Enhanced Expression, Native Purification, and Characterization of CCR5, a Principal HIV-1 Coreceptor*

Seven-transmembrane segment, G protein-coupled receptors (GPCRs) play important roles in many biological processes in which pharmaceutical intervention may be useful. High level expression and native purification of GPCRs are important steps in the biochemical and structural characterization of these molecules. Here, we describe enhanced mammalian cell expression and purification of a codon-optimized variant of the chemokine receptor CCR5, a GPCR that plays a central role in the entry of the human immunodeficiency virus-1 (HIV-1) into immune cells. CCR5 could be solubilized in its native state as determined by its ability to be precipitated by 2D7, a conformation-dependent anti-CCR5 antibody, and by the HIV-1 gp120 envelope glycoprotein. The 2D7 antibody recognized immature and mature forms of CCR5 equally, whereas gp120 preferentially recognized the mature form, a result that underscores a role for posttranslational modification of CCR5 in its HIV-1 coreceptor function. The methods described herein contribute to the analysis of CCR5 and are likely to be applicable to many other GPCRs.

The entry of human immunodeficiency virus-1 (HIV-1) 1 into host cells usually requires the sequential interaction of the gp120 exterior envelope glycoprotein with the CD4 glycoprotein and a chemokine receptor on the cell membrane (1,2). CD4 binding facilitates virus attachment to the cell surface and mediates conformational changes in gp120 that allow a high affinity interaction with the chemokine receptor (3,4). Chemokine receptor binding is believed to trigger further conformational changes in the viral envelope glycoproteins that allow the gp41 transmembrane envelope glycoprotein to fuse the viral and host cell membranes (5,6).
Chemokine receptors are members of the large family of seven-transmembrane segment, G protein-coupled receptors (GPCRs). The ␤-chemokine receptor CCR5 is the principal HIV-1 coreceptor used during natural infection (7)(8)(9)(10)(11). Individuals with genetic defects in CCR5 expression are relatively resistant to HIV-1 infection (12,13). Some HIV-1 isolates can be adapted in tissue culture to replicate on cells lacking CD4 (14,16), 2 but binding to either CCR5 or to the other common HIV-1 coreceptor, CXCR4 (17), is essential for the entry of these viruses. The necessity of the HIV-1 gp120-chemokine receptor interaction to virus replication makes an understanding of the structural basis of this binding a high priority. Structures of unbound CD4 (18 -20), an HIV-1 gp120 derivative complexed with CD4 (21,22), and segments of the HIV-1 gp41 ectodomain (23)(24)(25) have been resolved by x-ray crystallography. Resolution of the structure of CCR5, either alone or in a complex with HIV-1 gp120, would provide information important for guiding attempts at intervention.
The GPCRs play central roles in a wide variety of physiological, neurological, and immunological processes and represent major targets for current pharmaceutical therapies (26 -28). Nonetheless, structural information on this large family of proteins is very sparse. Generally low levels of expression and the dependence of the native conformation of these proteins on the hydrophobic, intramembrane environment have complicated attempts to study GPCR structure. With the exception of the light-sensitive opsins found in certain bacteria and in the retinae of higher organisms, most GPCRs are naturally expressed at low levels (29,30). No system has reproducibly resulted in levels of protein expression suitable for the purification of most GPCRs (31). Bacterial, yeast, or insect cell expression of GPCRs can result in protein misfolding, aggregation, and heterogeneity (32). Furthermore, some GPCRs, such as CCR5, require posttranslational modification, which may exhibit significant differences in nonmammalian cell types, for efficient function (33). Irreversible protein denaturation occurring during the solubilization of the cell membrane, a process required for purification of the GPCR, represents another obstacle to biochemical, biophysical, and structural studies.
Here, we report the development of a system for efficient expression of CCR5, and the identification of an approach that allows solubilization and purification of CCR5 in its native state. We demonstrate a cell-free association between CCR5 and the HIV-1 gp120 glycoprotein. These methods should facilitate the identification of small molecules that block the HIV-1-CCR5 association, the determination of CCR5 structure, and the study of other GPCRs.

