CCR5 HIV-1 Coreceptor Activity

Human (H-) CCR5 is the primary coreceptor for ENV-mediated fusion by R5 strains of human immunodeficiency virus type 1, whereas mouse (M-) CCR5 lacks this function. An array of 23 H/M-CCR5 hybrids containing increasing amounts of H-CCR5 extending from the N terminus generated by random chimeragenesis had a biphasic pattern of coreceptor activity with JRFL and 89.6, revealing active regions in the N-terminal extracellular domain (N-ED) and at the junction of cytoplasmic loop 3. The M-CCR5 mutant in which divergent residues were replaced with the corresponding H-CCR5 N-ED sequence (NyYTsE) gained coreceptor function in fusion but not infection experiments. A M-CCR5 double mutant with substitution of human sequences for divergent residues from the N-ED and cytoplasmic loop 3 had augmented coreceptor activity in fusion assays and gain of function in infection experiments. The SIV-251 ENV utilized H- and M-CCR5 and variants. Flow cytometric analysis of M-CCR5 mutants and bifunctional receptors composed of CD4 domains fused to M-CCR5 mutants excluded the possibility that differences in coreceptor activity resulted from variations in cell surface expression. These results demonstrate that the coreceptor activity of the H-CCR5 N-ED is modulated by intracellular residues, illustrating the complexity of CCR5 requirements for interaction with ENV.

humans (ccr5⌬32; Refs. 13 and 14) and primates (15,16) supports the primary role of CCR5. The precedent of protective alleles and the lack of significant consequences of CCR5 deficiency in humans and knockout mice (17) point to its relevance as a target for preventing HIV-1 infection.
The gp120 subunit of the ENV glycoprotein of R5 strains of HIV-1 binds to CCR5 following conformational activation by CD4 (18,19). Structure-function studies have not provided a clear picture of the domain(s) of CCR5 required for triggering fusion, and various regions have been implicated in this process. Multiple reports have demonstrated that the N-ED of CCR5 is sufficient but not necessary for coreceptor activity with R5 ENV (20 -25). Acidic and aromatic residues in this region have been implicated in coreceptor activity in loss of function experiments (24, 26 -31), and sulfation of active Tyr residues may also contribute to this function (32). The finding of differences in the binding of anti-CCR5 mAbs between receptors expressed in transfectants and primary cells implies that the conformation of this domain may have an impact on coreceptor function as well (25). Analysis of receptor chimeras and point mutants reveals that the body of the receptor may also play a role in the fusion process (20 -25, 28, 30), but it is not clear whether there are multiple sites or one complex structure for interaction with ENV. Interpretation of these findings is complicated by the observations that the domains of CCR5 critical for utilization by different ENV varies (20,(22)(23)(24)(25). In contrast, key regions for CXCR4 utilization by R5/X4 ENV have been localized to sequences in or adjacent to the second extracellular loop (ECL) (33,34), and monoclonal antibodies (mAb) to ECL2 of CXCR4 (35) may block the entry of X4 strains. Similarly, mAbs that bind the corresponding region of CCR5 have been shown to inhibit infection by R5 strains, whereas those that bind the N-ED have lower efficacy (36 -38). These findings raise questions regarding the identity of CCR5 residues that are critical for ENV binding and activation.
To elucidate the precise molecular structures involved in CCR5 coreceptor function that are lacking in M-CCR5, a random chimeragenesis strategy (39) using H-CCR5 and the inactive mouse homolog was employed to create an array of chimeras with junctions spanning the receptor. Analysis of these hybrids in fusion and infection experiments revealed active sequences in the N-ED and at the fifth transmembrane spanning domain (TM5)/C3 junction of H-CCR5. Here, we demonstrate that (i) the motif NyYTsE 2 in the N-ED is sufficient to confer coreceptor activity to M-CCR5 in fusion but not infection experiments; (ii) the substitution of TM5/C3 residues from H-CCR5 in the M-CCR5 (NyYTsE) mutant conferred coreceptor activity in infection experiments and enhanced this role in the fusion model; (iii) M-CCR5 and mutants were utilized by an SIV ENV indicating proper cellular trafficking; and (iv) flow cytometric analysis of M-CCR5 mutants and bifunctional receptors composed of two CD4 Ig domains fused to M-CCR5 variants demonstrated that the observed coreceptor activities do not result from variations in cell surface expression. These findings provide direct evidence that NyYTsE is the active motif in the N-ED of H-CCR5 through loss of function experiments, as previously shown, and in gain of function experiments by replacing the four divergent residues into M-CCR5. It provides direct evidence that amino acids predicted to be at the interface of TM5 and C3 are involved in coreceptor activity, based on the finding that they enhance the function of N-ED residues in the ENV-mediated cell-cell fusion assay model and are critical for sensitivity to infection mediated by a M-CCR5 mutant containing the active residues from the H-CCR5 N-ED.

