Functional Dissection of CCR5 Coreceptor Function through the Use of CD4-Independent Simian Immunodeficiency Virus Strains

Abstract

With rare exceptions, all simian immunodeficiency virus (SIV) strains can use CCR5 as a coreceptor along with CD4 for viral infection. In addition, many SIV strains are capable of using CCR5 as a primary receptor to infect CD4-negative cells such as rhesus brain capillary endothelial cells. By using coupled fluorescence-activated cell sorter (FACS) and infection assays, we found that even very low levels of CCR5 expression could support CD4-independent virus infection. CD4-independent viruses represent valuable tools for finely dissecting interactions between Env and CCR5 which may otherwise be masked due to the stabilization of these contacts by Env-CD4 binding. Based on the ability of SIV Env to bind to and mediate infection of cells expressing CCR5 chimeras and mutants, we identified the N terminus of CCR5 as a critical domain for direct Env binding and for supporting CD4-independent virus infection. However, the activity of N-terminal domain CCR5 mutants could be rescued by the presence of CD4, indicating that other regions of CCR5 are important for post-binding events that lead to viral entry. Rhesus CCR5 supported CD4-independent infection and direct Env binding more efficiently than did human CCR5 due to a single amino acid difference in the N terminus. Interestingly, uncleaved, oligomeric SIV Env protein bound to both CD4 and CCR5 less efficiently than did monomeric gp120. Finally, several mutations present in chronically infected monkey populations are shown to decrease the ability of CCR5 to serve as a primary viral receptor for the SIV isolates examined.

Simian immunodeficiency virus (SIV) is frequently used as a model system for human immunodeficiency virus (HIV) infection because it causes a syndrome very similar to human AIDS in several Asian macaque species. Like HIV, SIV uses chemokine receptors in conjunction with CD4 for viral entry (2, 6, 11, 13, 17, 21, 22, 24, 31, 45). With rare exceptions, all SIV strains identified to date use CCR5 for entry regardless of cellular tropism (11, 24, 37, 45, 47). CXCR4 is used as a coreceptor only very rarely by SIV isolates, although rhesus CXCR4 is functional for X4 HIV-1 strains and several recently described pathogenic SHIVs use CXCR4 as a coreceptor (24, 33, 36, 44, 47). Thus, SIV tropism does not follow the R5/X4 designations given to HIV-1 (5). Although both T- and M-tropic SIVs use CCR5, they may interact with CCR5 differently, and there is evidence to suggest that only M-tropic Env proteins signal through CCR5 (24, 62). SIV typically displays a broader coreceptor use pattern than HIV-1, with most SIV strains using GPR15/BOB and/or STRL33/Bonzo as coreceptors in addition to CCR5 (18, 25, 29, 47). Use of CCR8, ChemR23, GPR1, CCR2b, and APJ as coreceptors by SIV is more restricted (12, 25, 26, 29, 47, 54, 56).

Several lines of evidence support the hypothesis that chemokine receptors actually represent the primordial primate lentivirus receptors. A number of HIV-2 strains, including some primary isolates, are able to infect cells which express CXCR4 or CCR5 but lack CD4 (10, 13a, 14, 28, 52). Laboratory-adapted isolates of feline immunodeficiency virus also use CXCR4 as a receptor in the absence of CD4 (51, 64). Furthermore, a significant number of SIV isolates can use CCR5 as a primary receptor and can infect CCR5-positive, CD4-negative primary cells such as rhesus brain capillary endothelial cells (27). In keeping with their decreased dependence on CD4 for infection, HIV-2 and SIV envelope (Env) proteins have a lower affinity for CD4 than does HIV-1 Env and are more resistant to inhibition by soluble CD4 (sCD4) (8, 57, 58). In fact, low to moderate levels of sCD4 enhance both syncytium formation and infection by a variety of HIV-2 and SIVagm strains (4, 14, 63).

We have extended our studies of CD4-independent infection by SIV strains to better understand the factors that govern this entry process and to identify regions of CCR5 important for SIV Env binding in the absence of CD4. We found that CD4-independent SIV infection occurred over a broad range of CCR5 expression levels, including levels barely detectable by fluorescence-activated cell sorting (FACS). The addition of CD4 in soluble or membrane-bound form enhanced infection for most CD4-independent viruses, suggesting that CD4 is required for optimal entry efficiency. While both human CCR5 (Hu CCR5) and rhesus CCR5 (Rh CCR5) functioned equally well as coreceptors in the presence of CD4, in its absence Rh CCR5 supported infection more efficiently and for a larger number of virus strains. The increase in CD4-independent infection through Rh CCR5 mapped to the Asp at position 13 in the N terminus and correlated with direct Env binding. Several mutations in the CCR5 N-terminal domain were identified that blocked virus infection in the absence (but not in the presence) of CD4, suggesting that the N-terminal domain of CCR5 is more important for Env binding than for triggering the conformational changes that lead to membrane fusion. Interestingly, we found that the ability of uncleaved Env to bind both CCR5 and CD4 was markedly reduced relative to gp120. Finally, several mutations in CCR5 which exist as polymorphisms in chronically infected monkey populations were identified as limiting CD4-independent infection for the SIV strains tested.

