin-2 (GenBank accessionnumber NM_004945), 5'-GACATGATCCTGCAGTTCA-3'. The control shortinterfering RNA was a functional oligonucleotide with p53 asthe target sequence and described elsewhere (9).

Lentiviral packaging plasmid p8.91 and vesicular stomatitisvirus envelope plasmid (pMD.G) were kindly provided by D. Trono(Université de Genève, Geneva, Switzerland); transfervector TRIP ${Delta}$ U3-CMV-EGFP was kindly provided by P. Charneau (HôpitalNecker, Paris, France). A PCR-amplified WPRE cassette (GenBankaccession number J04514, nucleotides 1093 to 1685) (69) wasinserted into plasmid TRIP ${Delta}$ U3-CMV-EGFP at the unique XhoI sitedownstream of the EGFP stop codon, resulting in TRIP ${Delta}$ U3-CMV-GFP-WPRE.Next, the H1 RNA interference (RNAi) cassettes were EcoRI-EcoRItransferred from the pSUPER(EcoRI) vector into the EcoRI-digestedTRIP ${Delta}$ U3-CMV-EGFP-WPRE vector. Sequencing (ABI, Perkin Elmer,Foster City, CA) confirmed the integrity of all constructs.

Production of retroviral and lentiviral supernatants. The Phoenix-Amphotropic packaging cell line was transfectedwith the different retroviral constructs as previously described(61). For lentivirus production, 293T cells were seeded 24 hbefore transfection. Transfection of the three plasmids wasdone using a calcium phosphate transfection kit (Invitrogen)and viral supernatant was harvested 40 h later.

Cell culture and chemical products. All cells were cultured at 37°C in a humidified atmospherecontaining 7.5% (vol/vol) CO₂ in air. Peripheral blood mononuclearcells were isolated by density separation (Lymphoprep, Nyegaard,Oslo, Norway) of buffy coats or whole blood. SupT1 (AIDS Researchand Reference Reagent Program, National Institutes of Health,Bethesda, MD), Daudi, 293T, and other cells were cultured asdescribed previously (61). Freshly isolated peripheral bloodmononuclear cells were stimulated at day 0 with phytohemagglutinin(2 µg/ml) and interleukin-2 (50 IU/ml), and fresh interleukin-2was added every 3 days of culture. 4-Hydroxytamoxifen was purchasedfrom Sigma-Aldrich (Bornem, Belgium) and ikarugamycin was obtainedfrom LKT Laboratories (St. Paul, MN).

Viral gene transfer. For retroviral transduction of cell lines and peripheral bloodmononuclear cells, cells were mixed with viral supernatant whichwas incubated for 10 min with Dotap (Roche Diagnostics, Penzberg,Germany). To increase transduction efficiency, cells were spun(90 min, 2500 rpm, 32°C). Lentiviral transductions werecarried out in the presence of Polybrene (4 µg/ml; Sigma-Aldrich).Transduction efficiency was evaluated by flow cytometry 48 to72 h after transduction and varied between 20 and 50% with retroviralsupernatant and between 50 and 100% with lentiviral supernatant.The degree of Nef-mediated down- or up-regulation of a receptorwas evaluated by measurement of the surface level expressionof the receptor at day 2 after transduction. The fraction ofdown- (or up-) regulation of a receptor was calculated by subtractingthe mean fluorescence intensity (MFI) of cells with a fixedrange of high marker gene expression (EGFP or ${Delta}$ NGFR) from theMFI value of the cells lacking the marker gene and then dividingthis result by the MFI of the cells lacking the marker gene(62).

Monoclonal antibodies, flow cytometry, and cell isolation methods. Mouse anti-human monoclonal antibodies used were CD28 (Leu-28;phycoerythrin; Becton Dickinson Immunocytometry Systems [BDIS],Mountain View, CA), CD4 (Leu-3a phycoerythrin or allophycocyanin;BDIS), CD8 ${alpha}$ (SK1 phycoerythrin or allophycocyanin; BDIS), CD8 ${alpha}$ ß(2ST8.5H7 phycoerythrin; Coulter, Miami, FL), CD74 (LN2; Immucor,Heppignies, Belgium), HLA-DR (L243 allophycocyanin; BDIS), HLA-A,B, and C (G46-2.6 phycoerythrin; BDIS), and nerve growth factorreceptor (NGFR) low-affinity receptor (ME20.4 phycoerythrinor allophycocyanin; Chromaprobe, Maryland Heights, MO).

CD74 was detected through a phycoerythrin-conjugated goat anti-mouseF(ab')₂ antibody (Jackson ImmunoResearch Laboratories, WestGrove, PA). The cells were analyzed on a FACSCalibur flow cytometer(BDIS), as previously described (66). Daudi cells transducedwith CD8ß-IRES- ${Delta}$ NGFR or CD8 ${alpha}$ constructs were purifiedby positive selection using a phycoerythrin-conjugated monoclonalantibody assigned to NGFR and CD8 ${alpha}$ , respectively, followed byanti-phycoerythrin superparamagnetic microbeads (MACS; MiltenyiBiotec, Bergisch Gladbach, Germany). SupT1 cells, transducedwith (mutant) Nef-IRES-EGFP constructs, were sorted using aFACSVantage (BDIS).

Endocytosis assays. The fluorescence-activated cell sorting-based endocytosis assay(10), using a phycoerythrin-conjugated monoclonal antibody toCD8 ${alpha}$ ß, CD4, or CD28, was performed as described previously(50). For the calculation of internalization, a region was seton the EGFP-positive cells. The fraction of receptor internalizedwas calculated by subtracting the MFI of the background signal(the initial time zero acid wash) from all MFI values obtainedand then dividing this result by the MFI of the total boundantibody (41).