EXPERIMENTAL PROCEDURES
Random Chimeragenesis-The random chimeragenesis approach (39) was adapted to generate hybrids with H-CCR5 at the N terminus and M-CCR5 at the C terminus (H/M-CCR5; see Fig. 1A). cDNAs encoding H-CCR5 and M-CCR5 were cloned in tandem (5Ј-H-CCR5-M-CCR5-3Ј) in pcDNA3, which encodes ampicillin resistance. Two unique sites (NotI and XhoI) were engineered between the cDNAs and used to linearize the final construct and generate incompatible ends. The linear plasmid was transformed into Escherichia coli, and rare ampicillinresistant colonies, the result of circularization by homologous recombination between H-and M-CCR5, were grown and mapped to select plasmids containing inserts the size of a single (chimeric) CCR5 cDNA. The junction of each hybrid was determined by sequencing the open reading frame.
Site-directed Mutagenesis-CCR5 constructs served as templates for site-directed mutagenesis using the Chameleon double-stranded sitedirected mutagenesis kit (Stratagene, San Diego, CA). Clones containing the programmed mutation(s) were identified by sequencing. The sequence of the open reading frame was confirmed.
ENV-mediated Cell-Cell Fusion Assay-Coreceptor activity of the CCR5 variants was analyzed as described previously (24,40). Constructs encoding coreceptor candidates in the pcDNA3 vector, CD4, and luciferase, under the control of a T7 promoter, were cotransfected into QT6 cells using calcium phosphate precipitation. Effector cells were prepared by infection of QT6 cells with recombinant vaccinia viruses that program the expression of ENV on the cell surface and T7 RNA polymerase in the cytoplasm. After 16 h, the target cells were mixed with QT6 effector cells expressing HIV-1 ENV and T7 polymerase. Eight h following mixing of effector and target cells, the medium was aspirated, and detergent lysates were analyzed for luciferase activity using a LucLite Luciferase Reporter Gene Assay Kit (Packard, Downers Grove, IL) in a Top Count luminometer (Packard).
Infection Assay-Viral infection assays were performed as described previously (40) using pseudotyped recombinant viruses containing cloned env genes and a luciferase reporter gene (41,42). Viral stocks were prepared by co-transfecting 293T cells with plasmids encoding the ADA or SIV 251 ENV and the NL4 -3 luciferase virus backbone. U87-MG target cells were transfected with plasmids encoding CD4 and coreceptor candidates. After 24 h, the cells were infected with viral stocks in the presence of polybrene. Three days post-infection additional medium was added, and the cells were harvested by resuspension in 0.5% Triton X-100 in phosphate-buffered saline the following day. Aliquots were assayed for luciferase activity.
Genetic Engineering of Bifunctional Receptors-Composite receptors composed of the first two Ig-like domains of CD4 fused to CCR5 were prepared using a previously described strategy (43). A segment encoding the signal peptide and Ig domains 1 and 2 of human CD4 was amplified using a downstream primer containing a BamHI site to permit in frame fusion with CCR5. The introduction of this site added two residues (Gly-Ser). The final construct was cloned into pcDNA3 and sequenced.
Analysis of Cell Surface Expression-The level of cell surface expression of coreceptor candidates was determined by flow cytometry. QT6 cells transiently expressing CD4/M-CCR5 variants and Chinese hamster ovary cells transiently expressing M-CCR5 mutants were detached from the culture vessel using trypsin or citrate and incubated for 5 h at 37°C to regenerate receptor proteins. The cells were then incubated with an anti-CD4-mAb phycoerythrin conjugate (22°C) (MT310, DAKO, Denmark) or 227R (4°C), respectively, and control myeloma proteins. Cells stained with 227R were subsequently incubated with a phycoerythrin-labeled secondary antibody. Following washing, the cells were analyzed using a FACScan (Becton Dickinson).

