HIV-1 entry and macrophage inflammatory protein-1beta-mediated signaling are independent functions of the chemokine receptor CCR5.

The human immunodeficiency virus type 1 (HIV-1) requires the presence of specific chemokine receptors in addition to CD4 to enter its target cell. The chemokine receptor CCR5 is used by macrophage-tropic strains of HIV-1, which predominate during the asymptomatic stages of infection. Here we investigate whether the ability of CCR5 to signal in response to its β-chemokine ligands is necessary or sufficient for viral entry. Three CCR5 mutants with little or no ability to mobilize calcium in response to macrophage inflammatory protein-1β could nonetheless support HIV-1 entry and the early steps in the virus life cycle with efficiencies comparable with those of wild-type CCR5. Conversely, a chimeric receptor with the N terminus of CCR2 replacing that of CCR5 responded to macrophage inflammatory protein-1β and MCP-1 but did not efficiently support viral entry. These results demonstrate that chemokine signaling and HIV-1 entry are separable functions of CCR5 and that only viral entry requires the N-terminal domain of CCR5.

Chemokines are a family of small cytokines that share a common structure containing four conserved cysteines, the first two of which are adjacent (C-C or ␤ chemokines) or separated by one intervening residue (CXC or ␣ chemokines) (19). Chemokines are believed to be important in the trafficking of leukocytes in both basal and inflammatory states (20). Chemokine receptors are G-protein-coupled, seven transmembranespanning receptors (21,22). Chemokine ligation of receptor promotes the exchange of GDP for GTP in an associated heterotrimeric G-protein, dissociation of G␣ from the G␤ and G␥ subunits, and numerous downstream effector functions, including phospholipid hydrolysis and calcium mobilization (23). Gprotein ␣ subunits have been grouped in several classes based on sequence similarity and common effector functions (24). Chemokine receptors have been shown to be coupled to members of the G i and the G q families (25)(26)(27). Signaling through G i proteins is inhibited by pertussis toxin, whereas G q signaling is not affected by pertussis toxin (24).
Here we describe mutants of CCR5 that fail to mobilize calcium following chemokine ligation but that bind chemokine and support HIV-1 entry as well as wild-type CCR5. We also characterize a chimeric receptor of CCR2 and CCR5 that binds MIP-1␤ and mobilizes calcium in response to MIP-1␤ and MCP-1, the ligand for CCR2 (28), but fails to support efficient HIV-1 infection. These data demonstrate conclusively that CCR5 coupling to G-proteins is not a requirement for efficient HIV-1 entry. They also show that HIV-1 entry requires portions of the CCR5 receptor not required for MIP-1␤ binding or signaling. CXCR1 (IL8-RA) were cloned in a pcDNA3 vector (9). A pcDNA3 vector expressing FLAG epitope-tagged CCR2, was a generous gift of Dr. Israel Charo (28). The FLAG epitope is DYKDDDDK (FLAG tag, IBI-Kodak) inserted after the N-terminal methionine. Mutagenesis used to create the expressor plasmids for the D76N, R126N and D125N/R126N mutants was performed on CCR5 in a pcDNA3 vector using the QuikChange method of Stratagene, Inc., according to manufacturer's instructions. The 2M5 chimera was constructed by substituting the DNA encoding the epitope-tagged CCR2 N terminus for the corresponding section of the CCR5 gene in the pcDNA3 plasmid, using the common Msc-1 site as a junction.
Cell Lines-CF2Th canine thymocytes (ATCC CRL 1430), Bing (ATCC CRL 11554), and HEK293 cells were obtained from American Type Culture Collection. Hela-CD4 cells were obtained from Dr. Bruce Chesebro through the National Institutes of Health AIDS Research and Reference Reagent Program. Cells were maintained as described previously (9).
Env Complementation Assay-A single round of HIV-1 entry was assayed as described previously (9), except that 25,000 cpm reverse transcriptase activity of the recombinant viruses containing the ADA and YU2 envelope glycoproteins were used per assay, and cells were incubated with virus for 48 h. Briefly, HIV-1 virus with the nef gene replaced by the CAT gene was used to infect cells expressing CD4 and a chemokine receptor. Cells were lysed after infection, and CAT activity was measured, indicating the level of transcription from the integrated HIV-1 genome (5). In parallel to the infection assays, anti-FLAG and anti-CCR5 antibody 5C7 were used to quantify receptor expression by FACS analysis. 5C7 was generated against the CCR5 receptor stablyexpressed on a murine lymphocyte line. 2 Calcium Mobilization-HEK293 cells were transfected by the calcium phosphate method (33) with 30 g of plasmid DNA transiently expressing the chemokine receptors. Cells were suspended in 10 ml of buffer (Hanks' buffered saline solution, 25 mM HEPES, pH 7.2, 0.1% bovine serum albumin) per flask and incubated with 30 g of Fura-2/AM (Molecular Probes, Inc.) for 30 min at 37°C. Cells were then washed twice with phosphate-buffered saline and resuspended in buffer. Calcium flux measurements in response to MIP-1␤ and MCP-1 (R & D Systems) were taken at excitation wavelengths 340 and 380 nm and reported as a ratio of 340/380 nm. In parallel, an anti-CCR5 antibody, 5C7, was used to quantify receptor expression by FACS analysis. Pertussis toxin-treated cells were incubated for 18 h with 10 ng/ml pertussis toxin (CalBiochem).