Construction and Expression of Codon-optimized CCR5 (synCCR5)-
The analysis of codon usage for 45 GPCRs representing different pro-tein subfamilies was performed with GenBank TM data and software developed by the University of Wisconsin Genome Sequence Group. The sequence encoding human CCR5 was optimized for mammalian cell codon usage (28), utilizing the following codons: alanine (GCC), arginine (CGC), asparagine (AAC), aspartic acid (GAC), cysteine (TGC), glutamic acid (GAG), glutamine (CAG), glycine (GGC), histidine (CAC), isoleucine (ATC), leucine (CTG), lysine (AAG), methionine (ATG), phenylalanine (TTC), proline (CCC), serine (TCC), threonine (ACC), tryptophan (TGG), tyrosine (TAC), and valine (GTG). The 5Ј and 3Ј sequences flanking the CCR5 coding sequence were modified. Following restriction sites for EcoRV, EcoRI and HindIII, the Kozak consensus (GCCGCCACCATGG) was placed immediately 5Ј to the CCR5 reading frame. A sequence encoding a single glycine residue followed by the bovine rhodopsin C9 peptide tag (TETSQVAPA) was introduced immediately 5Ј to the natural stop codon of CCR5. At the 3Ј end of the epitope-tagged CCR5 gene, XbaI, SalI, and NotI restriction sites were introduced. Analogous constructs were made for the wild-type human CCR5 gene and the bovine rhodopsin gene, except that the codons were not altered and, in the latter case, the C-terminal C9 sequence was naturally present.
A total of 34 oligonucleotides, each approximately 70 nucleotides in length, corresponding to the complete sense and antisense strands of the synCCR5 gene and flanking sequences, were constructed so that approximately 50% of their sequences were complementary to those of each of the two complementary oligonucleotides from the opposite strand. Oligonucleotides were deprotected in pure ammonium hydroxide at 65°C for 4 h, after which the ammonium hydroxide was evaporated, and the oligonucleotides were dissolved in water at a final concentration of 2 nM. For gene synthesis, the 34 oligonucleotides were separated into five groups (6 or 8 oligonucleotides per group) and 25 cycles of polymerase chain reaction were performed using Pfu polymerase (Stratagene, La Jolla, CA) and a 3-fold molar excess of the 5Ј and 3Ј terminal oligonucleotides in each group. This step generated five small segments of the synCCR5 gene with complementary and overlapping ends. Equal amounts of each polymerase chain reaction product were combined with a 3-fold molar excess of the 5Ј and 3Ј terminal oligonucleotides of the complete synCCR5 sequence. A second round of 25 cycles of polymerase chain reaction yielded the complete synCCR5 sequence. The product was sequenced to ensure that the sequence was correct.
The synCCR5, wild-type CCR5, and bovine rhodopsin sequences were cloned into the following vectors: PMT4 (a gift from Dr. Reeves, Massachusetts Institute of Technology), PACH (a gift from Dr. Velan, Israel Institute for Biological Research), pcDNA 3.1(ϩ) and pcDNA4/HisMax (Invitrogen), and PND (a gift from Dr. Rhodes, University of California, Davis). After cloning of the synCCR5 gene into the pcDNA4/HisMax vector, the sequence encoding the N-terminal HisMax region was removed by QuikChange mutagenesis (Stratagene). Different cell lines were transfected with the synCCR5 and wild-type CCR5 genes using the GenePorter transfection reagent (San Diego, CA). Following transfection, cells expressing CCR5 were selected with 0.8 mg/ml of neomycin (G418). Cells expressing the highest surface levels of CCR5 were selected by FACS after staining cells with the R-phycoerythrin-conjugated anti-CCR5 antibody 2D7-PE (Pharmingen, San Diego, CA). Among all tested cells (canine thymocytes Cf2Th, human embryonic kidney cells HEK-293T, COS-1, and HeLa (American Type Culture Collection)), the highest CCR5 expression levels were observed in Cf2Th and HEK-293T cells transfected with synCCR5 gene in the PACH vector. The highest synCCR5-expressing clones were selected by FACS from a total of 76 clones of Cf2Th cells and 62 clones of HEK-293T cells.
Radiolabeling and Immunoprecipitation of CCR5-Approximately 4 ϫ 10 6 CCR5-expressing Cf2Th or HEK-293T cells grown to full confluency in 100-mm dishes were washed twice in PBS and starved for 1 h at 37°C in Dulbecco's modified Eagle's medium without cysteine and methionine (Sigma) or in sulfate-free media (ICN, Costa Mesa, CA). The lysate was incubated at 4°C for 30 min on a rocking platform, and cell debris was removed by centrifugation at 14,000 ϫ g for 30 min. CCR5 was precipitated with 20 l of 1D4-Sepharose beads (38) overnight, after which the beads were washed six times in the solubilization medium and pelleted. An equal volume of 2ϫ SDS-sample buffer was added to the beads, following by resuspension and incubation for 1 h at 55°C. Samples were run on 11% SDS-polyacrylamide minigels, which were visualized by autoradiography or analyzed on a Molecular Dynamics PhosphorImager SI (Sunnyvale, CA).