Random Chimeragenesis Produces an Array of H/M-CCR5
Chimeras with Junctions Spanning the Receptor-The linearized construct containing H-CCR5 and M-CCR5 in tandem was transformed into E. coli, resulting in a low frequency of ampicillin-resistant colonies (Fig. 1A). An insert the length of a single CCR5 cDNA was detected in plasmids from 54 of 114 colonies following digestion with HindIII and ApaI. Mapping and nucleotide sequencing of the 54 clones revealed 23 chimeras with different junctions between H-and M-CCR5, which spanned the receptor (Fig. 1B).
Multiple (H12/M) lacked coreceptor function with both JRFL and 89.6, but that containing the first 25 residues (H25/M) had approximately 50% of the activity of H-CCR5 with both ENV. This activity was consistent in hybrids with junctions between H-CCR5 residues 25-90. Hybrids in which cross-over occurred between residues 118 and 158 of H-CCR5 had coreceptor function similar to that of the human receptor with these ENV. Chimeras with junctions between amino acids 183 and 214 showed decreasing coreceptor activity, with the H214/M chimera having the nadir value (Ͻ15% of H-CCR5). The utilization of the remaining chimeras, beginning with H235/M, by JRFL and 89.6 returned to that of H-CCR5. Whereas hybrid receptors with less than 235 residues of H-CCR5 were not effectively utilized as a coreceptor by the RF ENV, H235/M and subsequent chimeras with increasing proportions of the human receptor had coreceptor activity similar to that of H-CCR5.
Analysis  (Fig. 2B). A H12/M variant with substitution by NyYTsE, which encodes a primary structure identical to H25/M, was utilized for ENV-mediated fusion with JRFL.
Intracellular Residues of H-CCR5 Modulate Coreceptor Function of N-ED Motifs-The H214/M chimera had minimal coreceptor function, and H235/M had activity similar to that of H-CCR5, although they differ only at two residues predicted to be at the TM5/C3 boundary. In H-CCR5 they are Lys-219 and Leu-222, which correspond to His-221 and Phe-224, respectively, in M-CCR5. 3 The role of Lys-219 and Leu-222 in coreceptor activity was investigated by site-directed mutagenesis, initially in H214/M and H235/M (Fig. 2C). Conversion of His-221 to Lys in H214/M resulted in an increase in coreceptor activity, which was augmented by the additional replacement of Phe-224 with Leu. Conversely, substitution of the mouse counterparts for these two residues in H235/M resulted in a loss of coreceptor function that was greatest in the double mutant.
Residues in the N-ED and TM5/C3 Region of H-CCR5 Cooperate to Confer Coreceptor Activity to M-CCR5 in the Cell-Cell Fusion Model-The structure-function analysis on the H/M-CCR5 chimeras identified the involvement of residues in two regions of H-CCR5 in coreceptor function. The critical residues in the N-ED and TM5/C3 region of H-CCR5 were introduced into M-CCR5 to determine whether they were sufficient to confer coreceptor activity to the mouse receptor in ENV-mediated fusion assays. The M-CCR5 mutants that were analyzed are listed in Table I

Introduction of H-CCR5 N-ED and Intracellular Residues into M-CCR5 Is Necessary to Confer Sensitivity to Infection-
These M-CCR5 variants were tested in infection assays using pseudotyped viruses containing a reporter gene and the R5 ADA ENV (Fig. 3A). The sensitivity of cells transiently expressing M(NyYT) to infection was similar to that of those expressing M-CCR5. Similarly, the sensitivity of target cells expressing M(NyYTsE) did not appear to be significantly different from those expressing M-CCR5. M-CCR5 mutants in which TM5/C3 residues were replaced with the corresponding amino acids from H-CCR5 had no effect on utilization in the infection experiments. Target cells transiently expressing M(NyYTsEϩK) had a significant increase in viral entry over wild type M-CCR5 and the mutants containing H-CCR5 sequences from either the N-ED or TM5/C3 regions. There appeared to be some enhancement of activity by the additional substitution of the second H-CCR5 TM5/C3 residue at position 224 (M(NyYTsEϩKL)).
M-CCR5 Is a Coreceptor for SIV-Representative H/M-CCR5 chimeras were tested for utilization by the SIV 251 ENV in infection assays. H-CCR5, M-CCR5, and representative chimeras demonstrated consistent levels of coreceptor activity (data not shown). Analysis of wild type M-CCR5 and the mutants revealed that all were utilized as coreceptors by SIV 251 ENV at similar levels (Fig. 3B).