Immunoblotting. Transduced SupT1 cells, sorted to homogeneity, were lysed inLaemmli sample buffer and equal amounts of proteins were runon 8 or 12% Novex Tris-glycine-polyacrylamide gels (Invitrogen)as previously described (61). Primary monoclonal antibodiesused included mouse anti-Nef EH-1 (AIDS Research and ReferenceReagent Program) (11), mouse anti-AP50 (Transduction Laboratories,Lexington, KY), mouse anti-clathrin heavy chain (Covance, Berkeley,CA), and rabbit anti-dynamin-2 (Affinity Bioreagents, Golden,CO).

Confocal microscopy. Transduced or transfected 293T cells were cultured on coverslips,fixed in methanol for 15 min at –20°C, and dried atroom temperature. After a 10-min rehydration in phosphate-bufferedsaline, cells were blocked for 30 min at room temperature (0.4%fish skin gelatin [Sigma-Aldrich] in phosphate-buffered saline),followed by incubation for 60 min with a mouse anti-human CD8ßprimary antibody (5F2 [15], Serotec, Oxford, United Kingdom)or mouse anti-HA antibody (HA.11, Covance), both 1:100 dilutedin blocking solution. After washing in phosphate-buffered saline,cells were stained with Alexa-Fluor 594-conjugated goat anti-mouseimmunoglobulin G secondary antibody (Molecular Probes Inc, Eugene,OR) for 1 h (1:100 diluted in blocking solution). After a fourfoldwash step, nuclei were counterstained with 4',6'-diamidino-2-phenylindole(DAPI)/methanol for 5 min, and after a final wash, coverslipswere mounted onto glass slides using Vectashield (Vector LaboratoriesInc, Burlingame, CA). Confocal images were collected with ablue diode Bio-Rad Radiance 2100 confocal laser scanning system(Bio-Rad) and were processed using the Confocal Assistant (CAS)program (Bio-Rad) and Adobe Photoshop (Adobe, San Jose, CA).

RESULTS

Top

Results

Nef induces a decrease in CD8 ${alpha}$ ß surface expression in peripheral blood mononuclear cells and cell lines. A Nef-induced decrease in CD8 ${alpha}$ ß surface levels hasbeen reported by our group in fetal thymic organ culture experiments,using human primary T-cell precursors (61, 67). In order toinvestigate whether this observation was conserved between primarycells, peripheral blood mononuclear cells were retrovirallytransduced with bicistronic constructs expressing HIV-1 Nefand EGFP as a reporter protein. For flow cytometric analysis,cells were stained with a monoclonal antibody directed againstthe CD8 ${alpha}$ ß heterodimer or an anti-CD8 ${alpha}$ , the latter recognizingboth ${alpha}$ ß and ${alpha}$ ${alpha}$ dimers. Surface staining of CD8 ${alpha}$ ßrevealed a dose-dependent decrease in the steady-state expressionlevels of this receptor with increasing Nef expression (allelesNA-7 [Fig. 1A], LAI, and NL4-3 [data not shown]), resultingin a 79% down-regulation in mean fluorescence intensity comparedto nontransduced cells, whereas the decrease in CD8 ${alpha}$ stainingwas less pronounced.

View larger version (37K):
[in this window]
[in a new window]
FIG. 1. Flow cytometric evaluation of Nef-mediated CD8 down-regulation in retrovirally transduced cells. (A) Bivariate dot plots (CD8 ${alpha}$ -allophycocyanin, CD8 ${alpha}$ ß-phycoerythrin versus EGFP) of flow cytometric measurement of Nef^– (control) and Nef⁺ (NA-7 allele) transduced peripheral blood mononuclear cells, gated on CD8⁺ cells, at day 3 after transduction. (B) Bivariate dot plots of flow cytometric measurement (CD8 ${alpha}$ -allophycocyanin, CD8 ${alpha}$ ß-phycoerythrin versus EGFP) of Nef NA-7 wild-type and NA-7 LLAA transduced SupT1 cells (left) and SupT1 cells overexpressing CD8 ${alpha}$ (right), at day 2 after transduction. (C) The solid and open histograms show the CD8 ${alpha}$ ß expression profile of SupT1 CD8 ${alpha}$ ⁺ cells and SupT1 CD8 ${alpha}$ ⁺ cells (CD8 ${alpha}$ -transduced population), respectively, gated as shown in the inset. (D) Daudi CD8 ${alpha}$ ${alpha}$ cells and Daudi CD8 ${alpha}$ ß cells were transduced with control, HIV-1 (NL4-3, LAI and NA-7), SIV (mac239), and HIV-2 (Rod) Nef. Percent down-regulation is shown with white bars for CD8 ${alpha}$ in Daudi CD8 ${alpha}$ ${alpha}$ , with gray bars for CD8 ${alpha}$ in Daudi CD8 ${alpha}$ ß, and with black bars for CD8 ${alpha}$ ß in Daudi CD8 ${alpha}$ ß. All percentages were calculated, as described in Materials and Methods, using the ranges E– and E+, as indicated in A.

