Palmitoylation-dependent control of degradation, life span, and membrane expression of the CCR5 receptor.

We have shown that the chemokine and HIV receptor CCR5 is palmitoylated on a cluster of cysteine residues located at the boundary between the seventh transmembrane region and the cytoplasmic tail. Single or combined substitutions of the three cysteines (Cys-321, Cys-323, and Cys-324) or incubation of wild-type CCR5-transfected cells with the palmitic acid analog 2-bromopalmitate prevented palmitoylation of the receptor. Moreover, failure of CCR5 to be palmitoylated resulted in both accumulation in intracellular stores and a profound decrease of membrane expression of the receptor. Upon metabolic labeling, kinetic experiments showed that the half-life of palmitoylation-deficient CCR5 is profoundly decreased. Bafilomycin A1, but not a specific proteasome inhibitor, prevented early degradation of palmitoylation-deficient CCR5 and promoted its accumulation in lysosomal compartments. Although membrane expression of the CCR5 mutant was diminished, the molecules reaching the membrane were still able to interact efficiently with the chemokine ligand MIP1 beta and remained able to function as HIV co-receptors. Thus we conclude that palmitoylation controls CCR5 expression through regulation of the life span of this receptor.

We have shown that the chemokine and HIV receptor CCR5 is palmitoylated on a cluster of cysteine residues located at the boundary between the seventh transmembrane region and the cytoplasmic tail. Single or combined substitutions of the three cysteines (Cys-321, Cys-323, and Cys-324) or incubation of wild-type CCR5-transfected cells with the palmitic acid analog 2-bromopalmitate prevented palmitoylation of the receptor. Moreover, failure of CCR5 to be palmitoylated resulted in both accumulation in intracellular stores and a profound decrease of membrane expression of the receptor. Upon metabolic labeling, kinetic experiments showed that the half-life of palmitoylation-deficient CCR5 is profoundly decreased. Bafilomycin A1, but not a specific proteasome inhibitor, prevented early degradation of palmitoylation-deficient CCR5 and promoted its accumulation in lysosomal compartments. Although membrane expression of the CCR5 mutant was diminished, the molecules reaching the membrane were still able to interact efficiently with the chemokine ligand MIP1␤ and remained able to function as HIV co-receptors. Thus we conclude that palmitoylation controls CCR5 expression through regulation of the life span of this receptor.
G-protein-coupled receptors (GPCRs) 1 are formed by seven transmembrane hydrophobic ␣-helixes connected by extra-and intracellular loops. A glycosylated extracellular amino-terminal domain and a cytoplasmic carboxyl-terminal domain complete the structure of this superfamily of proteins, which transduces signals from a large and disparate number of extracellular ligands (1)(2)(3). Among GPCRs, chemokine receptors mediate the biological effects of chemokines that are structurally related to pro-inflammatory cytokines and are primarily involved in the control of constitutive and pathological migration of leukocytes (4,5). However, a rapidly growing body of evidence shows that functions of chemokine receptors are not limited to locomotion and extends to a vast array of immune responses, organ and tissue development, as well as the pathogenesis of inflammatory, cardiovascular, and infectious diseases (6,7).
Two of these chemokine receptors, CXCR4 and CCR5, have been identified as co-receptors that allow efficient entry of HIV into CD4ϩ host cells (8). Upon binding to CD4, the interaction of HIV isolates with CXCR4 or CCR5 (X4 and R5 isolates, respectively) relies on the structure of the HIV envelope (Env) glycoprotein, gp120 subunit. The binding of the HIV Env complex (gp120/gp41 subunits) to the receptors CD4 and CXCR4 or CCR5 initiate conformational changes in the gp120/gp41 oligomer, leading to membrane fusion and, ultimately, internalization of the viral capsid into the host cell (9,10).
CCR5 plays a critical role in the transmission and early propagation of HIV-1 (11)(12)(13)(14). Thus, R5 isolates predominate for years after primo-infection and are very often the only variants detected in AIDS patients. Identification of a CCR5 gene polymorphism results in expression of severely truncated molecules that fail to reach the cell surface and render individuals highly resistant to viral infection (15)(16)(17). This pointed to the essential role of this chemokine receptor in HIV-1 entry.
