INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

In a critical sequence of events leading to viral entry, the human immunodeficiency virus type 1 (HIV-1) surface glycoprotein, gp120, binds its primary cellular receptor CD4 and undergoes a conformational change that exposes a coreceptor binding site (3). Subsequent binding of gp120 to a chemokine receptor coreceptor allows fusion of viral and cellular membranes to occur. The function performed by the binding of gp120 to a coreceptor in promoting membrane fusion is unknown, but it is evident that productive infection of the host ultimately depends on this association (15).

The chemokine receptors CCR5 and CXCR4 are the predominant coreceptors for HIV-1 in vivo, and all HIV-1 strains are now classified phenotypically as R5, X4, or R5X4 depending on whether they preferentially utilize CCR5, CXCR4, or either, respectively (3). R5 strains are the predominant species transmitted and may be isolated at any stage of HIV infection. X4 viruses evolve in a subset of patients through mutations in the envelope gene, and their emergence is associated with an accelerated course of disease (12, 37, 42, 44). This capacity to alter coreceptor usage increases the number of CD4⁺ cell populations that are susceptible to HIV infection and complicates the development of antiviral strategies targeting HIV coreceptors in vivo. Additionally, viruses may switch coreceptor in the presence of CCR5- or CXCR4-specific chemokines or small molecule inhibitors in vivo and in vitro (19, 38, 42), and replication of X4 virus in the presence of chemokine or bicyclam ligands can force the emergence of escape mutants with unaltered coreceptor usage (43). To counteract this adaptability we need to elucidate the mechanisms underlying the interaction of gp120 with CCR5 and CXCR4, which should facilitate the development of more precise and effective coreceptor antagonists. Furthermore, by investigating the similarities and differences in the ways these chemokine receptors associate with gp120 and CD4, we may be able to define requirements for a universal HIV coreceptor inhibitor.

As members of the seven transmembrane domain G protein-coupled receptor superfamily, CCR5 and CXCR4 share common structural features, including an extracellular N terminus, three extracellular loops (ECLs), three intracellular loops, and an intracellular C-terminal tail (3). The chemokine receptor and HIV-1 coreceptor functions are separable for both proteins; although the binding sites for chemokine ligands and gp120 overlap, they are discrete (1, 6, 17, 20, 26, 35, 50). An additional site in each protein for interaction with CD4 is suggested by the results of coimmunoprecipitation (34, 51) and colocalization (46) studies. CCR5 appears to be constitutively associated with CD4 (51), while the interaction of CXCR4 and CD4 requires the presence of gp120 (46). Although CCR5 and CXCR4 perform essentially the same function in HIV entry and promote infection with similar efficiency, their mode of interaction with the envelope glycoprotein differs, and the affinity of gp120 for CCR5 may be significantly higher than that for CXCR4 (17, 28).

The gp120 binding site is a complex structure formed primarily by the extracellular domains of the coreceptor molecule, and interactions with this site on CCR5 and CXCR4 are notably strain dependent (3). Studies utilizing chimeric and mutated chemokine receptors have failed to reveal a common mechanism for the association of gp120 with either CCR5 or CXCR4, although certain regions, found mainly within the N terminus and ECL2, appear to contribute significantly to gp120 recognition (4-8, 10, 16, 21, 31, 35, 40, 41, 47, 48, 50). The importance of these domains is also suggested by the ability of targeted peptides and monoclonal antibodies (MAbs) to inhibit infection by certain viral strains (23, 29, 35, 50). Mutations of specific tyrosine residues in the N-terminal region of CCR5 (Y15A) (40) and CXCR4 (Y7A, Y12A) (6) impair viral entry, yet deletions of these residues in CXCR4 through N-terminal truncation can be tolerated (7). Multiple acidic amino acid residues in the N-terminal regions of CCR5 and CXCR4 and especially in ECL2 of CXCR4 have been predicted to form electrostatic associations with basic residues in the hypervariable regions of gp120 (4, 18, 31). Mutational analyses have both supported and contradicted this hypothesis (4, 5, 8, 31, 48). Interestingly, substitution with the CCR5 N-terminal region (36) or a D187A point mutation in ECL2 (8, 10, 48) allows CXCR4 to function as a coreceptor for certain R5 HIV-1 strains, suggesting that the CXCR4 structure contains the determinants for R5 gp120 binding. The converse has not been demonstrated for CCR5, however. Either replacement of the CCR5 N terminus with the CXCR4 N-terminal domain (27, 48) or introduction of aspartic acid into the corresponding position in ECL2 of CCR5 (8) abolishes or greatly diminishes CCR5 coreceptor function.