We also verified this observation in cell lines. In the CD8 ${alpha}$ ß⁺T-cell line SupT1, expression of Nef evoked a similar decreasein CD8 ${alpha}$ ß surface levels (Fig. 1B, left panels). However,since Nef down-regulates several surface markers on T cells,decreased CD8 ${alpha}$ ß expression could be the consequenceof Nef-mediated internalization of another surface molecule,such as CD4 or CD28, in close proximity to the CD8 receptor.Therefore, we generated Daudi B-cell lines stably expressingthe CD8 ${alpha}$ ${alpha}$ or CD8 ${alpha}$ ß receptor after retroviral gene transferof only CD8 ${alpha}$ or both CD8 chains, further referred to as DaudiCD8 ${alpha}$ ${alpha}$ and CD8 ${alpha}$ ß, respectively.

Figure 1D gives an overview of the ability of different wild-typeHIV-1, HIV-2, and SIVmac239 Nef alleles to down-modulate CD8in Daudi cells. Wild-type Nef-transduced cells are characterizedby a strong decrease in CD8 ${alpha}$ ß surface expression levels,i.e., an 8- to 13-fold decrease in MFI for HIV-1 and SIVmac239.A lower functional activity was found with HIV-2 Rod Nef (atwofold decrease), an observation that was also made for othersurface marker modulations (data not shown and (57)). Theseobservations in the Daudi B-cell line demonstrate that down-regulationis independent of the presence of neighboring T-cell surfacemolecules that are known Nef targets. Nef-mediated down-modulationof the CD8 ${alpha}$ -chain was three- to fourfold lower than that ofthe ß-chain in Daudi CD8 ${alpha}$ ß cells (Fig. 1B).In the Daudi CD8 ${alpha}$ ${alpha}$ cell line, NL4-3- and HIV-2-mediated CD8 ${alpha}$ down-regulationdid not differ significantly from control transduced cells,while LAI, NA-7, and SIVmac239 Nef caused CD8 ${alpha}$ down-regulationto a degree comparable to that observed in the Daudi CD8 ${alpha}$ ßcells (Fig. 1B).

An intriguing finding in Nef-transduced SupT1 cells was that,besides staining with CD8 ${alpha}$ ß monoclonal antibody, CD8 ${alpha}$ monoclonal antibody staining also showed a prominent decreasein surface staining (88% and 71% down-regulation, respectively)(Fig. 1B, left panels). Additional transduction of these cellswith the CD8 ${alpha}$ chain resulted in a distinct population with ahigher CD8 ${alpha}$ staining profile (Fig. 1C). However, only a slightincrease in CD8 ${alpha}$ ß surface staining could be observedin these cells (Fig. 1C and 1B, right panels), suggesting thatthe amount of CD8ß in these cells limits the formationof CD8 ${alpha}$ ß heterodimers, the excess of CD8 ${alpha}$ likely givingrise to CD8 ${alpha}$ ${alpha}$ homodimer expression. Whereas this CD8 ${alpha}$ ^high populationshowed a weakened down-regulatory response (33%), Nef-mediateddown-regulation of the CD8 ${alpha}$ ß receptor remained strong(65%) (Fig. 1B, right panels). In conclusion, these data suggestthat Nef strongly reduces CD8 ${alpha}$ ß surface receptor expression,while it reduces the CD8 ${alpha}$ -chain surface expression only to alower degree.

Concordance between CD8 ${alpha}$ ß down-regulation and Nef-mediated down-/up-regulation of other surface receptors. Retroviral transduction of SupT1 and Daudi cells with wild-typeand mutant Nefs, aligned in Fig. 2A, was performed to assesstheir ability to down-regulate cell surface expression of CD4,MHC class I, CD28, CD8 ${alpha}$ ß (SupT1), and HLA-DR (Daudi)(Fig. 2B) or to up-regulate CD74 (Daudi) (Fig. 2C). As judgedfrom Western blotting (Fig. 2D), all constructs were correctlyexpressed, their expression levels correlating with the meanfluorescence intensity of the sorted SupT1 cells (data not shown).Mutations within Nef that diminished its modulating capacitymore than 50%, compared to the respective wild-type Nef, areunderlined in Fig. 2B and 2C. The results indicate that thesame mutations within Nef that abrogate down-regulation of CD4also abolish down-regulation of CD8 ${alpha}$ ß.

View larger version (40K):
[in this window]
[in a new window]
FIG.2. Flow cytometric mutational analysis of Nef-mediated down- or up-regulation of cell surface molecules in retrovirally transduced SupT1 and Daudi cells. SupT1 cells and Daudi cells were transduced with control virus or wild-type or mutant HIV-1 Nef (alleles NL4-3, LAI, and NA-7). Percent down-regulation was calculated, as described in the Materials and Methods, using the ranges indicated in Fig. 1A. (A) Alignment of the amino acid sequences of the HIV-1 Nef alleles LAI, NL4-3, and NA-7. Underlined amino acids represent the positions that are changed within the indicated mutant Nef protein, and shaded amino acids in NL4-3 and NA-7 are different from the LAI sequence. (B) Comparison of the down-regulating activity of control and Nef constructs in SupT1 cells by using monoclonal antibodies against CD4-allophycocyanin, CD28-phycoerythrin, CD8 ${alpha}$ -allophycocyanin, and CD8 ${alpha}$ ß-phycoerythrin. (C) Comparison of the modulating activity of control and Nef constructs in Daudi cells by using monoclonal antibodies CD74/goat anti-mouse-phycoerythrin and HLA-DR-allophycocyanin. The values in B and C are means and standard deviations calculated from the data generated from three independent experiments. Mutations that abrogate the modulating capacity of Nef for more than 50%, compared to the respective wild-type Nef, are underlined. (D) Western blot analysis of Nef expression in transduced SupT1 cells. Lysates of sorted SupT1 cells, transduced with virus as indicated, were stained for Nef. The marker indicates the position of 31 kDa.