It has been shown that functional interaction of CCR5 with its natural ligands or with HIV Env involves preferentially the first two extracellular loops of the receptor, as well as the amino-terminal domain (18,19). Besides this structural role, it has been postulated that CCR5-mediated signal transduction upon binding of the viral Env is required for infection by R5 isolates. However, when overexpressed, ectopic CCR5 molecules that cannot support agonist-dependent signaling or undergo endocytosis still exhibit an HIV co-receptor function (20 -23). In contrast, a more relevant biological system using primary T lymphocytes showed that inhibition of G␣ i or specific desensitization of CCR5 prevents replication of R5 isolates in these cells (24 -26). Thus, the participation of CCR5-dependent signaling in HIV infection remains controversial (10).
This situation is especially unclear because other aspects of CCR5 biology that may influence the capacity of the receptor to transduce signals have not been investigated. Such is the case of palmitoylation, which occurs at cysteine residues located at the boundary of the seventh transmembrane domain with the cytoplasmic tail of many GPCRs (27). The rhodopsin receptor was the first GPCR for which palmitoylation was demonstrated (28,29). It was suggested that this post-translational modification may lead to structural changes of the cytoplasmic tail and promote the formation of a fourth cytoplasmic loop that could influence GPCR function (30). This thioesterification of cysteine residues by palmitate is distinct from other lipid modifications, such as prenylation and myristoylation, by its re-* This work was supported in part by a grant from Ensemble contre le SIDA (sidaction). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  versibility and its dynamic regulation (31). Palmitoylation is particularly prevalent among soluble proteins implicated in cell adhesion, growth, and signaling, allowing their attachment to the cytosolic face of the plasma membrane (32). However, in the case of transmembrane receptors, the high frequency of cysteine residues in the carboxyl-terminal region of many GPCRs is suggestive of an additional role for this type of modification.
Palmitoylation of GPCRs has been found to cover a broad spectrum of biological activities (33), including G-protein coupling efficiency (34 -38) and control of receptor phosphorylation and desensitization (39 -41). Furthermore, palmitoylation has been shown to regulate intracellular trafficking of some GPCRs (42)(43)(44)(45). Although palmitoylation of GPCRs is well conserved, its biological role varies from one member of the superfamily to another, and a general model cannot be proposed. This applies especially to chemokine receptors, because palmitoylation of no member of this particular family has been investigated.
The present study was undertaken to examine whether the three cysteine residues present in the carboxyl-terminal tail of CCR5 undergo palmitoylation. The possible functional role of this lipid modification in the stability of the receptor, its proper membrane expression, and its capacity to interact with its natural and viral ligands were assessed.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection-Parental HeLa cells or indicator HeLa CD4 LTRLacZ (46) and HEK-293T human embryonic kidney cells stably expressing large T antigen were grown in Dulbecco's modified Eagle's medium (DMEM) (Life Technologies, Inc.), supplemented with 10% fetal calf serum (FCS) and antibiotics. Cells in exponential growth were transiently transfected using LipofectAMINE plus (Life Technologies, Inc.; HeLa) and the calcium phosphate co-precipitation method (HEK-293T). All transfections (1 ϫ 10 6 cells) were done with 1 g of DNA, except when indicated otherwise.
Construction of Mutated CCR5 cDNAs-Mutagenesis of the plasmid encoding WT CCR5 cDNA cloned into pcDNA 3 was done to replace CCR5 cysteine residues at positions 321, 323, and 324 of the carboxylterminal tail by overlap extension using T7, Sp6, and two internal primers containing the mutation (47). The forward primer used to mutate the three cysteines, either separately (simple mutation) or together (double and triple mutations), to alanine residues (Cys to Ala) was , and Z 1 Z 2 Z 3 (Cys-324) correspond in the WT sequences to TGC, TGC, and TGT, respectively, and to GCC, GCC, and GCT when the cysteine residue was substituted by an alanine. The reverse primer sequence used was X 4 X 5 X 6 Y 4 Y 5 Y 6 T-TTZ 4 Z 5 Z 6 GAAGCGTTTGGCAATGTGCTTTTGGAA, where X 4 X 5 X 6 , Y 4 Y 5 Y 6 , and Z 4 Z 5 Z 6 were ACA, GCA, and GCA, respectively, in the WT and AGC, GGC, and GGC when the cysteine residue was substituted by an alanine. For serine replacement, (Cys to Ser) WT codons in each position were substituted by AGC, AGC, and AGT on the forward primer and by ACT, GCT, and GCT on the reverse primer. cDNA mutants were confirmed by sequence analysis.