Purification of CCR5-Stable Cf2Th/PACH/synCCR5 cells grown to full confluency in a 150-mm dish were incubated with medium containing 4 mM sodium butyrate for 40 h, washed in PBS, detached by treatment with 5 mM EDTA/PBS, pelleted, and again washed in PBS. Cells were solubilized for 30 min with 3 ml of the solubilization medium containing Cymal TM -5 and centrifuged for 30 min at 14,000 ϫ g. The cell lysate was incubated with 50 l of 1D4-Sepharose beads on a rocking platform at 4°C for 10 -12 h. The Sepharose beads were washed five times with the washing buffer (100 mM (NH 4 ) 2 SO 4 , 20 mM Tris-HCl (pH 7.5), 10% glycerol, and 1% Cymal TM -5) and once with washing buffer plus 500 mM MgCl 2 . CCR5 was eluted from the beads by three successive washes with 50 l of medium containing 200 M C9 peptide (TETSQVAPA), 500 mM MgCl 2 , 100 mM (NH 4 ) 2 SO 4 , 20 mM Tris-HCl (pH 7.5), 10% glycerol, and 0.5% Cymal TM -5. The total quantity of harvested CCR5 was estimated by Coomassie Blue staining of an SDS-polyacrylamide gel run with standard quantities of bovine serum albumin.
Binding of HIV-1 gp120 Envelope Glycoproteins to Solubilized CCR5-Approximately 4 ϫ 10 6 Cf2Th/PACH/synCCR5 cells were labeled for 12 h with [ 35 S]Met/Cys and lysed in solubilization buffer containing 1% Cymal TM -5. One ml of cleared cell lysate was incubated with 100 -500 l of the gp120-containing solutions. The unlabeled JR-FL gp120 was produced in Drosophila cells (3), and the ADA and 190/197 R/S gp120 glycoproteins were produced from transiently transfected 293T cells that had been radiolabeled with [ 35 S]Met/Cys overnight. 2 Except in the case of the CD4-independent gp120 variant, 190/197 R/S, the gp120 glycoproteins (2-4 g) were preincubated with sCD4 (2-4 g) in 20 ml of PBS for 1 h at 22°C prior to addition to the CCR5-containing lysates. After 12 h at 4°C, the gp120-CCR5 complexes were precipitated with either the C11 anti-gp120 antibody (kindly provided by Dr. James Robinson, Tulane University Medical School) or with the 1D4 antibody.

RESULTS
Expression of CCR5 in Mammalian Cells-We compared the codon usage for opsins, the only GPCRs that are naturally highly expressed, with the codon usage for 45 other GPCRs representing a spectrum of different GPCR subfamilies. Opsin codons are biased toward those shown to be optimal for efficient translation in mammalian cells (34), whereas other GPCRs, including CCR5, are associated with codons that are more random and, in many cases, inefficiently translated (data not shown). A codon-optimized CCR5 gene was designed, synthesized using the polymerase chain reaction, and transiently expressed in several different cell lines, using five different expression vectors (pcDNA 3.1, PACH, PND, PMT4, and pcDNA4/HisMax). The level of CCR5 expression directed by the codon-optimized gene was 2-5 times that directed by the wildtype CCR5 gene (Fig. 1, A and B, and data not shown). Among the cell lines tested, CCR5 expression was the highest in Cf2Th canine thymocytes (data not shown), so these cells were used to generate stable cell lines. The PACH vector was used to express the codon-optimized gene encoding human CCR5 containing a 9-residue C-terminal epitope tag (the C9 tag) derived from bovine rhodopsin. The presence of the C9 tag allows recognition of the CCR5 protein by the 1D4 antibody (35). CCR5 expression in the stable cell line, designated Cf2Th/PACH/synCCR5, could be enhanced 2-3-fold by treatment of the cells with sodium butyrate (Fig. 1C). Following this treatment, approximately 3-5 g of CCR5 of high purity could be isolated from 10 7 Cf2Th/PACH/ synCCR5 cells, using techniques described below (Fig. 1D).