MATERIALS AND METHODS

Reporter virus infections.

All SIV Env proteins, CD4, and CCR5 constructs were cloned into the pcDNA3 or pcDNA3.1 vectors (Invitrogen). Luciferase reporter viruses were produced by cotransfecting 293T cells with the env gene of interest under the control of the cytomegalovirus promoter and with the NL-Luc-R⁻-E⁻ backbone plasmid in the presence of 7.5 mM sodium butyrate (9, 15). Supernatant containing virus was harvested 2 days posttransfection and used to infect 48-well plates of 293T target cells transfected with 1 μg of CD4 and 0.5 μg of CCR5 construct or vector as indicated. Infections were performed in Dulbecco’s modified Eagle’s medium (DMEM; Gibco-BRL) with 10% fetal calf serum (FCS; Hyclone) and 1% penicillin-streptomycin (Gibco-BRL) and supplemented with 8 μg of DEAE-dextran (Sigma) per ml, except for infections with the alanine scan mutants, which were done in the absence of DEAE. Cells were lysed in 0.5% Triton X-100 in phosphate-buffered saline (PBS) 3 days postinfection, and luciferase activity in the cell lysate was quantified in a luminometer. sCD4 was kindly provided by Jim Hoxie, University of Pennsylvania.

Flow cytometry.

293T cells were lifted with PBS, transferred to FACS tubes, and washed once with FACS staining buffer (PBS plus 2% FCS). The cells were incubated with the anti-CCR5 monoclonal antibody CTC5 (Protein Design Labs, Mountain View, Calif.) at 10 μg/ml for 30 min at 4°C. They were then washed twice with FACS staining buffer and incubated with phycoerythrin-conjugated horse anti-mouse antibody (Vector Laboratories) at 1 μg/ml for 30 min at 4°C in the dark. They were washed twice with FACS staining buffer and analyzed with a Becton Dickinson FACScanner immediately following staining. Antibody quantification was performed with a QuickCal kit (Sigma) and antibody 2D7 to CCR5 (Pharmigen). Hu CCR5 and RhCCR5 expression levels on transfected 293T cells were assessed by FACS with two monoclonal antibodies: 455519, a generous gift from R&D Systems, and mCR35.4, kindly provided by Protein Design Labs. These antibodies bind to regions in CCR5 that are identical between Rh CCR5 and Hu CCR5 (43).

Direct Env binding assay.

Secreted Env for the binding assay was produced by infecting 293T cells with the recombinant vaccinia virus vTF1.1 (which expresses T7 polymerase [1]) for 1 h in DMEM containing 2.5% FCS and then transfecting the cells in DMEM–10% FCS with gp120 or gp140 constructs in pcDNA3 under the control of the T7 promoter. At 4 h posttransfection, the cells were washed twice with PBS and placed in serum-free DMEM. Approximately 24 h posttransfection, the medium was harvested and cleared by centrifugation, and vaccinia virus inactivated by the addition of Triton X-100 to a final concentration of 0.1% (wt/vol). Env was purified from the medium by lectin chromatography (wheat germ agglutinin coupled to agarose beads [Vector Laboratories]) and eluted with 0.5 M N-acetyl-d-glucosamine (Sigma) in PBS. Two buffer exchanges were performed in a 50-ml Amicon stir cell with a YM30 membrane by using morpholineethanesulfonic acid (MES) buffer, (pH 7.0). For the Env binding assay, 293T cells were transfected with 3 μg of CCR5 construct or vector DNA per 24-well plate as indicated and 24 h later were incubated in DMEM–10% FCS containing 4 μg of soluble Env protein with or without 100 nM sCD4. The cells were incubated for 2 h at 37°C, washed once with 1 ml of serum-free DMEM, lysed in 0.5% Triton X-100 in PBS containing protease inhibitors, and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis followed by Western blotting with an anti-gp120 monoclonal antibody and enhanced chemiluminescence (Amersham) detection. Alternatively, Env was quantified in cell lysates by an antibody capture enzyme-linked immunosorbent assay (ELISA). ELISA plates were prepared by an overnight incubation at 4°C of 96-well plates with capture buffer (20 mM Tris, 0.1 M NaCl, 0.05% NaN₃ [pH 8.0]) containing three monoclonal antibodies generated against SIVmacCP-MAC gp140 and mapped to the SU subunit, i.e., 36D5, 115G2, and 168B2. The plates were then washed with a plate washer and ELISA wash solution (0.05% Tween 20 in PBS), blocked for 1 to 2 h at room temperature in BLOTTO (5% nonfat dry milk, 0.1% Tween 20, and 0.1% NaN₃ in PBS), washed, and incubated with 100 μl of lysate/well for 2.5 h at room temperature. After being washed, the plates were incubated with a 1:500 dilution of serum from a macaque infected with SIVmac251 (from the AIDS Research and Reference Reagent Program) for 1.5 h at room temperature, washed, incubated with rabbit anti-monkey horseradish peroxidase-conjugated secondary antibody (Sigma) at a 1:4,000 dilution for 30 min, and given a final wash. 2,2′-Azinobis(3-ethylbenzthiazolinesulfonic acid) (ABTS) tablets (Pierce) were dissolved in substrate buffer (0.1 M sodium acetate) supplemented with 0.05% H₂O₂, the solution was added to the plate, and the absorbance was measured at 405 nm. AOP-RANTES was a generous gift of Timothy Wells and Amanda Proudfoot (Serono Pharmaceutical Research Institute) (59), and 12G5 (28) was kindly provided by Jim Hoxie.

