Derivation and Characterization of a CD4-Independent, Non-CD4-Tropic Simian Immunodeficiency Virus

ABSTRACT

CD4 tropism is conserved among all primate lentiviruses and likely contributes to viral pathogenesis by targeting cells that are critical for adaptive antiviral immune responses. Although CD4-independent variants of human immunodeficiency virus (HIV) and simian immunodeficiency virus (SIV) have been described that can utilize the coreceptor CCR5 or CXCR4 in the absence of CD4, these viruses typically retain their CD4 binding sites and still can interact with CD4. We describe the derivation of a novel CD4-independent variant of pathogenic SIVmac239, termed iMac239, that was used to derive an infectious R5-tropic SIV lacking a CD4 binding site. Of the seven mutations that differentiate iMac239 from wild-type SIVmac239, a single change (D178G) in the V1/V2 region was sufficient to confer CD4 independence in cell-cell fusion assays, although other mutations were required for replication competence. Like other CD4-independent viruses, iMac239 was highly neutralization sensitive, although mutations were identified that could confer CD4-independent infection without increasing its neutralization sensitivity. Strikingly, iMac239 retained the ability to replicate in cell lines and primary cells even when its CD4 binding site had been ablated by deletion of a highly conserved aspartic acid at position 385, which, for HIV-1, plays a critical role in CD4 binding. iMac239, with and without the D385 deletion, exhibited an expanded host range in primary rhesus peripheral blood mononuclear cells that included CCR5⁺ CD8⁺ T cells. As the first non-CD4-tropic SIV, iMac239-ΔD385 will afford the opportunity to directly assess the in vivo role of CD4 targeting on pathogenesis and host immune responses.

IMPORTANCE CD4 tropism is an invariant feature of primate lentiviruses and likely plays a key role in pathogenesis by focusing viral infection onto cells that mediate adaptive immune responses and in protecting virions attached to cells from neutralizing antibodies. Although CD4-independent viruses are well described for HIV and SIV, these viruses characteristically retain their CD4 binding site and can engage CD4 if available. We derived a novel CD4-independent, CCR5-tropic variant of the pathogenic molecular clone SIVmac239, termed iMac239. The genetic determinants of iMac239's CD4 independence provide new insights into mechanisms that underlie this phenotype. This virus remained replication competent even after its CD4 binding site had been ablated by mutagenesis. As the first truly non-CD4-tropic SIV, lacking the capacity to interact with CD4, iMac239 will provide the unique opportunity to evaluate SIV pathogenesis and host immune responses in the absence of the immunomodulatory effects of CD4⁺ T cell targeting and infection.

INTRODUCTION

The primate lentiviruses human immunodeficiency virus type 1 (HIV-1), HIV-2, and simian immunodeficiency virus (SIV) share a mechanism of target cell entry by interacting with CD4 and a member of the chemokine receptor family (1–3). CD4 binding to the envelope glycoprotein (Env) trimer initiates a cascade of conformational changes, resulting in the formation and exposure of the coreceptor binding site on the gp120 subunit of Env. Following coreceptor binding, the gp41 subunit is released to interact with the target cell membrane, leading to the formation of a fusion intermediate and, ultimately, the 6-helix bundle, which drives membrane fusion and viral entry (1, 3–9). While CCR5, CXCR4, and, less commonly, other coreceptors can be used by these viruses during entry, CD4 tropism, mediated by a highly conserved binding site on gp120, is an invariant feature (1, 10), indicating that it plays a major role in pathogenesis. CD4 binding enables HIV-1 to evade host neutralizing antibody responses by limiting antibody access to neutralizing epitopes once the virion has attached to CD4 on the cell surface (11, 12). In addition, CD4 tropism in vivo focuses viral infection onto CD4⁺ T cell subsets that are critical in mediating adaptive antiviral immunity (13–16). These cells include Th1, Th17, T follicular helper, and T regulatory cells that collectively contribute to the coordinated induction, maturation, and maintenance of cellular and humoral immune responses (17–26) and (for Th17 cells) to the integrity of the epithelial barrier at mucosal surfaces (20, 27, 28).