Precursor and Mature Forms of CCR5-CCR5 synthesis and turnover in Cf2Th cells were studied by pulse-chase analysis (Fig. 2). A precursor of approximately 40 kDa chased into the mature form of CCR5, which migrated as a wide band of approximately 43 kDa. The CCR5 precursor exhibited a half-life of approximately 25 min (Fig. 2A). The half-life of the mature form of CCR5 was 11-14 h, regardless of whether CCR5 expression was directed by the wild-type or codon-optimized CCR5 gene (Fig. 2, B and C). The half-lives of the precursor and mature forms of CCR5 in HEK-293 cells were similar to those in Cf2Th cells (data not shown). In several different cell lines, a lower molecular mass (approximately 36 kDa) form of CCR5 appeared in parallel with the mature protein (Fig. 2, B and C). This lower molecular mass form of CCR5 was expressed at lower levels than the mature form of CCR5 and has not been completely characterized. Its identity as a CCR5 isoform was confirmed by its precipitation by the 1D4 antibody and the anti-CCR5 antibody 2D7 and by mass spectrometry (Figs. 2 and 3A and data not shown).
Solubilization of Native CCR5-Membrane protein purification requires solubilization of the membranes, typically through the use of detergents. A broad spectrum of conditions was studied to arrive at the composition of the buffer that allowed solubilization and isolation of native CCR5. This opti-mization was guided by a comparison of the amount of solubilized CCR5 capable of being precipitated by the 2D7 antibody, which recognizes a conformation-dependent CCR5 epitope (36), with that able to be precipitated by the 1D4 antibody directed against the linear C9 epitope tag. In this manner, the percentage of solubilized CCR5 remaining in a native conformation could be estimated (Fig. 3A). Eighteen detergents, most of which were designed specifically for the extraction and purification of membrane proteins, were studied. In terms of the quantity of isolated CCR5 protein, as well as the percentage of protein in a conformation able to be recognized by the 2D7 antibody, the most effective detergents were DDM, Cymal TM -5, and Cymal TM -6 ( Fig. 3B). Of these detergents, Cymal TM -5 exhibits the highest critical micelle concentration (2.4 mM), facilitating dialysis of the detergent from the protein solution for the purposes of membrane reconstitution and/or crystallization. We also found that a CCR5 conformation competent for binding HIV-1 gp120 was best preserved in buffers containing Cymal TM -5 (see below). Therefore, Cymal TM -5 was used for further refinement of the CCR5 solubilization/isolation protocol, examining a number of variables (salt composition and concentration, pH, temperature, and minor additives) known to influence the stability of solubilized proteins (37). Ammonium

FIG. 1. Expression of CCR5 at the cell surface (A) or in cell lysates (B) of Cf2Th cells transfected by plasmids expressing wild-type (Wt) and codon-optimized CCR5 (synCCR5 (Syn)).
Forty-eight hours after transfection, cells were suspended by treatment with 5 mM EDTA, washed in PBS, stained with the 2D7-PE anti-CCR5 antibody, and analyzed by FACS (A), or labeled with [ 35 S]Met/Cys, lysed, and precipitated with 1D4-Sepharose (B). C, increased CCR5 expression following sodium butyrate treatment of cells. Cf2Th/PACH/ synCCR5 cells were incubated with sodium butyrate for 40 h, and CCR5 expression was estimated by FACS. D, purified CCR5 isolated from cells. CCR5 was purified from 10 7 Cf2Th/PACH/synCCR5 cells grown in a 150-mm dish, using 1D4-Sepharose. CCR5 was eluted with the C9 peptide, run on an SDS-polyacrylamide gel, and stained with Coomassie Blue. In addition to the monomeric CCR5 bands at 36 -43 kDa, higher order CCR5 forms are evident. The amount of CCR5 in the gel was estimated by comparison with bovine serum albumin standards. The precipitates were analyzed on SDS-polyacrylamide gels. The gel in B was exposed to film for 48 h, and the gel in C was exposed for 19 h. The levels of expression of the 43-kDa (circles) and 36-kDa (squares) CCR5 proteins were quantitated by densitometry.
sulfate and glycerol were found to prolong the existence of a CCR5 conformation capable of being recognized by the 2D7 antibody (data not shown). The optimized CCR5 solubilization buffer was composed of 100 mM (NH 4 ) 2 SO 4 , 20 mM Tris-HCl (pH 7.5), 10% glycerol, and 1% Cymal TM -5.