CCR5 constructs.

The majority of the chimeras and N-terminal truncations have been described previously (20, 55). The additional chimeras 55(25)5 and 55(52)5 were constructed by replacing a ClaI-EcoRI cassette in wild-type CCR5 with a fragment generated by PCR as described previously for N-terminal domain chimeras (55). The junction between CCR5 and CCR2 sequences corresponded to the conserved cysteine (C178) that is presumably involved in a disulfide bond with the first extracellular loop. The mutant cassettes were transferred as needed in other backgrounds, such as CCR2b or a (52)222 chimera. All final constructs were verified by sequencing before being used in transfection experiments. The N13D reciprocal point mutants were constructed with a Quickchange mutagenesis kit (Stratagene). The cloning of vervet CCR5 is described in detail elsewhere (40). Briefly, total cellular DNA was isolated from peripheral blood by using the Puregene DNA Isolation kit (Gentra Systems, Inc.) and amplified with primers in the flanking untranslated regions of the Hu CCR5 gene 5′ (CCR5A; 5′-GGAGGGCAACTAAATACATTCTAGG-3′) and 3′ (CCR5B; 5′-GACTGGTCACCAGCCCACTTGAGTCC-3′). PCR-products were re-solved on a 1.0% agarose gel by electrophoresis, and the appropriately sized band was excised, extracted by the QIAquick gel extraction method, and inserted into the pCR2.1 vector (Invitrogen) by the TA method. The inserts were sequenced in both orientations by using ABI dye-terminator chemistry.

RESULTS

Effect of CCR5 expression levels on CD4-independent infection.

Previous studies have shown that multiple strains of SIV can use CCR5 as a receptor for viral infection in the absence of CD4 in both primary and transiently transfected cells (27). Since surface levels of CD4 and chemokine receptors can influence the efficiency of virus entry (38, 50, 54), we performed linked pseudotyped virus infection and FACS experiments on cells expressing decreasing amounts of Rh CCR5 in the presence and absence of CD4. 293T cells were transfected with variable amounts of Rh CCR5 plasmid and a constant amount of CD4 or control vector plasmid. The next day, cells were either infected with luciferase reporter virions bearing the SIV/17E-Fr Env protein or analyzed by FACS for CCR5 and CD4 expression. Virions containing the amphotropic murine leukemia virus Env protein (which does not use either CD4 or CCR5 for infection) served as positive controls.

CD4-independent infection was observed at all Rh CCR5 levels evaluated, including levels barely detectable by FACS (Fig. 1). It should be noted that transfected cells represent a population of cells expressing different levels of the transfected genes and mean channel fluorescence (MCF) reflects only the average cell in this population (Fig. 1a). Using the 2D7 antibody to CCR5 and a quantitative FACS assay, we estimated that less than an average of 400 copies of CCR5 were expressed on the surface of cells transfected with the smallest amount of Rh CCR5 plasmid (data not shown). CD4-independent infection by SIV/17E-Fr was roughly half as efficient as infection in the presence of CD4 at any given MCF, indicating that CD4-independent virus infection is not more dependent on CCR5 expression levels than is infection in the presence of CD4 (Fig. 1b). In fact, we noted that CD4 expression gradually declined as Rh CCR5 expression increased, with the MCF on Leu3A-stained cells changing from 205 on cells lacking Rh CCR5 to 128 when maximal levels of Rh CCR5 were expressed (data not shown). Since higher levels of CD4 may compensate for lower levels of coreceptor (50), the ratio of CD4-dependent to -independent infection at low levels of CCR5 expression may even be falsely elevated under these experimental conditions. Alternatively, coexpression of CCR5 may interfere with Leu3A binding if CD4 and CCR5 colocalize on the cell surface. Neither the presence or absence of CD4 nor the amount of Rh CCR5 on the cell surface significantly affected the ability of MLV pseudotypes to enter the target cells (data not shown). In summary, these results suggest that CD4-dependent and -independent infection pathways are similarly sensitive to CCR5 expression levels and that even low levels of CCR5 expression can be sufficient to support CD4-independent virus infection.