Structurally, these can be divided into three groups: the N-terminalanchor region (G2A and Del3), the W57L58 sequence (WLAA), knownas the binding site for the CD4 receptor (42), and the flexibleloop harboring a dileucine (LLAA and LLGG) and E/D174D175 (EDAAand DDGA) sequence. With regard to Nef-mediated CD28 down-regulation,we found one more mutant (E4A) capable of disturbing the wild-typeNef effect. Of note, up-regulation of CD74 in Daudi cells canbe prevented by the same mutations, with the exception of Del3,as those affecting down-regulation of CD4/CD8 ${alpha}$ ß. HLA-DRdown-regulation was a rather weak effect, only abrogated bythe G2A mutant, while the MHC class I down-regulating capacityof the panel of Nef mutants showed no resemblance to the otherreceptor modulations. In conclusion, these data suggest thatdown-regulation of CD4, CD28, and CD8 ${alpha}$ ß and the up-regulationof CD74 result from related functions of Nef.

Nef induces accelerated internalization of the CD8 ${alpha}$ ß receptor resulting in a down-regulated steady state surface expression. To explore the mechanism of down-regulation, a comparison ofNef-mediated CD4 and CD8 ${alpha}$ ß down-regulation was donein transduced SupT1 cells in function of time. Therefore, weused an inducible NA-7 Nef.ER fusion protein (61), activatedat time zero of the experiment by adding 1 µM 4-hydroxytamoxifento the cells. As shown in Fig. 3A, Nef-mediated down-regulationkinetics of both receptors were comparable: within 2 h, halfof the steady-state level (reached after approximately 10 h)of down-regulation is achieved. These fast kinetics are dueto accelerated internalization of cell surface receptors, asshown with a fluorescence-activated cell sorting-based assayusing the inducible Nef.ER (Fig. 3B). Nef induced an accelerationof the endogenous internalization of CD4, CD8 ${alpha}$ ß, andCD28. The basal (Nef-negative) and Nef-mediated (Nef-positive)internalization kinetics for CD4 and CD8 ${alpha}$ ß were almostidentical, while those measured for CD28 were both higher.

View larger version (13K):
[in this window]
[in a new window]
FIG. 3. Flow cytometric analysis of Nef-mediated receptor down-regulation and internalization in retrovirally transduced SupT1 cells. SupT1 cells were transduced with an inducible NA-7.ER construct. At time zero, 4-hydroxytamoxifen (1 µM) was added to the culture medium. (A) Percent down-regulation was calculated, as described in the Materials and Methods. CD4-allophycocyanin (solid line) and CD8ß-phycoerythrin (dashed line) were measured as a function of time. (B) The figure shows the percentage of CD4, CD8 ${alpha}$ ß, and CD28 molecules internalized by HIV-1 NA-7.ER, calculated as described in Materials and Methods. In each graph, Nef-positive (EGFP expressing Nef, solid line) and Nef-negative (EGFP not expressing Nef, dashed line) cells are depicted. The EGFP ranges used for calculation are indicated in Fig. 1A.

Mutational analysis of the CD8ß cytoplasmic tail. In other known Nef-targeted surface receptors, essential residuesin the cytoplasmic tail have been identified. To determine theseresidues in the cytoplasmic tail of the CD8 ß-chain,series of Daudi cells, coexpressing CD8 ${alpha}$ and a CD8ßdeletion mutant, were transduced with wild-type Nef to evaluatethe steady-state down-regulation levels (Fig. 4A and 4C) andinternalization kinetics (Fig. 4B and 4D) of the CD8 ${alpha}$ ßreceptors. In the absence of Nef, both wild-type and mutantCD8 ${alpha}$ ß showed low levels of down-regulation (Fig. 4C)and internalization kinetics (wild-type Nef-negative in Fig.4B and data not shown). However, in the presence of Nef, truncationof the CD8 ß-chain at amino acid 206 weakened down-regulationand internalization by Nef (Fig. 4A and 4B). Deletion of threemore amino acids further decreased the steady state down-regulation±2-fold, suggesting that amino acids 204 to 206 wereimportant for this Nef effect. Further truncating the CD8ßtail did not have any additional effect. Remarkably, deletionof the complete cytoplasmic ß-tail could not entirelyblock CD8 ${alpha}$ ß down-regulation.

View larger version (33K):
[in this window]
[in a new window]
FIG. 4. Mutations in the CD8 ß-chain and their effect on endocytosis and down-regulation. Daudi cells were cotransduced with CD8 ${alpha}$ , wild-type or mutant CD8ß, and control or wild-type Nef. (A and C) Alignment of amino acid sequences of wild-type (210*) and mutant CD8 ß-chain cytoplasmic tails. An asterisk indicates a stop codon. (A) The bar chart shows the percent down-regulation of (mutant) CD8 ${alpha}$ ß by HIV-1 Nef alleles NA-7, LAI, and NL4-3. In both B and D the percentage of (mutant) CD8 ${alpha}$ ß molecules internalized by wild-type HIV-1 NA-7 is shown, including in both the same data for a Nef-negative construct as a control (wild-type Nef–). (C) The bar chart represents the percentage of (mutant) CD8 ${alpha}$ ß down-regulation after transduction with either control virus or wild-type Nef NL4-3. Each bar represents a (mutant) CD8 ß-chain, as indicated by the changed amino acid sequence compared with CD8 ${alpha}$ ß. Percent down-regulation and internalization were calculated as described in Materials and Methods, using the ranges indicated in Fig. 1A. In A, B, and D, mean values are shown and standard deviations are calculated from the data generated from three independent experiments. In B the results for NL4-3 are representative of the results with HIV-1 alleles NA-7 and LAI.