Binding of Solubilized CCR5 to HIV-1 gp120 -To examine whether the solubilized CCR5 was capable of binding the HIV-1 gp120 glycoprotein, coprecipitation experiments were conducted using three different gp120 glycoproteins. The JR-FL and ADA gp120 glycoproteins, like all characterized wild-type R5 HIV-1 envelope glycoproteins, bind CCR5 efficiently only in the presence of CD4 (3,4). By contrast, the 190/197 R/S variant of the ADA gp120 glycoprotein was derived from a virus adapted in culture to replicate on CD4-negative, CCR5-positive cells, 2 and it therefore binds CCR5 in the absence of CD4. The mature CCR5 protein was precipitated from cell lysates by the C11 anti-gp120 antibody when unlabeled JR-FL gp120 and soluble CD4 (sCD4) were added to the lysates (Fig. 4A), but not when either gp120 or sCD4 were left out of the mixture (data not shown). The C11 antibody precipitated the mature CCR5 protein when radiolabeled 190/197 R/S gp120 was added to the lysates without sCD4 (Fig. 4B). Conversely, the 1D4 antibody precipitated the labeled 190/197 R/S gp120 glycoprotein from these lysates. In the coprecipitation experiments using the C11 anti-gp120 antibody, only the mature form of CCR5, and not the 36-kDa CCR5 isoform, was precipitated (Fig. 4).
Sulfation of CCR5-CCR5 has been shown to be posttranslationally modified by O-linked carbohydrates and by tyrosine sulfation (33). The latter modification has been suggested to facilitate the efficiency with which CCR5 is utilized as an HIV-1 coreceptor (33). To examine the relationship of sulfation to the observed conversion of the CCR5 precursor to the higher molecular mass, mature form of the protein, a pulse-chase analysis of Cf2Th/PACH/synCCR5 cells labeled with either [ 35 S]cysteine/methionine or [ 35 S]sulfate was performed. Sulfate label was incorporated only into the mature form of CCR5 (Fig. 5).
The recognition of the precursor and mature, sulfated forms glycerol) supplemented with a protease inhibitor mixture and 1% (w/v) of different detergents. After 30 min of solubilization and 30 min of centrifugation, the cleared cell lysates were separated into two equal portions. One portion was precipitated with 2D7 (a conformation-dependent antibody against CCR5) and other portion with 1D4 (an antibody that recognizes the linear C9 epitope tag). The precipitates were run on SDS-polyacrylamide gels, and two parameters were examined: 1) the total quantity of CCR5 precipitated by the 1D4 antibody, and 2) the ratio of CCR5 precipitated by the 2D7 antibody relative to that precipitated by the 1D4 antibody. A, precipitates from cells lysed in buffer containing Cymal TM -5, DHPC and Fos-choline ™ -14. Similar levels of CCR5 were precipitated by the 1D4 antibody from these lysates, but the percentage of conformationally intact CCR5 varied (98% in Cymal TM -5, 10% in DHPC, and 13% in Fos-Choline ™ -14). The sample run in the right-hand lane (asterisk) was the same as in the lane labeled 1D4 but was boiled prior to running on the gel, a procedure that results in the formation of high molecular weight multimers of CCR5. B, the amounts of CCR5 precipitated by the 1D4 (E) and 2D7 (q) antibodies from cell lysates containing different detergents, over a range of pH values. The JR-FL gp120-sCD4 complex was formed by incubation of 2 g of gp120 with 4 g of sCD4 in 20 ml of PBS for 1 h at 22°C. The mixtures were then precipitated either with the C11 anti-gp120 antibody (as in A) or with the 1D4 antibody directed against the C9 epitope tag (as in B). Some of the lanes (as indicated) were exposed to film for 12 h, but they are otherwise identical to the lanes shown with the matching antibodies that were exposed to film for 30 min. Under these labeling conditions, the mature 43-kDa form and the 36-kDa form of CCR5, but not the 40-kDa CCR5 precursor, are labeled. of CCR5 by the 1D4 and 2D7 antibodies and by gp120-sCD4 complexes was examined (Fig. 5). The conformation-dependent 2D7 antibody precipitated the CCR5 precursor at an efficiency approximately 20% lower than that seen for the 1D4 antibody. Complexes of the ADA gp120 glycoprotein, sCD4, and the C11 antibody preferentially precipitated the mature, sulfated form of CCR5; the precipitation of the CCR5 precursor by this complex was approximately 5-fold less efficient than that of the mature CCR5 protein (Fig. 5). Similarly, the mature CCR5 protein was recognized more efficiently than the CCR5 precursor by a complex of the C11 antibody and the 190/197 R/S gp120 glycoprotein, in the absence of sCD4 (data not shown).