Differences in Coreceptor Activities of M-CCR5 Mutants Do Not Result from Variation in Cell Surface
Expression-None of the available anti-H-CCR5 mAbs bound to M-CCR5, but mapping studies revealed that the epitope recognized by 227R, included residues between Asp-11 and Thr-16 of H-CCR5. Because Asp-11 and Ile-12 are conserved in the mouse homolog, M-CCR5 variants containing the NyYT mutation can be stained using this mAb. Therefore, a representative panel of M-CCR5 mutants containing this epitope were evaluated for levels of cell surface expression with 227R. Staining of cells transiently expressing M-CCR5 variants revealed no differences in the level of expression between the following mutations, NyYT, NyYTsE, NyYTsEϩK, and NyYTsEϩKL (Fig. 4), which demonstrate a spectrum of coreceptor activities.
A novel strategy was developed to provide further evidence 3 A two-codon insertion in the N-ED of M-CCR5 is responsible for the difference in residue number between the human and mouse receptors.  (Fig. 4). c Cell-cell fusion assay (Fig. 2D). d NA, not analyzed.
that differences in coreceptor activity of CCR5 variants did not result from variation in the level of cell surface expression. Bifunctional receptors were engineered in which the first two Ig loops of CD4 were fused to the N terminus of the M-CCR5 variants (Fig. 5A). Analysis of the CCR5 bifunctional receptors with a mAb to CD4 demonstrated that they were all expressed at similar levels on the cell surface (Fig. 5B). Analysis of the CD4/M-CCR5 hybrids in fusion assays using the JRFL ENV (Fig.  4C) revealed that the M(NyYTsEϩK) and M(NyYTsEϩKL) bifunctional receptors had greater activity than those containing only the N-ED or C3 sequences. The relative values of wild type H-and M-CCR5 and the mutant forms of M-CCR5 were similar to those obtained in standard cell-cell fusion assays, in which CD4 and the coreceptor candidates were transfected as individual proteins encoded by separate constructs.

DISCUSSION
In the present study, analysis of an array of H/M-CCR5 chimeras with increasing amounts of the human receptor, which was generated by random chimeragenesis, implicated the involvement of sequences in two regions of the protein, at the N-ED and at the TM5/C3 junction. Conversion of six divergent residues identified by this analysis, four in the N terminus and two in the TM5/C3 region, to the corresponding amino acids in H-CCR5 was sufficient to confer coreceptor function to the mouse homolog in the fusion model and infection assays. The NyYTsE motif in the N-ED of H-CCR5 was sufficient to impart limited coreceptor function to M-CCR5 in fusion but not infection experiments. Introduction of the two H-CCR5 residues in the TM5/C3 region had a modest effect on coreceptor activity of M-CCR5 alone, but their addition significantly augmented this function of M(NyYTsE).
Because substitution of the human TM5/C3 residues into M-CCR5 exerted a limited direct effect in either fusion or infection assays, it is proposed that they enhance coreceptor activity indirectly by influencing the conformation of the protein. The current finding of a biphasic pattern of coreceptor activity with JRFL and 89.6 ENV also favors the interpretation that cooperativity between the N-ED and other regions of  16 , which can be detected by the 227R mAb to the H-CCR5 N-ED, were transiently expressed in Chinese hamster ovary cells. Following transfection, the cells were cultured for 48 h, then detached with trypsin, incubated at 37°C to regenerate cell surface receptors, and stained with 227R or a control myeloma protein at 4°C, washed, and incubated with a mAb-conjugated secondary antibody. Fluorescence was measured using a FACScan and log fluorescence is shown on the X axis. Histograms obtained with cells transiently expressing M-CCR5 and mutants lacking the NYYT sequence were identical to that shown for H-CCR5 and the Ig control. CCR5 is important for this function. The significance of the nadir in coreceptor activity that was observed in the H214/M chimera is supported by the inability of RF, an R5/X4 ENV that is sensitive to changes in coreceptor structure (34), to utilize any of the chimeras containing fewer than 235 residues from H-CCR5. Molecular modeling predicts that residues His-221 and Phe-224 are in TM5 of M-CCR5. When His-221 is converted to Lys, the location of this residue is predicted to be at the boundary of TM5 and C3. Therefore, it is possible that conversion of His-221 to Lys in M-CCR5 removes a positive charge from TM5. This could have an impact on the interaction between TM ␣-helices, resulting in a conformational change of extracellular domains of the receptor. Such conformational changes could indirectly affect the structure of the N-ED or directly modulate the accessibility of cryptic motifs in or adjacent to ECL2 that are also influenced by the N-ED. The increase in coreceptor activity noted for the chimeras that contain ECL1 of H-CCR5 (H118/M) suggests that this domain may contribute to the utilization of the N-ED by JRFL. The absence of coreceptor activity in a chimera composed of extracellular domains of CCR3 in a background of CCR1 also supports the importance of TM and cytosolic domains in this function (44).