In a next step, we generated targeted amino acid substitutionsin this particular cytoplasmic region and performed down-regulationand internalization experiments. As shown in Fig. 4C and 4D,the weakest down-regulation and internalization potential ofNef was observed with the mutant 204/5/6 Ala, followed by themutants 204/5/6 Ser, 204/6 Ser, and 204 Ala, suggesting an importantfunction for residue 204. Interestingly, for reasons that areunclear, the 206A mutant showed an even higher internalizationrate than wild-type CD8ß. However, this was not reflectedby its Nef-induced down-regulation, which was lower than thecontrol. In summary, our mutational analysis of the CD8ßcytoplasmic tail revealed a C-terminal motif (FMK) implicatedin Nef-mediated internalization, in which phenylalanine residue204 seemed to be the most important. The residual down-regulation(approximately 20%) and internalization potential observed suggestsome role for the CD8 ${alpha}$ cytoplasmic tail.

CD8 ${alpha}$ cytoplasmic tail is not required for Nef-mediated down-regulation of the CD8 ß-chain. To investigate the role of the CD8 ${alpha}$ cytoplasmic tail in Nef-mediateddown-regulation, we designed a CD8 ${alpha}$ (EC+TM)-CD8ß(IC)chimera and stably expressed this molecule in Daudi cells. TheCD8 ${alpha}$ ß chimeric receptor and the CD8 ${alpha}$ ${alpha}$ chimeric controlreceptor showed comparable high expression levels, as judgedby MFI (Fig. 5, control). In contrast to the construct withthe CD8 ${alpha}$ cytoplasmic tail, which was only moderately affectedby HIV-1 Nef or SIVmac239 Nef, down-regulation of the chimericCD8 ${alpha}$ ß surface marker was very pronounced followingNef expression (Fig. 5). Indeed, by combining two CD8ßcytoplasmic tails (in the chimera), instead of a CD8 ${alpha}$ and a CD8ßcytoplasmic tail, the down-regulating capacity of Nef was elevateddrastically, pointing towards a preferential, direct or indirectinteraction between Nef and the short cytoplasmic CD8ßtail. We conclude that, while the CD8 ${alpha}$ cytoplasmic tail is dispensablefor Nef-mediated CD8 ${alpha}$ ß down-regulation, the CD8ßcytoplasmic tail harbors a potent target sequence for down-regulation.

View larger version (42K):
[in this window]
[in a new window]
FIG. 5. Chimeric constructs. Daudi cells were retrovirally transduced with the CD8 ${alpha}$ (EC-TM)-CD8 ${alpha}$ (IC) chimera (cyt tail) or a CD8 ${alpha}$ (EC-TM)-CD8ß(IC) chimera (cyt tail), using bicistronic constructs with ${Delta}$ NGFR as the reporter. Bivariate dot plots are gated on ${Delta}$ NGFR-positive cells, at day 2 after transduction of these cells with control virus, HIV-1 Nef (NA-7 allele), and SIV Nef (mac239), using bicistronic constructs with EGFP as the reporter. CD8 ${alpha}$ -phycoerythrin versus EGFP expression is shown.

Colocalization of Nef and CD8 ß-chain. To gain more insight into the intracellular localization ofCD8 ${alpha}$ ß in Nef expressing cells, 293T cells expressingboth CD8 ${alpha}$ ß and Nef.EGFP fusion protein were examinedby confocal microscopy. Nef was localized in the cytoplasm andat the cellular membrane. In the absence of Nef, the CD8 ß-chainwas found at the cellular membrane and showed some cytoplasmicstaining (Fig. 6, lower panel, Nef.EGFP-negative cells). Inthe presence of EGFP-tagged wild-type Nef and the PPAA Nef mutant,both being able to down-regulate HA-tagged CD8ß (datanot shown), the CD8 ß-chain was localized in submembranousvacuoles (Fig. 6, upper and lower panels). In contrast, membranousCD8ß expression was retained in cells coexpressingthe Nef mutant LLAA, which fails to downregulate CD8ß(Fig. 6, middle panel).

View larger version (64K):
[in this window]
[in a new window]
FIG. 6. Confocal images of 293T cells. Nef.EGFP was detected by direct fluorescence (green, left panels) and CD8ß.HA or CD8ß by monoclonal antibodies as indicated in Materials and Methods (red, middle panels). Nuclei were visualized by DAPI staining (blue). Right panels show the merged images from Nef.EGFP and CD8ß. Areas of colocalization of Nef.EGFP/CD8ß are shown in yellow. As indicated, the upper panels show cells expressing wild-type LAI, the middle panels show the LLAA mutant, and alower panels show the PPAA mutant. Scale bars represent 5 µm.