Metabolic Labeling-Prior to use, [1-14 C]palmitic acid (Amersham Pharmacia Biotech) was dried under vacuum and solubilized in Me 2 SO. Cells were washed with phosphate-buffered buffer (PBS) and exposed for 3 h either to DMEM medium containing 8% dialyzed FCS and [ 14 C]palmitate (100 Ci/ml) or to methionine (Met)-free DMEM medium (Sigma-Aldrich Inc.) containing 8% dialyzed FCS and [ 35 S]Met (100 Ci/ml) (Amersham Pharmacia Biotech). For pulse-chase experiments, after metabolic labeling, cells were washed twice and cultured for the indicated time in DMEM supplemented with 10% FCS (chase).
Western Blot Analysis and Immunoaffinity Purification-For Western blot experiments, cells were lysed in 20 mM Tris-HCl, pH 7.5, 100 mM NaCl, 10 mM EDTA, 1% Triton X-100. Immunoblotting was performed using an anti-human CCR5 monoclonal antibody, MC-5 (48). Simultaneous detection of ␤-tubulin or lactate dehydrogenase (LDH) was performed using monoclonal (BD PharMingen) or polyclonal (Interchim) antibodies, respectively. Immobilized antigen-antibody complexes were detected with secondary anti-species IgG-horseradish peroxidase conjugates (Pierce) and an enhanced chemiluminescence detection system (Pierce). Acquisition and quantification of chemiluminescence signals were done using a LAS-1000ϩ apparatus (Fuji). For immunoprecipitation, after labeling with [ 35 S]Met or [ 14 C]palmitate, cells were lysed by incubation for 1 h at 4°C in lysis buffer (20 mM Tris-HCl, pH 7.5, 100 mM ammonium sulfate, 10% glycerol, 1% Cymal-5 (Anatrace)), as previously described (49). After removing insoluble material, 1 mg of lysate was incubated with 2 g of anti-CCR5 antibody 2D7 (19) for 12-16 h at 4°C before adding protein A/G-Sepharose beads (Calbiochem) for 2 h at 4°C. Immunoprecipitates collected by centrifugation were electrophoresed directly on SDS-polyacrylamide gels (PAGE) at 4°C. After migration, [ 14 C]palmitate and [ 35 S]Met gels were first fixed in methanol/acetic acid/H 2 O (5:1:4, v/v) for 30 min, then in NAMP 100 Amplify (Amersham Pharmacia Biotech) for 30 min at room temperature before being dried and exposed for 3 days to 1 month ( 35 S and 14 C labeling, respectively). For quantification analysis, gels were recorded with a digital imaging system and computer-generated images were analyzed to obtain densitometric values (PhosphorImager 2850, Molecular Dynamics program). Statistical analysis was performed using the Spearman test for correlation between incorporation of [ 14 C]palmitate and CCR5 expression. A p value of Ͻ0.05 was considered significant.
Immunofluorescence Microscopy-CCR5-transfected HeLa (1 ϫ 10 5 ) cells were seeded onto glass coverslips in 24-well plates and stained by indirect immunofluorescence 24 h later. Cholera toxin (CTx) is pentavalent for ganglioside GM1 and causes formation of clusters of five GM1 molecules (50). Phycoerythrin (PE)-conjugated CTx (5 g/ml, Sigma Chemical Co.) was applied to transfected cells for 30 min at 4°C and removed by one wash in RPMI containing 1% bovine serum albumin (BSA) and 20 mM Hepes. Anti-CTx monoclonal antibody (1/150 dilution, Sigma) was applied to the cells for 30 min at 4°C, the cells were washed in PBS containing 0.5% BSA (blocking buffer), and indirect immunofluorescence was performed as previously described (51). Briefly, cells were fixed with 4% paraformaldehyde in PBS and, when indicated, permeabilized by incubation in blocking buffer containing 0.05% saponin. Specific antibodies (anti-CCR5 antibody 2D7 and anti-Lamp1 (PharMingen)) were applied for 1 h, and cells were washed in blocking buffer and incubated for 1 h with Texas Red-or fluorescein isothiocyanate (FITC)-conjugated secondary antibodies (1/300 dilution, Southern Biotech Corp.). Coverslips were mounted in Mowiol (Hoechst). Confocal laser scanning microscopic analysis was performed on a Leica TCS64D instrument using an X63 oil immersion PL APO objective.
Flow Cytometric Analysis of Cell Surface Antigens-To determine cell surface expression of mutants CCR5 molecules, 5 ϫ 10 5 cells were resuspended in PBS containing 1% BSA, 0.1% ␥ globulin, 0.1% sodium azide for 10 min at room temperature before staining with anti-CCR5 PE-conjugated 2D7 antibody (PharMingen). After three washes, cells were maintained in PBS containing 4% formaldehyde and analyzed on a Becton Dickinson FACSCalibur.