To investigate the functional differences between CCR5 and CXCR4 we generated a set of chimeric coreceptors for stable expression in the coreceptor negative cell line U87.CD4. Because we wished to address the question of coreceptor specificity, we focused on replacing the extracellular domains, although the substituted regions included transmembrane and intracellular sequences as well. Each coreceptor-expressing cell line was susceptible to HIV-1-mediated cell fusion and infection. We demonstrate common coreceptor usage patterns evident for a diverse panel of X4 viruses and show that entry of X4 and R5X4 viruses into our chimeric coreceptor cell lines was blocked by the CXCR4-specific inhibitor AMD3100. Our data clearly demonstrate that the CCR5 molecule contains features that allow utilization by all X4 viral strains tested. These findings suggest that a common mechanism underlies the gp120-coreceptor interaction, regardless of viral phenotype.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cells. U87MG-CD4 (U87.CD4) cells were obtained from the AIDS Research and Reference Reagent Program and were maintained in Dulbecco's modified Eagle's medium (DMEM)-4 mM L-glutamine-1 mM sodium pyruvate-100 units of penicillin per ml-100 µg of streptomycin per ml (complete DMEM; WUMS Tissue Culture Support Center) supplemented with 15% fetal bovine serum (FBS) and 200 µg of Geneticin per ml (Gibco BRL, Grand Island, N.Y.). BSC40 and 293T cells were maintained in complete DMEM containing 10% FBS. Human peripheral blood lymphocytes (PBL) were isolated on Ficoll-Hypaque from concentrated lympho-platelets obtained from normal donors by pheresis (BJC Hospital, St. Louis, Mo.). Human PBL were stimulated in complete RPMI medium containing 10% FBS and 10 µg of phytohemagglutinin per ml for 3 days and then were maintained in complete RPMI medium supplemented with 10% FBS and 50 units of interleukin 2 per ml.

Viruses. Recombinant vaccinia viruses expressing LacZ (vCB21R), T7 polymerase (vPT7-3), CD4 (vCB-3), and HIV-1 envelopes from the JR-FL (vCB-28) and SF162 (vCB-32) strains were obtained from the AIDS Research and Reference Reagent Program. Vaccinia viruses encoding an uncleaved HIV envelope (UNC; vCB-16), ADA envelope (vCB-39), and HXB2 envelope (vSC60) were a gift from Edward Berger. Recombinant vaccinia viruses encoding the HIV envelopes from the YU2 (vSP-5), 89.6 (vDC-1), and UG92046 (vCB-51) strains were a gift from Christopher Broder. The R5, X4, and R5X4 HIV-1 primary isolates utilized for infection were obtained from the AIDS Research and Reference Reagent Program. All R5 and R5X4 viruses utilized and the X4 strains HXB2 and HT92599 are envelope subtype B. The envelope subtypes of the remaining X4 isolates are as follows: CMU08, subtype B; UG92029, subtype A; UG92021, UG93053, and UG93059, subtype D.

CCR5/CXCR4 chimeric constructs. CCR5 lacking a termination codon and flanked by NheI and NotI restriction sites was generated by PCR using pBBS-CCR5 (from Stephen Peiper) as a template. A plasmid encoding a variant of green fluorescent protein (GFP) (Green Lantern Protein; Gibco BRL) was used as a template for PCR to produce GFP cDNA lacking an initiation codon and flanked by NotI and HindIII sites. The DNA encoding CCR5 and GFP was digested, ligated into the NheI and HindIII restriction sites of pBK-CMV (Stratagene), and sequenced. To assemble chimeras, CXCR4 cassettes were generated by PCR using human PBL cDNA as a template. CXCR4 cassettes were engineered by site-directed mutagenesis to contain NspI, ClaI, and EcoRI restriction sites at nucleotides 168, 394, and 798, respectively, in addition to flanking NheI and HindIII restriction sites. CXCR4 PCR products and CCR5-GFP were digested with appropriate enzyme pairs and ligated to produce CCR5/CXCR4 chimeras expressing all possible combinations of the four segments. The four nucleotide sequences exchanged correspond to the following stretches of amino acids in CXCR4: I, amino acids (aa) 1 to 53; II, aa 54 to 131; III, aa 132 to 267; and IV, aa 268 to 352. The substituted regions of CCR5 were as follows: 1, aa 1 to 45; 2, aa 46 to 123; 3, aa 124 to 263; and 4, aa 264 to 352.