Repression of Nef-mediated CD8ß endocytosis. Based on strategies, used to abrogate Nef-mediated CD4 endocytosis(37, 40, 55), repression of Nef-mediated CD8ß endocytosiswas tested. A decline in Nef-mediated CD8 ${alpha}$ ß internalizationwas observed after addition of ikarugamycin, a macrolide antibioticpostulated to be a general inhibitor of clathrin-coated pit-mediatedendocytosis (40), to Nef-expressing Daudi CD8 ${alpha}$ ß cultures(Fig. 7A). To block distinct steps in endocytic trafficking,lentiviral RNA interference (RNAi) constructs were designed,expressing EGFP as well as a short hairpin RNA, to induce knockdownof three key molecules in clathrin-mediated endocytosis, i.e.,AP-2 complexes by µ2 chain RNAi (AP-2 RNAi), clathrinheavy chain (Chc RNAi), and dynamin-2 (Dyn-2 RNAi).

View larger version (39K):
[in this window]
[in a new window]
FIG. 7. Blocking Nef-mediated internalization and down-regulation by ikarugamycin and RNA interference. (A) The figures show the percentage of CD8 ${alpha}$ ß molecules internalized in Daudi CD8 ${alpha}$ ß cells by HIV-1 NA-7.ER. Cells were incubated for 2 h with (IKA+) or without (IKA–) ikarugamycin (2 µM) prior to the internalization experiment. At time zero, 4-hydroxytamoxifen (1 µM) was added to NA-7.ER-transduced Daudi cells. (B) Western blot, performed as indicated in Materials and Methods, shows protein expression levels of AP-2 µ2 subunit, clathrin heavy chain (Chc), and dynamin 2 (Dyn-2, arrowhead) in control and RNAi-transduced SupT1 cells, with equal amounts of protein loaded. (C) Bivariate dot plots of flow cytometric measurement of SupT1 cells transduced with AP-2i and HIV-1 Nef (LAI). CD8 ${alpha}$ ß (phycoerythrin) versus ${Delta}$ NGFR (allophycocyanin) expression is shown, gated on EGFP-negative and EGFP-positive cells. (D) The bar charts represent the effect of AP-2, clathrin heavy chain (Chc), and dynamin 2 RNAi on HIV-1 Nef (LAI) and SIV (mac239) Nef-induced CD4 (left panel) and CD8 ${alpha}$ ß (right panel) down-regulation in transduced SupT1 cells, gated on EGFP-negative and EGFP-positive cells. Percent down-regulation was calculated, as described in Materials and Methods, using the ranges N– and N+, as indicated in C. Mean values and standard deviations are shown, calculated from data generated from three independent experiments.

Transduction of SupT1 cells with any of these RNAi constructsresulted in a strong depletion of the respective proteins (Fig.7B), which had functional implications, as evidenced by a morethan fivefold increase of the transferrin receptor surface expression(data not shown). Nef and control constructs were expressedfrom a bicistronic construct expressing a truncated human nervegrowth factor receptor ( ${Delta}$ NGFR) as the reporter, allowing simultaneousflow cytometric evaluation of the expression levels of RNAi(EGFP), Nef ( ${Delta}$ NGFR), and a surface molecule (e.g., CD4) (Fig.7C). Loss of AP-2 complexes in SupT1 cells markedly impairedHIV-1 or SIVmac239 Nef-mediated down-regulation of CD4 and CD8 ${alpha}$ ßcompared to the control (p53 RNAi) (Fig. 7D). The knockdownof clathrin heavy chain had no effect on Nef-mediated receptorendocytosis. Remarkably, knocking down dynamin-2 hampered CD4and CD8 ${alpha}$ ß down-regulation by HIV-1 Nef, but not bySIVmac239 Nef. This inhibition was less pronounced with theNA-7 allele (data not shown). In conclusion, these experimentsdemonstrate a common role for the AP-2 µ2 subunit in HIV-1and SIVmac239 Nef-mediated receptor internalization and pointto a similar mechanism used by Nef to down-modulate CD4 andCD8 ${alpha}$ ß surface expression.

DISCUSSION

Top

Discussion

In this study, we demonstrated that HIV-1, SIVmac239, and HIV-2Nef functionally interact with the CD8 ß-chain ofthe CD8 ${alpha}$ ß receptor, leading to an accelerated internalizationof this receptor. Mutational analysis of several HIV-1 Nef allelesand mechanistic exploration indicated that Nef-mediated CD8 ${alpha}$ ßand CD4 endocytosis are closely related effects, suggestingthat Nef utilizes similar mechanisms to internalize both receptors.

Based on our confocal images, showing colocalization of theCD8 ß-chain with the HIV-1 Nef protein in perinuclearvacuoles, and our observation that the CD8 ${alpha}$ -chain is dispensablefor the down-regulation effect, we suggest that Nef interactsdirectly or indirectly with the CD8ß cytoplasmic tail.Our attempts to demonstrate a molecular interaction betweenNef and CD8 in mammalian cells all failed (data not shown),which might not come as a surprise, as the interaction betweenNef and CD4, suggested to be weak and/or transient, has alsoonly been demonstrated in vitro and in insect cells (27, 29,52, 56).

Although the mechanism of Nef-mediated down-regulation of CD4,the receptor for HIV virions, has already been intensively studiedfor more than a decade, the ability of Nef to internalize theother T-cell coreceptor, i.e., CD8 ${alpha}$ ß, was never addressedin human cells. This can be explained by the fact that the mostconvincing evidence for the susceptibility of CD8 lymphocytesto HIV infection has been gathered only in the last years (7,18, 30, 39, 44, 60). Moreover, most laboratories are using anantibody assigned to the CD8 ${alpha}$ -chain for CD8 flow cytometricmeasurements instead of the 2ST8.5H7 clone, having an epitopeon the CD8 ${alpha}$ ß heterodimer. The use of the CD8 ${alpha}$ monoclonalantibody clone also explains why Garcia et al. claimed thatNef fails to affect the human CD8 receptor (20). However, ourresults in Daudi CD8 ${alpha}$ ${alpha}$ cells indicate that there is some effecton the CD8 ${alpha}$ -chain, suggesting a minor interaction of Nef withthe CD8 ${alpha}$ cytoplasmic tail.