MIP1␤ Binding Assays-HEK-293T cells transiently transfected with WT CCR5 cDNA or triple Cys 1,2,3 mutant cDNA, were first incubated 1 h at 37°C in DMEM containing 5 milliunits/ml heparitinase I (EC 4.2.2.8) (Seikagaku Corp.) to cleave and remove from the cell surface heparan sulfate that could attach MIP1␤ nonspecifically. After two washes in cold PBS, cells were incubated in PBS containing 0.25 nM iodinated MIP1␤ (2200 Ci/mmol, PerkinElmer Life Sciences), and various concentrations of unlabeled MIP1␤ for 1 h at 4°C in a total volume of 200 l. Cell pellets were washed twice in ice-cold PBS. Nonspecific binding was determined in the presence of 1 M unlabeled MIP1␤. Cell pellet-associated radioactivity was counted using an LKB-Wallac computer-controlled, 1272 CliniGamma counter. Binding data were analyzed using GraphPrad Prism 2.0 software.
HIV-1 Infection, Luciferase, and ␤-Galactosidase Assays-Infectious HIV-1 particles were produced by transfecting HEK-293T cells with either nef-deleted R5 HIV-1 HIV-1 JR-CSF-luc , which carries the firefly luciferase (luc) reporter gene instead of nef (a gift from V. Planelles, Rochester, NY), or the WT HIV-1 YU2 , proviral DNAs. For pseudotyped virus, HIV-1 pNL4 -3-luc envelope-deficient (env (Ϫ)) proviral DNA containing a luc reporter gene instead of nef was co-transfected with an HIV-Env-expressing vector, as previously described (52). Culture supernatants containing the infectious HIV particles were harvested 2 days after transfection, cleared of cellular debris by filtration through a 0.45-m pore size filter, and concentrated by centrifugation through filters (Centricon Plus-20, cut-off (100.000 nominal molecular weight limit), Millipore). The amount of viral particles was estimated by measuring the HIV-1 Gagp24 antigen using a capture assay (PerkinElmer Life Sciences). Target cells were infected for 2 h at 37°C with the indicated amounts of Gagp24, then washed, harvested at the indicated times post-infection, and lysed for luciferase or ␤-galactosidase assays. Experiments were performed in triplicate. Luciferase activity was measured using a luminometer (Lumat LB 9501, Berthold), and results were expressed as relative luciferase units per microgram of cell lysates protein. Background signal was subtracted from the value obtained for each sample. ␤-Galactosidase measurement on total cell extracts was performed using a kit according to the manufacturers' instructions (Roche Molecular Biochemicals).
Reagents-MIP1␤ (provided by F. Baleux, Institut Pasteur, Paris) was chemically synthesized by the Merrifield solid method on a fully automated peptide synthesizer (Pioneer, Perspective Biosystems, and PerkinElmer Life Sciences). The detailed procedure used for chemokine synthesis was described previously (53). The concentration of purified chemokine was determined by amino acid analysis on a 6300 Beckman amino acid analyzer after hydrolysis for 20 h in 6 N HCl and 0.2% phenol in the presence of a known amount of norleucine as an internal standard. All chemicals for the synthesis were purchased from Perspective Biosystems and PerkinElmer Life Sciences. Bafilomycin A1 (Sigma-Aldrich), ALLN (Sigma-Aldrich), or Z-LLL-H (provided by F. Baleux and previously described (51)) were used at 200 nM, 25 M, and 40 M, respectively. 2-Bromopalmitate (Sigma-Aldrich) was added overnight in 100 M in culture medium.

Identification of Palmitoylation Sites of CCR5 Receptor-
Most GPCRs have conserved cysteine residues in their carboxyl-terminal cytoplasmic domain that appear in many cases to be covalently attached to palmitate (27). Although palmitoylation of chemokine receptors has not been reported, the presence of a cluster of cysteine residues (Cys-321, Cys-323, and Cys-324) located at the boundary of the cytoplasm with the seventh transmembrane domain of CCR5 (Fig. 1A) raises the possibility that these residues are palmitoylated and influence receptor functions.
To test this prediction, HEK-293T cells lacking endogenous CCR5 expression were transiently transfected with a human CCR5 cDNA and metabolically labeled with [ 14 C]palmitate. Thereafter, CCR5 was immunoprecipitated from cell lysates by a monoclonal anti-CCR5 antibody (2D7), which recognizes an epitope in the second extracellular domain of the receptor (19). Electrophoretic separation of immunoprecipitated material allowed identification of a single band that migrates with the mobility and apparent molecular weight of CCR5 ( Fig. 2A, lane  1). Western blot analysis of the same samples confirmed the identity of the protein visualized in radioactive gels (Fig. 3A).