Stable CCR5/CXCR4 chimera-expressing cell lines. To produce U87.CD4 cell lines with stable expression of the chimeric coreceptors, the constructs were digested with NheI and HindIII, treated with Klenow, and blunt-end ligated into the retroviral vector pBABE (received as a gift from Dan Littman), which carries the gene for puromycin resistance. U87.CD4 cells were transfected with coreceptor DNA using Lipofectamine (Gibco BRL) and then selected in media containing 1 µg of puromycin per ml for 3 weeks. Bulk selected cells were stained with anti-CD4-phycoerythrin (PE)-labeled antibody (Calbiochem, La Jolla, Calif.) in phosphate-buffered saline (PBS) containing 5% FBS according to the manufacturer's directions and sorted (WUMS Flow Cytometry Facility) with a gate on cells expressing the highest levels of GFP and PE fluorescence. Cells were collected and maintained in media supplemented with 200 µg of Geneticin and 1 µg of puromycin per ml.

Envelope-dependent fusion assay. Coreceptor-expressing cell lines were tested for their ability to support HIV-1 envelope-mediated fusion using a modification of the assay developed in Edward Berger's laboratory (39). Briefly, the U87.CD4 cell lines were infected with vCB21R. Fusion partner BSC40 cells were infected with vPT7-3 and one of several HIV-1 envelope-expressing recombinant vaccinia viruses. Cells were infected for 1 h (multiplicity of infection, 1) at 37°C in PBS containing 1% FBS and 2 mM MgCl₂. Infected cells were cultured overnight, trypsinized lightly, and washed prior to mixing. For fusion, 10⁵ cells were mixed 1:1 with a fusion partner in DMEM containing 5% FBS in triplicate wells and incubated 3 h at 37°C. Fusion was stopped by addition of NP-40 to a final concentration of 1% and a single round of freeze-thaw. -galactosidase activity of reaction lysates was determined using chlorophenol red--D-galactopyranoside (Calbiochem) as described previously (24). The absorbance of each sample was determined at a wavelength of 590 nm.

HIV-1 luciferase reporter virus infection assay. To generate the luciferase reporter viruses, we utilized pNLluc ER⁺, which encodes the firefly luciferase gene in place of Nef within the HIV_NL43 sequence (11). The lab-adapted HXB2 envelope gene and several chimeric envelope constructs generated previously in our laboratory (30, 49) were substituted for the pNLluc ER⁺ env, utilizing the unique SalI and BamHI restriction sites for digestion and ligation. The chimeric envelopes express the V3 loop of R5 (ADA, SF162, YU2) and R5X4 (SF2) strains within the HXB2 backbone. To produce the 89.6 reporter virus we inserted the SalI/BsaBI fragment of the R5X4 strain 89.6 into the corresponding sites of the HXB2-luciferase construct. The coreceptor usage phenotype of each envelope construct was characterized previously in our laboratory (30), with the exception of 89.6, which exhibited the expected R5X4 phenotype (see Results). In addition to the full-length proviral reporter clones, we generated a pseudotyped HIV luciferase reporter virus by cotransfection of pNLluc ER⁺ and a plasmid encoding the vesicular stomatitis virus (VSV) envelope glycoprotein, VSV-g, as described below for preparation of viral stocks.