Chemokine Binding-HEK293 or BING293 cells were transfected by the calcium phosphate method with 30 g of plasmid DNA transiently expressing the chemokine receptors. In some cases, parallel transfections were performed with a ␤-galactosidase expression plasmid to assess transfection efficiency. Roughly 25-30% of the cells were transfected. Cells were resuspended in binding buffer (50 mM HEPES, pH 7.5, 1 mM CaCl 2 , 5 mM MgCl 2 , and 0.5% bovine serum albumin). Approximately 5 ϫ 10 5 cells were mixed with 0.1 nM 125 I-labeled MIP-1␤ (DuPont NEN) and varying concentrations of unlabeled MIP-1␤ (R&D Systems) in a total volume of 100 l. Cells were shaken at 37°C for 30 min, centrifuged, resuspended in 0.6 ml of the same buffer containing 500 mM NaCl, and centrifuged again, and bound ligand was quantitated by liquid scintillation counting. For affinity measurements, nonspecific binding was determined in the presence of 200 nM Mip-1␤ and subtracted from all points.

RESULTS
Calcium Mobilization through CCR5 Mutants-Changes in a conserved aspartic residue in the second transmembrane domain have been shown to block ligand-induced calcium mobilization by several seven transmembrane-spanning receptors (34,35). An analogous CCR5 mutant, D76N, was made. Mutations affecting a highly conserved region of the second intracellular loop have similarly blocked the coupling of other seven membrane-spanning receptors to G-proteins (36 -38), and we made two constructs, R126N and D125N/R126N, that substituted asparagine for conserved residues in this region of CCR5. These mutants were expressed at or near wild-type levels in both HEK293 and CF2Th cells ( Fig. 1 and 2 legends) but failed to mobilize calcium in response to 500 ng/ml MIP-1␤ (Fig. 1A). Wild-type CCR5 responded strongly at 250 and 500 ng/ml ( Fig.  1A and data not shown). When incubated 18 h with 10 ng/ml pertussis toxin, CCR2 and CCR5 expressing HEK293 cells responded to 500 ng/ml MIP-1␤ with 50 -60% of the peak values of the same cells in the absence of pertussis toxin ( Fig.  1C and data not shown). We conclude that D76N, D125N/ R126N, and R126N are expressed at the cell surface but are not coupled to a signaling pathway leading to calcium mobilization. We also conclude that, as previously reported for CCR2 (26), CCR5 can couple to a signaling pathway that is insensitive to pertussis toxin at high chemokine concentrations. were treated with 500 ng/ml MIP-1␤ at the time points indicated by the arrows. Flux is displayed as a ratio of the response at 340 nm to the response at 380 nm excitation wavelength. The average mean fluorescence values of cells stained by anti-CCR5 antibody for CCR5, D76N, R126N, and D125N/R126N were 106 Ϯ 50, 142 Ϯ 19.5, 100 Ϯ 10, and 52 Ϯ 1, respectively. Background staining observed with flouresceinconjugated second antibody only was 3.5 Ϯ 0.1. For some experiments, lower amounts (20 g/flask rather than 30 g/flask) of wild-type CCR5 DNA were used for transfection to obtain expression levels comparable with those of the CCR5 mutants. B, calcium mobilization of 2M5 chimera in response to MIP-1␤ and MCP-1. Shown is a representative response of 2M5 when treated with 500 ng/ml MIP-1␤ or 1 g/ml MCP-1 at the time points indicated with the arrows. C, calcium mobilization of CCR5 in the presence and the absence of pertussis toxin treatment. CCR5-expressing HEK293 cells incubated with or without 10 ng/ml pertussis toxin (PTX) for 18 h before measurements were taken. MIP-1␤ (500 ng/ml) was added at the time points indicated with the arrows.

Chemokine Signaling and HIV-1 Entry on CCR5 6855
Chemokine Response of the 2M5 Chimera-A chimeric molecule, 2M5, was also tested for responsiveness to MIP-1␤ and the CCR2 ligand MCP-1. The chimera was made by replacing the N terminus of CCR5 with that of CCR2, with a junction in the second transmembrane domain. The chimeric molecule responded like wild-type CCR5 to MIP-1␤ and also gave an appreciable calcium flux to MCP-1 at 1 g/ml, whereas wild-type CCR5 responded only to MIP-1␤ and wild-type CCR2 responded only to MCP-1 ( Fig. 1B and data not shown). Thus the 2M5 chimera has retained the binding and signaling specificity of CCR5 and has also acquired the ability to bind to and signal in response to MCP-1.