Following 3 h of [ 14 C]palmitate labeling, single Cys to Ala mutants showed reduced incorporation of palmitate ( Fig. 2A,  upper panel, lanes 3-5) compared with the wild-type (WT) molecule, thus indicating that all three cysteine residues are modified by palmitoylation. The contribution of each cysteine to the bulk of palmitoylation shown by CCR5 was analyzed using parallel incorporation of [ 35 S]Met by each CCR5 derivative for the same length of time ( Fig. 2A, lower panel) to normalize the corresponding values of palmitate labeling (Fig. 2B). The preponderant role of a single cysteine in the palmitoylation of a GPCR has been described for the rhodopsin receptor, where modification of the first cysteine of the cytoplasmic tail was proposed to be a prerequisite for subsequent palmitoylation of the downstream cysteine residue (54). However, in the case of CCR5, based on the slight differences observed between the reduction of CCR5 palmitoylation upon mutation of each one of the three cysteines (50%, 35%, and 30% for Cys 1, Cys 2, and Cys 3 mutants, respectively), a predominant role cannot be attributed conclusively to the first cysteine residue. All mutants carrying combined substitutions of two cysteines (CCR5 Cys 1,2; CCR5 Cys 1,3; and CCR5 Cys 2,3) exhibited increased reduction of palmitoylation compared with single Cys to Ala mutants. Replacement of the three cysteines (CCR5 Cys 1,2,3) led to a profound decrease of palmitate incorporation (95% inhibition compared with WT labeling), thus indicating that the carboxyl-terminal Cys-321, Cys-323, and Cys-324 are the residues modified by palmitoylation in CCR5. The residual palmitoylation observed in CCR5 Cys 1,2,3 indicates that cytoplasmic Cys-58 and Cys-224 located in the first and third intracellular loops contribute little, if at all, to the bulk of CCR5 palmitoylation. 35 Slabeled, immunoprecipitated proteins revealed that introduction of a single mutation of the first cysteine (CCR5 Cys 1) or combined mutations (CCR5 Cys 1,2; Cys 1,3; and Cys 1,2,3), resulted in a decreased recovery of newly synthesized CCR5 ( Fig. 2A, lower panel). The markedly reduced expression of CCR5 might be attributed to either the lack of palmitoylation per se or to structural changes introduced by Cys to Ala substitutions. Experimental evidence in support of the first possibility is as follows.

Lack of CCR5 Palmitoylation Prevents Accumulation and Cell Surface Expression of the Receptor-Comparison of
First, CCR5 [ 14 C]palmitate incorporation was shown to be sensitive to 2-bromopalmitate (2BP) (Fig. 2A, upper panel, lane  10), a palmitic acid analog that acts as an inhibitor of fatty protein acylation with specificity for palmitoylation (55). Moreover, we show that overnight treatment with 2BP precluded accumulation of newly synthesized CCR5 ( Fig. 2A, lower panel,  compare lane 10 with lane 1). It follows that 2BP treatment reduced the steady-state level of CCR5 to that of the triple Cys to Ala CCR5 mutant (Fig. 3A, compare lanes 12 and 11 to lane  10), whereas cell viability and the amount of the constitutively expressed lactate dehydrogenase (LDH) protein remained unaffected (Fig. 3A). In keeping with the capacity of 2BP to prevent [ 14 C]palmitate incorporation, incubation of CD4-positive T lymphoblastoid cells with 2BP abolished palmitoylation of CD4, a receptor known to be palmitoylated (56), without modifying its cell surface expression (data not shown). Collectively, these results showed that metabolic inhibition of palmitoylation of CCR5 and combined mutation Cys to Ala of the three carboxyl-terminal cysteine residues have similar effects on the biology of the receptor. Thus, it can be proposed that reduced expression of non-palmitoylated CCR5 ( Fig. 2A, lower  panel, lanes 10 and 9, and Fig. 3A, lanes 12 and 11) is specifically related to a lack of palmitoylation of the receptor's carboxyl-terminal tail.