FIG. 1.

Effect of CCR5 expression levels on pseudotyped virus infection. 293T cells were transiently transfected with from 0.5 to 0.0005 μg of Rh CCR5 plasmid per 48-well and 1 μg of CD4 or vector plasmid as indicated. The next day, cells were either evaluated for CCR5 expression by FACS (a) or infected with luciferase reporter viruses pseudotyped with the SIV/17E-Fr Env protein (b) as described in Materials and Methods. Data from one representative experiment of six are presented.

Rh CCR5 is a more efficient primary receptor for SIV than is Hu CCR5.

Hu CCR5 and Rh CCR5 function equally well as SIV coreceptors in the presence of CD4 (11, 24, 45). To determine if this is the case for CD4-independent virus infection, we expressed Hu CCR5 or Rh CCR5 in 293T cells with or without CD4 and infected them with a panel of luciferase reporter viruses bearing different SIVmac or SIVsm Env proteins (Fig. 2). Rh CCR5 and Hu CCR5 were expressed at nearly identical levels as judged by FACS with two different CCR5 monoclonal antibodies (data not shown). To compare data from different experiments, relative light unit values are presented as the percentage of the luciferase signal obtained with Rh CCR5 in the presence of CD4. We found that Hu CCR5 and Rh CCR5 functioned equally well as SIV coreceptors in the presence of CD4 but that Rh CCR5 functioned more efficiently in its absence. In fact, Envs derived from SIVmac1A11 and SIVsmΔB670 clone 12 mediated the infection of cells expressing Rh CCR5 equally well in the presence and absence of CD4. Interestingly, SIVmacCP-MAC Env displayed complete CD4 dependence with Hu CCR5 while efficient CD4-independent infection was observed with Rh CCR5. Only the SIVmac239 Env protein required CD4 for infection via both Hu and Rh CCR5. These findings indicate that CD4-independent interactions between Rh CCR5 and SIV Env are more efficient than those with Hu CCR5.

FIG. 2.

Differential use of Rh CCR5 and Hu CCR5 by SIV. 293T cells were transiently transfected with 0.5 μg of CCR5 plasmid and 1 μg of CD4 or vector plasmid as indicated and infected on the following day with luciferase reporter viruses bearing the indicated SIV Env proteins. Averaged values for data normalized to Rh CCR5 with CD4 signal are presented from at least four separate experiments. Error bars represent standard error of the mean.

Aspartic acid 13 is responsible for the increased efficiency of CD4-independent infection through Rh CCR5.

Hu CCR5 and Rh CCR5 are highly homologous, with only three amino acid changes in the extracellular domains (I9T, N13D, and R171K) (Fig. 3a). The N13D change was shown by Martin et al. to be important for CD4-independent binding of SIVmac239 gp120 to CCR5 (46). To determine whether the increased ability of Rh CCR5 to support CD4-independent infection by SIVmacCP-MAC was associated with this change, we generated reciprocal point mutations at position 13 in Hu CCR5 and Rh CCR5 and performed luciferase virus infections, in the presence and absence of CD4 (Fig. 3b). In the presence of CD4, all receptor constructs served equally well as virus coreceptors. However in the absence of CD4, Hu CCR5(N13D) and Rh CCR5 functioned as efficient primary receptors whereas Hu CCR5 and Rh CCR5(D13N) did not. Thus, consistent with its proven role in the direct binding of SIVmac239 Env to CCR5 (46), Asp13 plays an important role in supporting CD4-independent infection by some SIV strains.

FIG. 3.

Role of residue 13 in CCR5 receptor function. (a) The putative extracellular domains of Rh CCR5 are shown, with the differences from Hu CCR5 indicated by reverse shading. (b) Luciferase reporter viruses bearing the SIVmacCP-MAC Env protein were used to infect cells expressing the indicated CCR5 constructs in the presence and absence of CD4. Error bars represent standard error of the mean.

We next examined the ability of these CCR5 constructs to bind the SIV/17E-Fr Env protein. We used a direct Env binding assay in which purified, soluble gp120 was incubated with 293T cells transiently transfected with CCR5 constructs in the presence or absence of 100 nM sCD4 for 2 h at 37°C. The cells were then washed once with serum-free medium and lysed, and bound Env was detected by Western blotting. Because the enhanced chemiluminescence signals were nonlinear beyond a twofold concentration range, we also used an ELISA to quantify gp120 binding. In the absence of sCD4, SIV/17E-Fr gp120 bound poorly to Hu CCR5 and to Rh CCR5(D13N) but moderately well to Hu CCR5(N13D) and to Rh CCR5 (Fig. 4). Binding to all CCR5 constructs was enhanced by sCD4, with Hu CCR5(N13D) and Rh CCR5 supporting binding most efficiently. AOP-RANTES, 2D7 (an antibody to the ECL2 of CCR5), and CTC5 (an antibody to the CCR5 N terminus) all blocked the binding of SIV/17E-Fr gp120 to Hu CCR5 with various levels of efficiency, demonstrating the specificity of this assay, while an antibody to CXCR4 (12G5) had no effect (data not shown). These results confirm that Asp13 in Rh CCR5 plays an important role in the direct binding of a CD4-independent SIV Env (46).