MIP-1␤ Binding to CCR5 Variants-Each of the CCR5 mutant proteins was tested for its ability to bind MIP-1␤ specifically. Unlabeled MIP-1␤ competed for 125 I-labeled MIP-1␤ binding to cells expressing wild-type and mutant proteins with very similar efficiencies ( Fig. 2A), yielding apparent dissociation constants of 6.8, 6.3, and 4.6 nM for wild-type CCR5, D76N, and D125N/R126N, respectively. The 2M5 chimera also bound MIP-1␤ at an affinity near that of wild-type CCR5 (Fig. 2B) with an apparent dissociation constants of 1.6 and 1.2 nM for the chimeric and wild-type proteins, respectively. We conclude that 2M5, D76N, and D125N/R126N each bind MIP-1␤ with affinities near that of wild-type CCR5.
HIV-1 Entry into Cells Expressing CCR5 Variants-We tested the ability of each of the CCR5 mutants to support HIV-1 entry into Hela-CD4 cells and CF2Th cells. Recombinant viruses containing the YU2 and ADA envelope glycoproteins infected Hela-CD4 cells expressing the D76N mutant at levels comparable with that seen for cells expressing wild-type CCR5. By contrast, both viruses inefficiently infected cells expressing the 2M5 receptor, near the levels seen for cells expressing the control receptor CCR4 (Fig. 3A). On CF2Th cells cotransfected with CD4, each of the signaling defective mutants D76N, D125N/R126N, and R126N supported efficient HIV-1 entry at a level proportionate to their surface expression, as documented by FACS analysis (Fig. 3B and its legend). Thus, the ability to support HIV-1 infection is not significantly impaired in cells expressing D76N, D125N/R126N, or R126N but is impaired in cells expressing the 2M5 chimera. Chemokine Signaling and HIV-1 Entry on CCR5 6856 DISCUSSION In this work we asked whether there is a necessary relationship between the signaling response of the chemokine receptor CCR5 to its natural ligand and the role of CCR5 as a coreceptor for the HIV-1 virus. Although other investigators have attempted to probe this relationship with pertussis toxin (39), a mutagenic approach was necessary because chemokine receptors, in particular the closely related CCR2, have been shown to be coupled to pertussis-insensitive as well as pertussis-sensitive pathways (26). Both sets of pathways are active in the natural target cells of HIV-1 (40). The D76N, D125N/ R126N, and R126N mutants described here are incapable of efficiently mobilizing calcium in response to high levels of chemokine but are expressed well and bind MIP-1␤ with affinities close to that of wild-type CCR5. The fact that these mutants support HIV-1 entry excludes an obligate role for calcium mobilization and its sequelae in promoting viral entry.
The results with the 2M5 chimera demonstrate that the binding site for MIP-1␤ is distinct from that used by HIV-1 entry and that binding of and efficient signaling through MIP-1␤ does not ensure a receptor that supports HIV-1 entry. A second property of this chimera, the ability to signal in response to the CCR2 ligand MCP-1, is consistent with reports implying a strong requirement by MCP-1 for the N-terminal domain of CCR2 (41). This contrasts with CCR1 and, as we have shown here, CCR5, whose natural ligands are relatively insensitive to perturbations in the N terminus of the receptor.
Rucker et al. (42) have used constructs similar to 2M5 and observed HIV fusion activity comparable with that of wild-type CCR5. Several possibilities could account for this inconsistency with our data. Unlike constructs used in the Rucker et al. report, 2M5 contains the first intracellular loop of CCR2 and is epitope-tagged at the N terminus. In addition, we used a singlestep entry assay that has a definite linear range and that may be more accurate than a syncytium forming assay for quantifying the ability of different receptors to support HIV-1 envelope-mediated membrane fusion. Other data in Rucker et al. (42) indicate that HIV-1 envelope glycoprotein-induced syncytium formation is sensitive to modifications of the CCR5 N terminus. This conclusion is supported in this study with a receptor whose expression and structural integrity are verified.
Ligands for many G-protein-coupled receptors, including chemokine receptors, are thought to bind at least in part in a pocket formed by the transmembrane helices and induce in the receptor a conformational change that promotes guanine nucleotide exchange in G-proteins (32). Chemokines are thought to bind this pocket at their N termini, and chemokines with N-terminal truncations function as receptor antagonists. Our data imply that the HIV-1 envelope need not induce an activated conformation in CCR5 to enter and thus could bind away from this pocket. Although chemokine inhibition of HIV-1 entry and gp120 binding might imply some overlap of the MIP1␤ and gp120 binding sites on CCR5, our data suggest that at least some of the elements of the binding site are distinct. These differences may need to be considered when designing strategies for therapeutic intervention. Further understanding of the interaction of CCR5 with HIV-1 and with its natural ligands could contribute to these efforts.