HIV-1 primary isolate infection assay. To generate viral stocks, phytohemagglutinin-stimulated human peripheral blood mononuclear cells (PBMCs) were inoculated with 1 ml of primary isolate stocks with various titers. Cells were fed fresh media every 3 to 4 days, and supernatants were harvested 7 and 10 days p.i. (d.p.i.). The p24 antigen content of each stock was quantitated by p24 antigen ELISA. Infections were performed with chimera-expressing U87.CD4 as described for the reporter viruses, except that the viral inoculum contained 2 to 10 ng of p24 and cells were washed extensively following the adsorption period to remove residual input virus. To assess viral infectivity, the p24 content of supernatants collected 2, 5, or 8 d.p.i. was quantitated by ELISA, and cell cultures were scored for the presence of syncytia 5 d.p.i. Infections with AMD3100 followed the protocol given for the reporter viruses.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Generation of CCR5/CXCR4 chimeric constructs and stable cell lines. To study the structural determinants of coreceptor specificity, we constructed a set of chimeric coreceptors expressing all possible combinations of the four extracellular domains from CCR5 and CXCR4. For the analysis of coreceptor localization and expression levels, all the constructs were tagged at the C-terminal end with GFP. To generate the chimeric constructs, three unique restriction sites within the CCR5 cDNA which effectively separate the four encoded extracellular domains were engineered into similar positions in the CXCR4 sequence by site-directed mutagenesis. Restriction enzyme digest of CXCR4 PCR products yielded four cassettes for the exchange of CXCR4 and CCR5 sequences. The amino acid sequence of the predicted extracellular region encoded by each cassette was aligned with the corresponding CCR5 sequence as shown in Fig. 1A. The entire sequence of each extracellular domain was contained within a single cassette except for ECL3, for which the first three amino acids were encoded by cassette 3, with the remainder encoded by cassette 4. As depicted graphically in Fig. 1B, the structures produced from each cassette included the following: (i) the extracellular N terminus and first transmembrane (TM) domain; (ii) the first intracellular loop, the second and third TM domains, and ECL1; (iii) ECL2, the second and third intracellular loops, and TM domains 4 to 6; and (iv) ECL3, the seventh TM domain, and the C-terminal tail. As indicated, chimeras were named by listing the parental coreceptor for each segment in order, with CCR5-GFP represented by 5555 and CXCR4-GFP represented by 4444. The total charges of the extracellular domains were derived using the sequences given in Fig. 1A and ranged from 9 to +4 (Fig. 1B).

View larger version (53K): [in this window] [in a new window]   FIG. 1.   (A) Amino acid sequences of putative extracellular domains of CCR5 and CXCR4 contained within the four substituted regions were aligned using Jellyfish software (Biowire). Sequence identity is highlighted in bold type. Charged amino acid residues are indicated. (B) Schematic representations of CCR5, CXCR4, and the CCR5/CXCR4 chimeras. The total charge of the extracellular domains is shown for each chimera. Chimera nomenclature is described in the text.

View larger version (38K): [in this window] [in a new window]   FIG. 2.   Flow cytometric analysis of stable coreceptor-expressing cell lines. U87.CD4 cell lines were stained with control immunoglobulin G-PE (A) or anti-CD4-PE antibody (B) and analyzed for GFP (FL1) and PE (FL2) fluorescence in a flow cytometer. The coreceptor construct expressed by each cell line is indicated in the upper right corner of each panel. Control U87 CD4 cells express vector alone.

View larger version (53K): [in this window] [in a new window]   FIG. 3.   Subcellular localization of CCR5/CXCR4 chimeric coreceptors. U87.CD4 cell lines were fixed and viewed by confocal microscopy. The chimeric coreceptor expressed by each cell line is indicated in the lower right corner of each panel. Some cells expressing 4454 exhibit an entirely intracellular fluorescence pattern (arrow).

Activity of chimeric coreceptors in envelope-dependent cell fusion. To assess the abilities of the chimeric coreceptors to promote membrane fusion, we mixed the CCR5/CXCR4 chimera cell lines with BSC40 cells expressing a panel of vaccinia-encoded HIV envelope glycoproteins. In addition, the BSC40 and chimera-expressing cells were infected with recombinant vaccinia viruses carrying either the T7 polymerase gene or the lacZ gene linked downstream of the T7 promoter, respectively. Positive fusion events resulted in the mixing of cellular contents and expression of the lacZ gene and were monitored by a colorimetric assay for -galactosidase activity. Background levels of fusion were determined by using either BSC40 cells expressing an uncleaved envelope glycoprotein defective in fusion activity or vector control cells which express only CD4.