FIG. 4.

Binding of SIV gp120 to CCR5. 293T cells were transiently transfected with 3 μg of CCR5 or vector DNA per well of a 24-well plate and incubated on the following day for 2 h at 37°C with medium containing purified, soluble SIV/17E-Fr gp120 with or without 100 nM sCD4. The cells were washed once with serum-free medium and lysed, and the lysate was analyzed for bound gp120 by Western blotting with a monoclonal antibody to gp120 (a) or by antibody capture ELISA (b). ELISA results represent the mean of duplicate determinations from three separate experiments. Error bars represent standard error of the mean.

Binding of oligomeric, non-cleaved Env gp140 to CCR5.

The binding assays described above were performed with monomeric gp120. Since SIV Env is oligomeric, we sought to determine whether the Env quaternary structure affected CCR5 binding. To investigate this, we generated a soluble SIV/17E-Fr Env containing all of the gp120 and the gp41 ectodomain. The bulk of this protein is recovered in a noncleaved, oligomeric form (Fig. 5 and data not shown). The remainder is cleaved to generate gp120 and gp41 ectodomain subunits which exist either as monomers or unstable oligomers. This protein mixture was tested for the ability to bind to Hu CCR5 and Rh CCR5, as well as the N13D and D13N reciprocal mutants, in the presence and absence of sCD4. Both gp140 and gp120 bound directly to Rh CCR5 and Hu CCR5(N13D) in the absence of sCD4 (Fig. 5). However, gp120 bound far more efficiently to CCR5 than did uncleaved gp140, as shown by the inversion of the gp140/gp120 ratio in the Env-bound to CCR5 relative to the input mixture. It is unclear whether the bound, uncleaved gp140 truly interacted with CCR5 or was bound as a consequence of forming mixed oligomers with gp120. Addition of sCD4 enhanced the binding of gp120 and gp140 to all of the CCR5 constructs, with gp120 binding again exceeding that of gp140. The gp120 also bound membrane-anchored CD4 more efficiently than did gp140, which could partially account for differences in gp120 and gp140 binding to CCR5 in the presence of sCD4 (Fig. 5). Taken together, our results indicate that uncleaved, oligomeric SIV/17E Fr Env interacts much less efficiently than does gp120 with both CD4 and CCR5.

FIG. 5.

Uncleaved, oligomeric SIV Env binds to CCR5 less efficiently than does gp120. Purified, soluble Env containing predominantly oligomeric SIV/17E-Fr gp140 and some gp120 (see input lanes, 50, 100, and 500 ng) was used in direct Env binding assays as described in the legend to Fig. 4 and Materials and Methods. Bound Env was detected by Western blotting using a monoclonal antibody to gp120.

CCR5 determinants for Env binding and conformational change induction.

In the absence of CD4, CCR5 must bind Env and trigger the conformational changes that lead to membrane fusion. In the presence of CD4, high-affinity binding of Env to CCR5 may be less important since CD4 may supply this function in trans. If so, mutations in CCR5 which primarily affect Env binding should have a greater effect on CD4-independent virus infection. To examine this, we selected two CD4-independent SIV Env proteins, SIV/17E-Fr and SIVsmΔB670 clone 3, for infection studies with receptor chimeras generated between Hu CCR5 and CCR2b in the presence or absence of membrane-anchored CD4 or sCD4. As always with analyses of chimeric receptors, greater emphasis should be placed on chimeras which function rather than those which do not, since loss of function could be due to specific alteration of an important region or to nonspecific, conformational effects.

We found that ECL2 of Hu CCR5 was necessary (5525 does not function as a coreceptor) and sufficient (2252 is an inefficient but functional coreceptor) for viral entry in the presence of membrane-anchored CD4 for both virus strains (Fig. 6a and b). Inclusion of ECL1 (compare 2555 to 2255) or the N-terminal domain of CCR5 (chimera 5252) in conjunction with ECL2 increased coreceptor efficiency to nearly wild-type levels. The third ECL of CCR5 is nearly identical to that of CCR2b; as expected, substitution of this region (chimera 5552) had little effect on coreceptor function. It is important to note that the chimeras were expressed at equivalent levels except for 2522 and 2555 (50 to 60% of wild-type MCF), 5525 (40 to 50%), and 5552 (10 to 20%) (43). Analysis of additional receptor chimeras revealed that the second half of ECL2 was critical for coreceptor function whereas the first half of this region was less important (Fig. 6c). Once again, more efficient coreceptor function was obtained when both the N-terminal domain (the first 20 residues) and the second half of ECL2 were derived from CCR5. To further investigate the role of the N-terminal domain, a series of CCR5 N-terminal truncations were examined. Removing four or eight residues from the N terminus of Hu CCR5 had no effect on coreceptor function (Fig. 6a and b). However, the deletion of 12 and 16 amino acids reduced coreceptor function to 30 to 40% that observed with wild-type Hu CCR5.