The conformation of ECL2 may play a critical role in coreceptor activity. The ability to unmask cryptic coreceptor activity of CXCR4 for R5 ENV by the conversion of a specific acidic residue in ECL2 without altering its utilization by X4 ENV (40) suggests that (i) the conformation of this region is a key factor in coreceptor activity and (ii) motifs with broad reactivity for various ENV may exist in a nonpermissive conformation or are not accessible. A divergent residue in the TM4 region of the African green monkey CCR5 homolog (Arg-163 versus Gly in H-CCR5) exerts a similar effect on coreceptor activity but does not alter ligand binding or signal transduction by the receptor (45). Point mutations in ECL2 of H-CCR5, including Ser-180 (30), Pro-183 (28), Tyr-184 (30), Ser-185 (30), and Lys-197 (24) have been associated with decreased coreceptor activity with various ENV. This raises the possibility that sequences in structures adjacent to ECL2, including TM4, TM5, and C3, may play a critical role in triggering ENV-mediated fusion by presenting a specific conformation. ECL2 of CCR5 has been identified as the binding site for antagonists of coreceptor utilization: (i) the cognate chemokine ligands (46,47) and (ii) inhibitory mAbs (36 -38). The importance of the conformation of the ECL2 region is highlighted by the increased efficacy of mAbs to epitopes in this domain of CCR5 to inhibit HIV-1 infection (36 -38), particularly those that recognize complex epitopes.
The finding that M-CCR5 variants and the H/M-CCR5 chimeras were utilized as coreceptors by an SIV ENV indicates that they undergo appropriate cellular trafficking and suggests that differences in utilization by HIV-1 ENV were not the result of variable cell surface expression. This was directly demonstrated by flow cytometric analysis of M-CCR5 mutants having a range of coreceptor activity, which revealed that the substitution of human TM5/C3 residues in M(NyYTsE) increased coreceptor activity without altering the level of cell surface expression. Experiments with bifunctional receptors in which Ig domains 1 and 2 of CD4 were fused to M-CCR5 mutants also revealed that the relative coreceptor activity correlated with the nature of the mutation and reflected the efficacy observed in standard fusion assays, although fusion proteins were expressed at similar levels on the cell surface.
The N-ED of H-CCR5 has been demonstrated to be a major determinant that is sufficient but not necessary for coreceptor function (20 -25). Mutagenesis studies have revealed that substitution of Ala for multiple acidic, Asp-2 (26), Asp-11 (24,26,27), and Glu-18 (26), and Tyr residues (28 -31) result in loss of coreceptor function. These studies have similarly demonstrated the involvement of Ser-7 and Ser-17 (30,31), Asn-13 (28,30), and Cys-20 (31). Dragic et al. (26) and Farzan et al. (29) have shown that removal of acidic residues and aromatic residues, respectively, from the CCR5 N-ED results in loss of ENV binding. The conformation of this domain may also be dynamic, because Hill et al. (25) have noted differences in the binding of mAbs to the N-ED of CCR5 expressed in transfectants and primary cells. Fusion of the intact ectodomain of CD4 to CCR5 has recently been reported to yield an active bifunctional coreceptor but not to compensate for loss of function resulting from truncation of the majority of the N-ED of CCR5 (25).
Multiple regions of H-CCR5 contribute to coreceptor function, either directly or indirectly. We used random chimeragenesis to pinpoint important regions and gain of function experiments to identify sequences in the N-ED and two residues in the TM5/C3 region that are sufficient to confer coreceptor activity to M-CCR5. The N-ED motif appears to contribute directly to this function, whereas the TM5/C3 residues enhance the activity while exerting a limited direct effect. The involvement of residues predicted to be in segments of the receptor that are not exposed on the extracellular surface in this function suggests that conformation is critical to coreceptor activity. These findings invite the premise that the presence of active residues is not sufficient to impart coreceptor activity unless they are in the proper context.