DISCUSSION
GPCRs serve as therapeutic targets for over 30% of the pharmaceutical agents currently used in clinical practice (26). Specific GPCRs, most of which are chemokine receptors, are necessary cofactors for HIV-1 entry. Progress toward intervening in virus-coreceptor interaction would be greatly expedited by the development of screening assays in which relevant ligands bind purified, native chemokine receptors. We provide herein an experimental approach to the purification of human CCR5 that yields a native protein capable of binding the HIV-1 gp120 envelope glycoprotein and a conformation-dependent antibody. The approach should be generally applicable to other GPCRs.
Achieving adequate levels of CCR5 in an expression system is essential for attempts to establish screening assays and for structural analyses. Codon optimization, empirical screening of plasmids and cell types, and FACS screening yielded stable lines expressing a level of CCR5 approaching that achieved for bovine rhodopsin. In studies not shown, we compared the synthesis, steady-state levels, and turnover of the codon-optimized CCR5 with those of bovine rhodopsin transiently expressed in the same cell type by the same vector and detected using identical epitope tags. The results indicated that the synthesis and steadystate levels of rhodopsin were approximately 5-fold higher than those of CCR5, whereas the half-lives of the two proteins were comparable. This difference was apparently compensated for by the establishment and selection of the stable Cf2Th/PACH/ synCCR5 cell lines. Scale-up of these cells could allow the purification of milligram quantities of the CCR5 protein.
We evaluated the conformational integrity of CCR5 solubi-lized in several different detergents, including DDM, which has been employed in the study of some other GPCRs (38,39). We show that Cymal TM -5 was at least as effective as DDM in solubilizing CCR5 in a native conformation. Human CCR5 retained a native conformation in Cymal TM -5-containing buffers for a few days at 4°C, and bound HIV-1 gp120 more efficiently than CCR5 solubilized in DDM (data not shown). A further advantage of Cymal TM -5 is its critical micelle concentration, which is 14 times higher than that of DDM, the mild detergent commonly used for membrane protein biochemistry. The high critical micelle concentration of Cymal TM -5 should facilitate its removal during reconstitution of the GPCR into membranes or for crystallization. The C-terminal C9 peptide provides an easy and rapid means for purification of CCR5 using the anti-peptide antibody 1D4. The presence of the C9 epitope tag does not affect the ability of CCR5 to function as an HIV-1 entry cofactor or as a receptor for chemokine ligands (data not shown). The purified CCR5 can be gently eluted from the 1D4 antibody by using the free C9 peptide as a competitor (35,38).
The synthesis and turnover of CCR5 were examined in the course of this study. Approximately 80% of the 40-kDa CCR5 precursor is competent for binding the conformation-dependent 2D7 antibody, indicating that a native structure is achieved rapidly after synthesis of the protein. The 40-kDa CCR5 precursor is rapidly converted into a more slowly migrating form. This 43-kDa form of CCR5 is evident within 10 min of labeling the CCR5 precursor. Most of this shift in molecular mass derives from the addition of O-linked carbohydrates to the CCR5 protein (27). The single potential N-linked glycosylation site on CCR5 is not utilized (27). Once synthesized, the mature CCR5 protein exhibits a half-life of 11-14 h. The tyrosines in the N terminus of the 43-kDa CCR5 form are modified by sulfation, suggesting that sulfation occurs after O-glycosylation. The incorporation of the sulfate label into the 43-kDa CCR5 protein continuously increases over a 4-h labeling period. Because tyrosine sulfation typically occurs in the trans-Golgi network, late in the secretory pathway (15), some sulfate addition may be occurring on CCR5 molecules that have been recycled from the cell surface back to the Golgi. Our results indicate that the HIV-1 gp120 glycoprotein preferentially recognizes the mature, sulfated form of CCR5, consistent with the proposed role of sulfated tyrosines in gp120 binding and HIV-1 entry (33).
The availability of methods to purify adequate amounts of CCR5 should expedite progress on understanding the structure and function of this key HIV-1 coreceptor and facilitate the search for effective inhibitors of virus-coreceptor interactions. ), or ADA gp120-sCD4 complexes and the C11 anti-gp120 antibody (lane gp120 ϩ sCD4 ϩ C11) (exposure time, 60 h). The ADA gp120 glycoprotein in the latter complex was metabolically radiolabeled.