FIG. 6.

Role of CCR5 domains in CD4-independent SIV infection. Transiently transfected 293T cells were used as target cells in luciferase reporter virus infections as described in the legend to Fig. 2 and in Materials and Methods, except that infections were performed in the presence of 50 nM sCD4 where indicated. Chimera expression levels were evaluated with a panel of CCR5 monoclonal antibodies as described previously (43). Infections were performed with viruses pseudotypes with the SIV/17E-Fr Env (a), the SIVsmΔB670 clone 3 Env (b), or the indicated Env proteins (c). mCD4, membrane-anchored CD4. Chimera designations reflect whether each of the four extracellular domains is derived from CCR5 or CCR2b.

These results show that residues within ECL2 of Hu CCR5 are critical for SIV coreceptor function in the presence of CD4 but that the N-terminal domain after residue 8 also plays a significant role. We next examined the chimeras for their ability to serve as primary receptors in the absence of CD4 and found that none of the chimeras were able to serve as primary receptors at wild-type levels (Fig. 6) with the exception of 5552, which is nearly identical to Hu CCR5. Consistent with the idea that the efficiency of Env binding to CCR5 is a critical determinant of its ability to serve as a primary receptor, none of the chimeras supported SIV/17E-Fr gp120 binding, with the exception of 5552, to which weak Env binding was detectable (data not shown). The N-terminal truncation mutants were inefficient primary receptors for SIV/17E-Fr, with complete loss of activity in the Δ12 mutant. Clone 3 was able to use the Δ4 mutant at wild-type levels, showed decreased ability to use Δ8, and did not use Δ12. These findings underscore the critical importance of the N terminus for direct binding of Env to Hu CCR5 and suggest that Env binding determinants lie within the first 12 residues of CCR5.

Effects of sCD4 on virus infection.

The addition of sCD4 enhances SIV infection under some circumstances (4, 63). We reasoned that conformational changes induced in Env by sCD4 might rescue the ability of some receptor chimeras to serve as primary viral receptors for infection, possibly by increasing the efficiency of the interaction with CCR5. Indeed, we found that sCD4 enhanced the ability of 5255, 5252 (chimeras which contain the CCR5 N terminus but do not function CD4 independently), and 5552 to serve as primary receptors for one or both of the viruses tested (Fig. 6a and b). The addition of sCD4 did not affect the use of chimeras lacking the CCR5 N-terminus or ECL2, consistent with their critical roles in virus infection as described above. However, sCD4 could rescue infection through Δ4 and Δ8 for SIV/17E-Fr but not through Δ12 or Δ16 for either SIV/17E-Fr or SIVsmΔB670 clone 3. These results suggest that sCD4 enhances the affinity of Env for the N terminus of CCR5 and that the N-terminal domain supporting direct Env binding is disrupted in the Δ12 and Δ16 mutants. Because sCD4 enhancement of SIV infection has been demonstrated with a variety of CCR5-negative cell lines, we tested whether sCD4 could allow the use of GPR15 or STRL33 as primary receptors. While GPR15 and STRL33 could serve as efficient coreceptors for infection by SIV/17E-Fr and SIVsmΔB670 clone 3, they did not serve as primary receptors even in the presence of sCD4 concentrations as high as 400 nM (data not shown). In addition, direct Env binding to GPR15 and STRL33 could not be detected in the presence of 100 nM sCD4 (data not shown).

Residues in the N terminus and ECL2 are critical for primary receptor function.

To further map the regions important for CD4-independent coreceptor function, we examined CCR5 point mutants in which the charged residues in the extracellular domains were individually changed to alanine (20). In the presence of CD4, all of these mutants served as efficient coreceptors (Fig. 7a). However, in the absence of CD4, D2A and, in particular, D11A showed decreased activity as primary receptors (Fig. 7b). Binding of SIV/17E-Fr gp120 was not detected to D2A or D11A with or without sCD4 (data not shown). These results are consistent with the fact that coreceptor function declined beginning with Δ4 for SIV-17E/Fr, which is sensitive to the D2A mutation (Fig. 6a), and that both viruses tested lost CD4-independent and sCD4-rescued infection between Δ8 and Δ12 (Fig. 6a and b). Since these point mutations were constructed in a background of Hu CCR5 (which lacks Asp13), the effects of D2A and D11A may be rendered more apparent. Furthermore, a point mutation which changes the Ser at position 180 in ECL2 of Hu and Rh CCR5 to a Pro affected CD4-independent but not CD4-dependent viral infection (Fig. 7). Since this mutant functions at wild-type levels in the presence of CD4, it suggests that ECL2 contributes in some way to the Env binding site. Interestingly, the S180P mutation is present in the sooty mangabey population (10). In addition, a CCR5 clone isolated from an African green monkey (AGM) (vervet) containing the D13N mutation found in the AGM population (39) was evaluated (40). Vervet CCR5 failed to serve as a primary receptor for SIV/17E-Fr and allowed only inefficient infection by SIVsmΔB670 clone 3, although it functioned as an efficient coreceptor in the presence of CD4 for both viruses (Fig. 7). These results are potentially interesting since sooty mangabeys and African green monkeys become chronically infected with SIV but do not progress to simian AIDS (3, 7, 49).