Second, when considered as a function of palmitoylation, the decreased steady-state levels of single-, double-, and triple-CCR5 mutants correlated significantly (r 2 ϭ 0.9472, p ϭ 0.014) with their respective incorporation of labeled palmitate (Fig.  3B). This is in keeping with the fact that the combined mutation of the three Cys to Ser residues (CCR5 Cys 1,2,3 Ser) led to a degree of accumulation intermediate between the WT and palmitoylation-deficient CCR5 (CCR5 Cys 1,2,3) counterparts (compare in: Fig. 2A, lower panel, lanes 1 and 9 with lane 2, and in Fig. 3A, lanes 1 and 2 with lane 3). This effect could be explained by the reduced, yet significant, palmitate incorporation into the CCR5 Cys 1,2,3 Ser mutant (40% of WT CCR5, Fig. 2B and Fig. 2A, upper panel, lane 2). The ability of a serine residue substituted for a cysteine to serve as an alternative palmitoylation site in membrane-associated proteins has been documented for the human transferrin receptor (57) and proposed as an alternative palmitoylation site for the transmembrane Lymphoma Proprotein Convertase (58). However, the mechanism of this acylation via an ester linkage is unknown.
Thus, it appears that, in the case of CCR5, the lesser but significant ability of the three serines CCR5 mutant to bind palmitate partially mimics the effect of natural palmitoylation of the receptor.
Collectively, these results suggest that CCR5 carboxyl-terminal palmitoylation is a key post-translational event that controls the proper expression of the receptor. Indeed, singleand double-cysteine CCR5 mutants exhibited altered cell surface expression (Fig. 3C) compared with WT CCR5, although the reduction was more dramatic for the triple-Cys to Ala mutant. In support of these results, the cell surface expression of single-, double-, and triple-CCR5 mutants could be significantly correlated (r 2 ϭ 0.8651, p ϭ 0.01671) to the extent of the relative [ 14 C]palmitate labeling of these receptors (Fig. 3D). Interestingly, and in keeping with its degree of palmitoylation, the cell surface expression of the CCR5 Cys 1,2,3 Ser mutant was found to be in an intermediate range between Cys 1-associated single (Cys 1) and double (Cys 1,2 and Cys 1,3) CCR5 mutants (Fig. 3C).
Taking into account the altered cell surface expression of palmitoylation-deficient CCR5 Cys 1,2,3, the subcellular localization and distribution of this molecule was investigated. Immunofluorescence microscopic analysis revealed that WT CCR5 was detected predominantly at the plasma membrane (Fig. 4, left panel). Moreover, cholera toxin B-subunit (CTx) detection of ganglioside GM1, widely used as a marker of plasma raft microdomains (59), showed that CCR5 and GM1 partially co-localized (Fig. 4, left panel, overlay) suggesting that ectopic WT CCR5 is associated with rafts. Whether or not this situation reflects the sub-membrane localization of CCR5 in primary lymphocytes remains unknown.
In contrast, the palmitoylation-deficient mutant (CCR5 Cys 1,2,3) exhibited no apparent cell surface expression but, instead, displayed a punctuate cytoplasmic staining (Fig. 4, right  panel). This abnormal distribution was also observed when WT CCR5-expressing cells were treated with 2BP (data not shown). This suggests that interfering with CCR5 palmitoylation, either by combined mutation of the three carboxyl-terminal cysteine residues or by metabolic inhibition of palmitate incorporation, promotes reduced cell surface expression of the receptor. This phenomenon, which is associated with a decrease in steady-state expression (Fig. 3A, lanes 11 and 12), supports the conclusion that a palmitoylation-deficient CCR5 molecule is prone to instability.
Palmitoylation-defective CCR5 Shows a Rapid Turnover and Reduced Half-life-The half-life of WT and palmitoylation-deficient CCR5 receptors were assayed in transiently transfected HeLa cells pulse-chased with [ 35 S]Met (Fig. 5). Analysis of immunoprecipitated cell lysates revealed that the half-life of the CCR5 mutant was dramatically shortened (t1 ⁄2 ϭ 1.5 h) compared with the WT counterpart (t1 ⁄2 ϭ 4.5 h). This result, together with the observation that 2BP treatment affected WT receptor turnover to the same extent as mutation of the three cysteine residues (data not shown), suggests that palmitoylation of CCR5 carboxyl-terminal tail prevents the receptor from entering a degradation pathway and/or protects it from accelerated proteolysis.
The contribution of the palmitoylation to GPCRs stability has previously been observed for the adenosine receptor A1 (60) and suggested for the human thyrotropin receptor (45). It was also shown that palmitoylation of the cation-dependent mannose 6-phosphate receptor prevents this protein from entering lysosomes during its trafficking between endosomes and the plasma membrane (61,62). The spacing between the TM domain and the lipid anchor site was proposed to contribute to receptor degradation (63). A recent study demonstrated that the proteasome is involved in the degradation of the fraction of newly synthesized human delta opioid receptor (64) that is retained intracellularly (65).