Previous reports have provided insight into the mechanisms involvedin Nef-mediated CD4 down-regulation. Nef accelerates normalCD4 clathrin-dependent internalization by functioning as a connectorbetween CD4 and the AP component of the clathrin complex. Ina next step, the CD4 receptor is not recycled to the membrane,as occurs in normal conditions, but is misrouted by Nef to thelysosomes. In contrast to the dynamic trafficking propertiesof CD4, the CD8 ${alpha}$ ${alpha}$ and CD8 ${alpha}$ ß receptors are slowly endocytosed,regardless of the presence of p56^lck or phorbol esters (6, 43).In our internalization assays 0 to 4% of CD8 ${alpha}$ ß wasfound intracellularly in Daudi CD8 ${alpha}$ ß and SupT1 cellsat equilibrium (25 min), increasing to 25 to 35% in the presenceof Nef.

Efficient endocytosis of cell surface glycoproteins throughclathrin-coated vesicles requires the presence of endocytosissignals (49). Whereas the cytoplasmic tails of CD4, CD28, CD74,and MHC class I all contain a putative sorting signal (35, 59,63), this is not the case for the CD8 ${alpha}$ and ß-chaincytoplasmic tails. However, as Nef can bring its own dileucinesorting signal, its recruitment, directly or indirectly, tothe cytoplasmic tail may be sufficient to induce acceleratedCD8 ${alpha}$ ß endocytosis, raising the question of whethera native CD8 endocytic signal would even be involved in thisprocess. In this respect, mutational analysis of the CD8ßtail showed a sequence (FMK), close to the CD8ß carboxyterminus, to be essential for efficient Nef-mediated CD8ßendocytosis.

Remarkably, we still observed a residual CD8 ${alpha}$ ß down-regulationof this mutant, as well as of the cytoplasmic tail-deleted CD8ß-chain mutant, probably reflecting the modulatingcapacity of Nef on the CD8 ${alpha}$ -chain. The FMK sequence is to ourknowledge not part of a known endocytic consensus sequence and,remarkably, is present in all membrane-associated splice variantsof CD8ß which, through alternative splicing, onlystart to differ immediately C-terminal of this sequence (48).Based on information from secondary-structure prediction methods(45), this FMK sequence would be part of a helix structure whichcomprises almost the complete CD8ß cytoplasmic tail.Similarly, residues in the cytoplasmic tail of CD4, necessaryfor Nef-mediated CD4 down-regulation, are located within thehelical part of the CD4 cytoplasmic tail. Based on mutationalstudies, a correlation was demonstrated between the presenceof an ${alpha}$ -helix in CD4 and its susceptibility to down-regulationby Nef (53), which might also be the case for Nef-mediated CD8 ${alpha}$ ßdown-regulation.

Our results demonstrate that at least three sites within Nefare required for CD8 ${alpha}$ ß down-regulation, the myristoylationsignal and N-terminal anchor regions, the C-terminal flexibleloop, and amino acid positions 57 to 58. Consistent with allreported Nef functions, the myristoylation signal was foundto be essential for CD8 ${alpha}$ ß down-modulation. The flexibleloop contains a dileucine-based internalization motif, whichis flanked by acidic clusters (¹⁵⁴EEX₈LLX₈DD¹⁷⁵), and is involvedin enhanced internalization of the Nef-CD4 complex (23). Thisdileucine sequence, located at amino acid positions 164 to 165of the Nef molecule, is involved in the association of HIV-1Nef with the AP complex (8). Consequently, mutations in theflexible loop at these dileucine sequences or at acidic positions174 to 175 drastically changed the modulating capacities ofNef, preserving only MHC class I down-regulation (Fig. 2B) (57).

Mutation of amino acid positions 57 to 58, denoted the CD4 interactionsite, abolished the capacity of Nef to modulate CD8 ${alpha}$ ßas well as CD4 surface expression. Furthermore, we found thissequence to be required for CD74 upregulation and, in agreementwith Swigut and coworkers, for CD28 internalization (63). HowNef can use the same acceptor site (or overlapping sites) toform a complex with both CD4 and CD28, as well as the CD8 ß-chainand CD74, even when the target sequences within these receptorsare different, remains an open question. Even more intriguingis our observation that Nef apparently can use the same mechanismto down-regulate and to up-regulate (CD74) receptor surfaceexpression levels.

A number of studies demonstrate that most if not all Nef functionsare genetically separable (57). However, based on the resultsof our structure-function analysis, we could not geneticallyseparate the ability of HIV-1 Nef to down-regulate CD8 ${alpha}$ ßand CD4. Mutation of the complete acidic cluster at amino acidpositions 62 to 66 separated the effect of HIV-1 Nef on CD8 ${alpha}$ ß/CD4and CD28 expression, while amino acids 41 to 49 (removed inDel3) were required for CD8 ${alpha}$ ß/CD4/CD28 down-regulationbut were dispensable for down-regulation of MHC class II andup-regulation of the Ii chain. Knockdown of the traffickingmolecules AP-2, clathrin heavy chain, or dynamin-2 showed noprofound difference in reversal of the HIV-1 Nef effect on CD4and CD8 ${alpha}$ ß, corroborating our results with the mutantNef panel.