FIG. 7.

Role of charged CCR5 residues in CD4-independent SIV infection. Transiently transfected 293T cells were used as targets in luciferase reporter virus assays as described in the legend to Fig. 2 and Materials and Methods in the presence (a) or absence (b) of membrane-anchored CD4. The graphs represent the mean of at least three experiments, and error bars show standard error of the mean.

DISCUSSION

Sequential interactions with CD4 and a coreceptor ultimately result in conformational changes in HIV-1 Env that lead to virus-mediated membrane fusion. In contrast to HIV-1, a number of laboratory-adapted and primary HIV-2 and SIV strains have been identified that short-circuit this normal entry process and require only CCR5 or CXCR4 to infect cells (10, 14, 27, 28). CD4-independent infection may influence viral pathogenesis in several ways. First, infection of CD4-negative, CCR5-positive cells such as brain capillary endothelial cells may play a role in neurotropism (27); B cells and cytotoxic T lymphocytes represent other CCR5⁺ CD4⁻ cell types which might be infected by CD4-independent viruses in vivo (34, 66). Second, since multimeric CD4 interactions are thought to be required for HIV-1 entry (42, 48), viruses that can enter independently of CD4 may be able to broaden their cellular tropism by more efficiently infecting cells that express low levels of CD4. Third, shed gp120 may affect CD4-negative cell types which express chemokine receptors by mechanisms such as Env signaling and receptor down-regulation (16, 62). In addition, these viruses represent a means of examining the interaction between Env and chemokine receptors in the absence of CD4, potentially allowing the identification of determinants in both Env and CCR5 necessary to support high-affinity binding. In this study, we extended our previous observations that many SIV strains can infect CCR5⁺ CD4⁻ cells, explored the consequences of receptor expression levels on CD4-independent infection, and used CD4-independent virus strains to identify domains in CCR5 important for supporting Env binding.

Coreceptor expression levels can influence the efficiency of virus entry (38, 50, 54). However, we found that CD4-dependent and -independent SIV infection pathways exhibited a similar dependence on CCR5 expression levels and that CD4-independent infection occurred even at CCR5 levels barely detectable by FACS (Fig. 1). Therefore, gross overexpression of CCR5 is not required for infection in the absence of CD4, at least not for infection by SIV/17E-Fr. This finding is consistent with our previous work, which showed that levels of CCR5 on brain capillary endothelial cells are sufficient to support CD4-independent virus entry (27).

In our previous study, we demonstrated that some SIV strains can infect cells that transiently express Hu CCR5 (27). Furthermore, we and others have found that Rh CCR5 and Hu CCR5 function equally well as SIV coreceptors in the presence of CD4 (11, 24, 45). However, we found that Rh CCR5 supported CD4-independent infection more efficiently and for a larger number of virus strains than did Hu CCR5 (Fig. 2). In fact, with the exception of SIVmac239, all SIV strains tested exhibited some degree of CD4 independence with cells transiently expressing Rh CCR5. We mapped the enhanced infection through Rh CCR5 in the absence of CD4 to residue 13 of the CCR5 N-terminal domain. These findings are consistent with a previous study by Martin et al. which showed that SIVmac239 gp120 can bind directly to Rh but not Hu CCR5 and that residue 13 was responsible for this phenotype (46). In summary, CD4 independence by SIV strains is more widespread than was first suspected due to the more efficient use of Rh CCR5 than Hu CCR5, and this difference maps to the CCR5 N terminus.

For most HIV-1 strains, CD4 binding to Env induces conformational changes in gp120 that confer the ability to interact with the coreceptor (41, 61, 65). These changes may include increased exposure of a highly conserved region in gp120 thought to play a critical role in coreceptor binding (53). Subsequent Env-coreceptor interactions are believed to trigger additional conformational changes in Env that lead to membrane fusion. It is likely that Env remains bound to CD4 when it associates with the chemokine receptor (41). Consequently, direct, high-affinity binding of gp120 to CCR5 might not be required in the presence of CD4 which could provide this function in trans. If true, mutations in CCR5 which predominantly affect Env binding should have an enhanced effect in the absence of CD4. Mutations which are defective in both the presence and absence of CD4 may not induce conformational changes in Env required for membrane fusion or may disrupt CCR5 binding more dramatically (46).