The involvement of intracellular proteolytic systems, such as calpains, lysosomal proteases, and the ATP-proteasome-dependent pathway, in the regulation of the short-lived, non-palmitoylated CCR5 mutant protein was investigated. Peptide-aldehyde Z-LLL-H (carbobenzoxyl-leucinyl-leucinyl-leucinal-H), a highly potent and selective inhibitor of the proteasome (66), failed to affect CCR5 Cys 1,2,3 levels in transfected HEK-293T cells (Fig.  6A, compared lanes 9 -12 to lanes 5-8). Interestingly, the peptide aldehyde ALLN (N-acetyl-Leu-Leu-norleucinal), a modest proteasome inhibitor (66,67), which in contrast potently inhibits a class of cysteine proteases, the cathepsins, promoted some accumulation of the CCR5 mutant protein over time (Fig. 6A, compared lanes 13-16 to lanes 5-8). Thus, and given the failure of Z-LLL-H to prevent CCR5 degradation, the effects of ALLN might be the consequence of its well-known capacity to inhibit proteases that, like cathepsins, function optimally at acidic pH and are usually sequestered within lysosomes (68,69). To further investigate the possibility that palmitoylation-deficient CCR5 is degraded in lysosomes, CCR5 Cys 1,2,3-expressing cells were treated with the mycelium secondary metabolite, bafilomycin A1, which is known to specifically inhibit vacuolar type H ϩ -ATPase and thereby affects acidic proteases by raising the pH of endocytic organelles, including lysosomes (70,71). As shown in Fig.  6A (lanes 17-20), steady-state amounts of CCR5 mutant protein were greatly increased in the presence of bafilomycin A1, whereas cell viability and the amount of the cytoskeletal protein ␤-tubulin used as control were not significantly affected. Indirect immunofluorescence and confocal microscopic analysis of cells permeabilized with saponin corroborated these results by showing that colocalization of the CCR5 Cys 1,2,3 mutant with the integral lysosomal membrane protein Lamp1 was enhanced by treatment with either ALLN or bafilomycin A1 (Fig. 6B). Based on these results, we propose that palmitoylation of carboxylterminal cysteines diverts and/or protects CCR5 from entering a degradation pathway involving acidic proteases in the lysosomal compartment.
The relevance of these findings to the natural catabolism of the CCR5 receptor is suggested by the observation that both ALLN and bafilomycin A1 promoted a slight, but significant, accumulation of WT CCR5 molecules ( Fig. 6 A, compare lanes  23 and 24 with lanes 21 and 22). The dynamic and reversible nature of palmitoylation is widely accepted (32), and turnover of this lipid modification on CCR5 could imply that palmitoylated and depalmitoylated forms of CCR5 co-exist. Thus, the depalmitoylated WT CCR5 form could enter a proteolytic pathway. However, the mechanism underlying this degradation remains to be investigated. It is possible that palmitoylation reduces one or more of the many steps in the natural degradation pathway of membrane proteins, including conveyance to intracellular endocytic organelles or sensitivity to proteases. An intriguing possibility would also be that palmitoylation stabilizes CCR5 membrane expression through proper folding and/or interaction of the carboxyl terminus with other CCR5 molecules or specific chaperones that escort receptors to the cell surface.
Non-palmitoylated CCR5 MIP1␤ Binding Properties and HIV-1 Co-receptor Functions-To address these issues, HEK-293T cells were transiently transfected with either WT or palmitoylation-deficient Cys 1,2,3 CCR5 receptors. Binding assays were performed under equilibrium conditions, in which radioiodinated MIP1␤, a high affinity ligand (8,72,73), was displaced by increasing concentrations of unlabeled MIP1␤. Interestingly, MIP1␤ bound strongly and specifically to cells that expressed the Cys 1,2,3 mutant. Quantitative analysis of the displacement of receptor-bound [ 125 I]MIP1␤ by increasing concentrations of unlabeled MIP1␤ allowed measuring the pK i and B max values. The affinities of MIP1␤ for the WT and mutant CCR5 were found to be similar (pK i : 7.30 Ϯ 0.27 and 7.54 Ϯ 0.19, respectively) (Fig. 7). Moreover, Scatchard analy- Lysates were subjected to SDS-PAGE followed by immunoblotting using CCR5 (MC-5) and ␤-tubulin antibodies. The NF-B inhibitor, IB␣, was found to be protected from its well-known inducible degradation (51) by Z-LLL-H at the concentrations and incubation times used (results not shown). Molecular mass markers are indicated (kDa). B, immunofluorescent localization of CCR5 and Lamp1 in HeLa cells expressing Cys 1,2,3 and WT CCR5. Cells were left untreated (CT) or were incubated for 6 h with ALLN or bafilomycin A1. CCR5 and Lamp1 staining were detected after cell permeabilization using FITC-labeled CCR5 2D7 and monoclonal antibodies to Lamp-1, followed by staining with anti-mouse Texas Red-labeled secondary antibody. Merging of the signals is shown in overlay panels. face, represents functional chemokine receptors. This result implies that cytoplasmic tail palmitoylation is not required for high affinity binding of MIP1␤ to CCR5.