Activities of chimeric coreceptors in infection by HIV-1 reporter viruses. The CCR5/CXCR4 chimera cell lines were assayed for their ability to support infection by a panel of luciferase reporter viruses expressing V3 loop chimeric envelopes (Table 2). In this highly sensitive assay, 13 of 14 chimera-expressing cell lines were efficiently targeted by the HXB2 reporter virus, including U87.CD4.5455, which promoted low levels of fusion with HXB2 envelope-expressing cells. Therefore, HXB2 was able to utilize CCR5 as a coreceptor for infection upon substitution of any single CXCR4 domain. This virus was not utilizing single CXCR4 domains for entry, however, as U87.CD4.4445 was refractory to infection by HXB2 virus, although this line was readily infected by the 89.6 virus stock. The 89.6 reporter virus infected every chimera-expressing cell line. A second R5X4 virus, expressing the V3 loop of the SF2 envelope, exhibited infectivity for 8 of 14 chimeric coreceptor cell lines, suggesting that this virus was less tolerant of changes in coreceptor sequence than was the 89.6 reporter virus. Consistent with the results of the fusion assay, none of the chimera-expressing cell lines allowed infection by viruses displaying R5 envelope proteins.

Activities of chimeric coreceptors in infection by HIV-1 primary isolates. To determine whether the use of lab-adapted X4 or chimeric R5 and R5X4 envelope glycoproteins had affected the outcome of the reporter virus assay, we infected the U87.CD4 cell lines with several X4, R5, and R5X4 HIV-1 primary isolates. The kinetics of infection varied with both virus and coreceptor; at early time points some positive cultures failed to produce p24 antigen in quantities detectable above the background of input virus, while at later time points other cultures were characterized by abundant syncytium formation, cell death, and low p24 content in the media (data not shown). Therefore, positive infection was detected by p24 antigen ELISA of supernatants and by observation of syncytium formation. Results from a representative experiment, in which cells were assessed for infection on day 8 p.i., are shown in Table 3.

Inhibition of chimera coreceptor activity by AMD3100. To investigate the mechanism behind the plasticity of the X4 envelope interaction with the CCR5/CXCR4 chimeras, we tested the efficacy of AMD3100 inhibition of infection. U87.CD4 cell lines were preincubated for 30 min with AMD3100 (1 µM) prior to infection with the SF162V3, HXB2, and 89.6 reporter viruses (20 ng of p24 per well) or the primary isolates HT92599 and UG92021 (5 ng of p24 per well). The representative results from one of two similar experiments are shown in Table 4 and confirm the specificity of AMD3100 for CXCR4. Infection of U87.CD4.4444 cells by X4 and R5X4 viruses was fully inhibited in the presence of AMD3100 at this concentration. No effect of the inhibitor was detected for any virus with U87.CD4.5555 (Table 4 and data not shown), and the inability of SF162V3 reporter virus to infect the other cell lines remained unaltered. Unexpectedly, AMD3100 completely inhibited infection of every chimera-expressing cell line by all X4 and R5X4 viruses tested. Syncytium formation was similarly blocked in the presence of AMD3100 for all cell lines in which HIV-mediated cell fusion was observed in the absence of drug (data not shown).

View larger version (15K): [in this window] [in a new window]   FIG. 4.   Dose-dependent inhibition of HIV infection by AMD3100. U87.CD4 cells expressing 4444 or the indicated chimeras were pretreated with increasing concentrations of AMD3100 prior to infection with HXB2 (A) or 89.6 (B) reporter viruses. Results of the luciferase assay performed 48 h p.i. are given as the average relative light units (RLU) of cell lysates from duplicate wells ± standard deviations.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Methodological issues in studies of coreceptor domains. To investigate the structural features of the HIV-1 coreceptor-gp120 interaction, we constructed a set of CCR5/CXCR4 chimeras and generated stable cell lines normalized for chimeric coreceptor and CD4 levels. Based on the assumption that the major determinants of HIV coreceptor specificity lie on the extracellular face of the membrane and given that these regions exhibit the greatest sequence diversity between chemokine receptors, our chimeras express all combinations of the four extracellular domains of CCR5 and CXCR4. However, we cannot exclude the possibility that transmembrane or intracellular sequences contribute to chimeric coreceptor activity. Our study is novel in the use of chimeras of CCR5 and CXCR4, and this has allowed us to define, for the first time, distinct patterns of usage of each coreceptor domain. We would like to highlight particularly novel findings with regards to CXCR4 function.