We found that in the presence of CD4, numerous single-amino-acid substitutions as well as truncations of the N-terminal domain of CCR5 were well tolerated. However, in the absence of CD4, many of these mutations abolished or strongly suppressed virus infection. Several previous studies have implicated the N terminus of CCR5, including residues D2 and D11, as important for interaction with HIV-1 Env and, in the case of D11, with SIVmac239 (20, 23, 30). Based on our findings with the N-terminal domain mutants and the ability of gp120 derived from SIV/17E-Fr to efficiently bind to Rh CCR5 in the absence of CD4, we conclude that Env proteins derived from CD4-independent SIV strains exist in a partially triggered conformation that allows them to interact directly with CCR5, with the N-terminal domain playing an important role. The enhancement of Env binding in the presence of sCD4 is probably due to conformational changes in Env that enable it to bind to CCR5 with even higher affinity.

The finding that membrane-anchored CD4 can rescue the activity of N-terminal CCR5 mutants, presumably by providing a high-affinity binding site in trans, but has no effect on the activity of chimeras lacking CCR5 ECL2 suggests that ECL2 may play a role in triggering membrane fusion subsequent to Env binding, although ECL2 could also affect Env-CCR5 binding. In fact, the S180P mutation in ECL2 affected CD4-independent but not CD4-dependent virus infection. In an effort to identify a CCR5 mutant which was able to support Env binding but not virus infection, we examined the ability of SIV/17E-Fr gp120 to bind 5525. Surprisingly, not only did Env fail to bind to 5525, but also binding could not be detected to 5552, a chimera which can serve as a primary receptor for infection even in the presence of sCD4. These results suggest that the Env binding site on CCR5 involves the N terminus but that binding is highly dependent on the overall conformation of CCR5 and other domains are probably involved.

It is important to note that we used monomeric gp120 to measure the ability of Env proteins to bind wild-type and mutant forms of CCR5. As a consequence, low-affinity gp120-CCR5 interactions may not have been detected by our assay. Since Env exists as an oligomer in the viral membrane, multimeric Env-CCR5 interactions could stabilize otherwise weak gp120-CCR5 binding events by increasing the avidity of the interaction. We attempted to address this point by using a secreted oligomeric form of SIV Env, gp140, that exists largely in an uncleaved state. Surprisingly, oligomeric gp140 interacted with CD4 and CCR5 much less efficiently than did gp120, suggesting that the chemokine receptor binding site in gp120 is affected by Env cleavage. This may have implications for vaccine development, since uncleaved Env proteins may not elicit neutralizing, broadly cross-reactive antibodies to the chemokine receptor binding site (60). If uncleaved forms of Env are less able to interact with coreceptors, this may also prevent nonproductive gp160-coreceptor interactions in the biosynthetic pathway.

For many years, it has been recognized that AGMs and sooty mangabey monkeys are chronically infected with SIV in the wild and, while these viruses cause no disease in their natural host, they produce an immunodeficiency syndrome when transmitted to rhesus or pigtailed macaques (3, 7, 19, 32, 35, 49). It is interesting that the AGM population harbors a CCR5 allele which contains the D13N mutation, which plays an important role in CD4-independent SIV infection (39). Since SIV is thought to have existed in the AGM population for many generations, it has been hypothesized that mutations in CCR5 are likely to occur in regions of CCR5 which limit SIV pathogenicity (39). Interestingly, a separate study found that in four of four sooty mangabeys examined, CCR5 contained the S180P mutation, which we also found to limit CD4-independent virus infection (10). It is tempting to hypothesize that in AGMs and sooty mangabey monkeys, host adaptation to limit viral pathogenesis includes the limitation of CD4-independent infection, and it will be important to evaluate primary AGM and sooty mangabey isolates for CD4 independence with AGM and sooty mangabey CCR5.

In conclusion, we found that multiple SIV strains can utilize Rh CCR5 independently of CD4 for virus infection, that some virus strains can distinguish between CCR5 molecules that differ by only one amino acid in the N terminus, and that the N-terminal domain of CCR5 plays an important role in Env binding. In future studies, the CD4-independent SIV phenotype can also be used to identify Env determinants that are important for receptor interactions. By using closely related pairs of virus strains, it will be possible to identify residues that are involved in direct CCR5 interactions and in distinguishing Rh from Hu CCR5. It will also be interesting to determine what role, if any, CD4 independence plays in vivo through further examination of SIV infection in AGMs and sooty mangabeys, evaluation of the extent of infection of CCR5⁺ CD4⁻ cells such as B cells and cytotoxic T lymphocytes in vivo, and infection of rhesus macaques with closely related SIV strains that differ only in their dependence upon CD4.