Surprisingly, Nef-mediated endocytosis could still occur afterknockdown of clathrin heavy chain but not AP-2, suggesting thattrace amounts of clathrin heavy chain are sufficient for clathrin-coatedendocytosis, while AP-2 is present at rate-limiting amounts.Although the LAI allele of HIV-1 Nef was affected more stronglyby dynamin-2 RNAi than the NA-7 allele (data not shown), wedid not observe any effect on the SIVmac239 Nef-mediated down-regulatorypotency, pointing to a difference between HIV-1 and SIVmac239Nef in the need for endocytic cargo molecules. However, we didnot observe this difference after knockdown of AP-2, suggestinga central role for AP-2. This result is in contrast with Roseet al., who demonstrated a stronger inhibition of SIV than ofHIV-1 Nef-mediated CD4 down-regulation (55). This dissimilarityin potency can be due to the different experimental set-ups(several electroporation steps of short interfering RNA) andcells (HeLa) used by Rose and coworkers.

Various mechanisms for HIV entry into CD8 lymphocytes have beenproposed, with entry through a conventional CD4-dependent pathwayas the most plausible one. This infection route can occur duringintrathymic CD8 lymphocyte development, at the CD4⁺ CD8⁺ doublepositive stage (14), or upon activation of the mature CD8 lymphocyte,which leads to the coexpression of the CD4 receptor on the cellsurface (18, 36). Recently, several groups provided evidencethat, while HIV-infected CD8 precursors from the thymus rarelyreach the periphery, the majority of circulating infected CD8lymphocytes acquired HIV through expression of CD4 during activation(7, 12).

HIV-infected CD8⁺ T-cell precursors have been suggested to bedepleted intrathymically. However, as previously reported, CD8 ${alpha}$ ß/CD4down-regulation alone cannot explain Nef-mediated impaired T-celldevelopment, and other functions of Nef and/or HIV are neededfor its induction of thymic depletion (61). Although CD8 ${alpha}$ ß⁺CD4^lowT lymphocytes account for a minor fraction of peripheral bloodCD8 lymphocytes (<5%), HIV patients were found to harborhigh levels of infection in this CD8 subset, approaching thosefound in CD4 lymphocytes, whereas CD8 ${alpha}$ ß⁺CD4^–T lymphocytes showed very low or even undetectable viral DNAloads (12). Possibly, the ability of HIV to infect this activatedCD8⁺ population may contribute to the decline in CD8 lymphocytefunction which is at present predominantly ascribed to the lackof CD4 lymphocyte help and viral escape (reviewed by McMichaeland Rowland-Jones [46]).

Remarkably, although the two chains of the CD8 ${alpha}$ ß receptorare covalently linked, suggesting a joined internalization ofCD8 ${alpha}$ and CD8ß, Nef modulates the surface expressionof CD8 ${alpha}$ to a much lower extent than CD8ß in peripheralblood mononuclear cells. A possible explanation may be the simultaneousexpression of CD8 ${alpha}$ ${alpha}$ and CD8 ${alpha}$ ß on the cell surface, withNef predominantly affecting the CD8 ${alpha}$ ß surface expression.Whereas CD8 ${alpha}$ ${alpha}$ is not expressed on resting naïve T cells inthe periphery (38), coexpression of CD8 ${alpha}$ ${alpha}$ and CD8 ${alpha}$ ß onactivated conventional T cells and in T-cell leukemias has beenreported (64). Hence, in the presence of Nef, activated lymphocytesmight get skewed towards a predominant CD8 ${alpha}$ ${alpha}$ receptor surfaceexpression.

In conclusion, this work describes the down-modulation of theCD8 ${alpha}$ ß surface expression by Nef. Analysis of the mechanismby which Nef down-regulates CD8 ${alpha}$ ß revealed an acceleratedAP-2 complex-mediated internalization of the receptor, closelyrelated with the mechanism of Nef-mediated CD4 down-modulation.Further functional experiments are warranted to elucidate whetherNef impairs CD8 lymphocyte function, thus affecting the immunecontrol of both HIV and opportunistic pathogens.

ACKNOWLEDGMENTS

We thank N. De Cabooter, E. Demecheleer, and S. De Keyser fortechnical assistance, Y. Taes for help with immunoblot imaging,C. Bonini for ${Delta}$ NGFR, O. Schwartz, Y. Collette, D. Olive, W. Greene,R. Benarous, J. Zack, and the AIDS Research and Reference ReagentProgram, Division of AIDS, NIAID, for the gift of the wild-typeand mutant nef constructs, P. Charneau for providing the pTRIPvector, D. Trono for the p8.91 and pMD.G plasmids, S. Bonattifor the CD8 ${alpha}$ and CD8ß constructs, and H. Göttlingerfor the dynamin-2 RNAi sequence. We also acknowledge J. Hoxiefor providing the hybridoma cell line EH-1.

V.S. is a Research Assistant, E.C. and C.S. are PostdoctoralFellows, and B.V. is a Senior Clinical Investigator of the Fundfor Scientific Research-Flanders (Belgium) (FWO -Vlaanderen).This work was supported by grants from the FWO -Vlaanderen.

FOOTNOTES

* Corresponding author. Mailing address: Department of Clinical Chemistry, Microbiology and Immunology, Ghent University Hospital, 4 Blok A De Pintelaan 185, B-9000 Ghent, Belgium. Phone: 32 9 240 22 26. Fax: 32 9 240 36 59. E-mail: bruno.verhasselt{at}UGhent.be.

REFERENCES

Top