HeLa cells were transiently co-transfected with CD4 and either WT or CCR5 Cys 1,2,3 molecules and tested for their ability to support entry of recombinant luciferase reporter HIV-1 virus pseudotyped with the R5 strain BaL-Env glycoprotein. Palmitoylation-deficient CCR5 was impaired in its ability to support a single round of virus replication (Fig. 8A). This handicap was confirmed for replication of either HIV-1 luciferase reporter JR-CSF-luc (Fig. 8B) or WT HIV-1 YU2 (Fig. 8C) viruses. These results suggest that the reduced HIV co-receptor function of palmitoylation-deficient CCR5 is secondary to reduction of its cell surface expression. This effect is consistent with the observation that changes in cell surface concentration of CCR5 may control HIV infection (74).
Interestingly, among the polymorphisms found in the CCR5 coding region, a single-nucleotide deletion (75) results in a receptor that lacks the entire cytoplasmic tail and is greatly impaired in its cell surface expression and HIV-1 co-receptor activity (76). Altered cell surface trafficking was proposed for the mutant molecule. However, the abnormal protein was still expressed at the cell surface, and its cytoplasmic levels were comparable to that of the WT CCR5 molecule (76). In this study (76) the massive expression of CCR5 from the potent Sendai virus or vaccinia-derived recombinant expression vectors might overcome supposed mechanism of receptor proteolysis, which opposes to cell surface expression. This assumption is supported by our observation that increasing the CCR5 mutant DNA/cell ratio favored cell surface expression of the receptor (Fig. 8B, and legend).
Our data demonstrate that the fraction of palmitoylationdeficient CCR5 that escaped rapid degradation and was expressed at the cell membrane could be used by HIV as a functional co-receptor. As shown in Fig. 8C, MIP1␤ displayed the strongest anti-HIV-1 activities on mutant compared with WT CCR5 co-receptor function (90% and 35% inhibition, respectively). This is likely to merely reflect the reduced number of CCR5-mutated receptors available on the membrane for MIP1␤ binding. However the saturating concentration of the chemokine used in the infection assay should overcome differences related to the amount of receptor at the cell surface, thus making this possibility unlikely. Alternatively, the efficient HIV-1 inhibition promoted by the chemokine in CCR5-Cys 1,2,3-expressing cells may result from agonist-induced, sustained down-regulation of the receptor. Indeed, our results suggests that, in opposition to the endocytosis and subsequent recycling, which occurs upon binding of WT chemokine receptors to natural agonists (53,77,78), the internalized CCR5 mutant may be directed, as a consequence of its lack of palmitoylation, to a degradation pathway thus precluding re-expression of the receptor at the cell surface.
Conclusion-Collectively, the results presented in this study indicate that CCR5 is palmitoylated at its carboxyl-terminal tail and that interfering with this normal lipid modification promotes rapid proteolytic degradation of the molecule and decreases proper cell surface expression of this receptor. Once expressed at the cell surface, the palmitoylation-defective mutant is fully active as a receptor for both HIV Env and for its ligand MIP1␤. While completing this manuscript, we became aware of another study (C. Blanpain, et al. (79)) also reporting an influence of palmitoylation on membrane expression of CCR5. The altered signaling activity observed by these authors, who also transfected palmitoylation-deficient CCR5, might result from decreased membrane expression of the receptor, although a direct influence of palmitoylation on CCR5-mediated signal transduction can also be envisaged. In any case, the finding that membrane expression of CCR5 is controlled by palmitoylation-dependent degradation of the molecule represents a novel mechanism of chemokine receptor regulation and is likely to be relevant to the physiological and pathological phenomena involving expression of this important receptor.