Patterns of CXCR4 domain use. While other groups have studied the behavior of CXCR4 domains in chimeric coreceptors, our study is the first to demonstrate the ability of each extracellular domain to support X4 fusion and infection. Additionally, our data indicate that although X4 viruses apparently utilize CXCR4 in different ways, there are common, and perhaps a finite number of usage patterns. The most permissive viral strains, which included HXB2, HT92599, and UG93053, infected cells expressing chimeras containing any one of the four CXCR4 domains within the context of CCR5 (Tables 2 and 3). Each was sensitive to changes in the CXCR4 molecule through substitution of the CCR5 domain(s) expressing ECL2 and/or ECL3 sequences but tolerated the exchange of the CXCR4 N terminus and ECL1 for CCR5 domains. A second group of X4 viruses, UG92021 and UG92029, exhibited a similar chimera usage pattern but were unable to utilize 5455, which expresses ECL1 of CXCR4 within CCR5 (Table 3). The HXB2 strain was also inactive with U87.CD4.5455 in the cell fusion assay (Table 1). A third grouping of X4 viruses includes those isolates that were the most sensitive to changes in coreceptor sequence, CMU08 and UG93059 (Table 3). These viruses were able to utilize ECL2 or ECL3 and sometimes the N terminus of CXCR4 for infection, and they exhibited the same restrictions to coreceptor usage as those described above but were also unable to utilize other chimeras. It would be of interest to determine if these distinct patterns of CXCR4 usage relate to virus evolution during disease progression.

CXCR4 inhibition. The universal inhibition of X4 infection of our cell lines by AMD3100 shows unequivocally that a common, CXCR4-like structure is presented in each of the CCR5/CXCR4 chimeras. Inhibition of R5X4 viruses further demonstrates that these strains utilize the chimeras similar to CXCR4. Lu et al. (36) proposed that the evolution of the R5X4 phenotype in an R5 viral strain results from an acquired ability to utilize ECL1 and ECL2 of CXCR4 while retaining the capacity to interact with the N terminus of CCR5. However, the infection of U87.CD4.5554 by the 89.6 strain was also sensitive to AMD3100 inhibition (Fig. 4 and Table 4). Our data indicate that contacts made by AMD3100 with specific residues in ECL2 and TM domain 4 of CXCR4 (25, 33) either are mimicked in the presence of the CCR5 ECL2 or are not important to the antiviral action of the drug. Because CCR5 does not itself interact with this bicyclam (13), we favor the latter explanation. However, we cannot at this time provide an alternative rationalization for the interaction of AMD3100 with the CCR5/CXCR4 chimeras. Comparison of the overall charge presented at the extracellular face of the chimeras in Fig. 1 rules out the basic nature of AMD3100 as an underlying mechanism mediating the inhibition of infection. Therefore, some other intrinsic feature of the structure of AMD3100 must form the basis for its antagonistic effect on coreceptor function. We are currently investigating the activities of other coreceptor inhibitors against the CCR5/CXCR4 chimeras.

Requirements for CCR5 domain activity. Our study also demonstrates an important role for the CCR5 ECL2 domain expressed in the CXCR4 backbone. SF162, an R5 primary isolate, was able to infect U87.CD4 expressing 4454, 4455, or 4555, suggesting that the ECL2 of CCR5 is important for its activity with HIV gp120. Unexpectedly, no activity was seen with any of the chimeras expressing the N-terminal domain of CCR5. Several studies have demonstrated the importance of the extracellular N terminus to CCR5 coreceptor function (4, 16, 18, 21-23, 27, 32, 35, 41, 47). However, it has also been shown that ECL2 participates in CCR5 coreceptor function based on reports of studies using CCR5-specific MAbs (29, 35, 50). Furthermore, Chabot et al. (8-10) have demonstrated that the molecular elements promoting activity with R5 viral envelopes are contained within the CXCR4 sequence and that specific point mutations in ECL2 of CXCR4 can reveal cryptic CCR5 activity. Therefore, it is not surprising that ECL2 of CCR5 expressed in the context of CXCR4 is active with an R5 envelope. However, we were able to detect activity only with the primary isolate of SF162, which expresses the entire SF162 envelope. When the SF162 V3 loop was expressed as a chimeric envelope, activity with the chimeric coreceptors was attenuated. While there are too many differences between the primary isolate and the luciferase reporter virus to draw conclusions about the function of the envelope glycoprotein, a potential explanation for this observation is that regions of gp120 outside the V3 loop contribute to the interaction with ECL2 of CCR5. However, we were unable to demonstrate fusion between cells expressing chimeric coreceptors and the full-length SF162 envelope glycoprotein. Efforts are under way to determine which elements of SF162 gp120 and CCR5 are involved in this interaction and whether they are sensitive to the actions of coreceptor antagonists.