Although CD4 tropism is conserved, rare examples of CD4-independent viruses have been described that can utilize coreceptors, either CCR5 or CXCR4, for entry in the absence of CD4. These viruses, through mutations in gp120 and/or gp41, preform and expose a functional coreceptor binding site that typically is present only after CD4 binding occurs (29–40). By cryoelectron microscopy, Env trimers on CD4-independent viruses exhibit more open conformations than CD4-dependent viruses, and in the absence of CD4 they acquire conformations typically seen only after CD4 binding and triggering occur (41, 42). Although CD4-independent viruses have been derived in vitro (29–37, 43), they have only rarely been observed in vivo, as in rhesus macaques during late-stage disease or following the depletion of CD4⁺ T cells prior to infection with anti-CD4 antibodies (44–46). CD4-independent viruses typically are highly neutralization sensitive as a result of their more open Env trimer conformations and exposure of neutralization epitopes on cell-free virions that are poorly accessible after binding to the cell surface (12, 32, 39, 40, 47). Thus, CD4-independent viruses likely are strongly selected against in vivo (32, 48, 49). Nonetheless, although not strictly CD4 independent, HIVs and SIVs with the ability to utilize low levels of CD4 for entry are well described, and this phenotype has been proposed to contribute to the infection of macrophages and microglial cells, which exhibit a lower density of CD4 than T cells (32, 43–45, 50–57). For one neuropathic SIV isolate, its ability to cause AIDS encephalopathy in macaques correlated with the infection of brain-derived endothelial cells that expressed CCR5 but not CD4 (58). In addition, CD4-independent infection has been proposed to contribute to SIV infection of macrophages in the context of cell-to-cell transmission (59, 60). Of note, viruses that are CD4 independent typically retain their CD4 binding site and the ability to engage CD4, if present (30, 34, 58, 61, 62), and have been shown to exhibit faster fusion kinetics in the presence of CD4 than CD4-dependent viruses (63).

For HIV-1, the CD4 binding site has been resolved at the atomic level and shown to be a deeply recessed pocket on gp120 formed by regions within the inner and outer domains that interact cooperatively with CD4 and gp41 during CD4 binding (1, 4, 41, 64, 65). Among HIV-1 and SIV isolates, some variability exists in these interactions. For SIVmac, the bulky side chain at Trp-375 has been shown to fill a space in the CD4 binding pocket, reducing its dependency on CD4 binding, while HIV-1, containing a serine at this position, requires additional contributions from a layer on the gp120 inner domain (64). In HIV-1, residues Asp-368, Glu-370, and Trp-427 are highly conserved and make multiple contacts with CD4, particularly amino acids Phe-43 and Arg-59 in its outermost D1 domain. Among these, Arg-59 forms a salt bridge with Asp-368 on gp120, and mutations in gp120 (10, 66) or CD4 (67, 68) that disrupt this bond ablate CD4 binding. Although crystallographic resolution of an SIV gp120 in complex with CD4 has not been determined, an aspartic acid at the analogous position (i.e., amino acid 385 for SIVmac) is conserved in all SIVs except for two SIV mandrill env clones that contain a glutamic acid residue, suggesting that this aspartic acid also is critical for SIV gp120 interactions with CD4 (10) (see Fig. S1 in the supplemental material). Indeed, a brain-derived Env clone from an SIVmac239-infected macaque with an asparagine at this position exhibited a 40-fold reduction in CD4 binding and CD4-independent use of CCR5 in a cell-cell fusion assay (69).

We describe the in vitro derivation of a novel variant of SIVmac239 that is both CD4 independent and unable to interact with CD4. This variant was adapted to replicate in a CD4-negative clone of SupT1 cells that expressed rhesus CCR5. Four mutations in gp120 and 2 in the gp41 ectodomain were associated with CD4 independence, and of these, a D178G mutation in the region analogous to the HIV-1 V1/V2 loops was shown to be necessary for this phenotype, while additional mutations were required to enable a spreading infection to occur. Although iMac239, like other CD4-independent viruses, was highly neutralization sensitive, mutations in gp120 were identified that conferred CD4 independence in the absence of increased neutralization sensitivity. Notably, when Asp-385 was deleted to disrupt the CD4 binding site, fusion and infectivity of parental SIVmac239 were ablated while iMac239 remained fully replication competent on CD4⁺, CCR5⁺, and CD4⁻, CCR5⁺ cell lines and primary macaque lymphocytes. In addition, iMac239 with and without the Asp-385 deletion exhibited an expanded host range in primary macaque peripheral mononuclear cells that included CCR5⁺, CD8⁺ T cells. Thus, iMac239 will provide a novel platform for exploring the molecular and structural determinants for CD4 independence and neutralization sensitivity and has enabled a novel non-CD4-tropic SIV variant to be derived that can be used to directly explore the role of CD4 binding and tropism in pathogenesis and in modulating host immune responses.

MATERIALS AND METHODS

Cell lines.

Human SupT1 and BC7 cell lines (29) were transfected with a lentiviral vector to express rhesus CCR5 (RhCCR5) (70). SupT1/RhCCR5, BC7/RhCCR5, CEMx174, and HUT-78 cell lines were maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS), 2 mM glutamine, and 2 mM penicillin-streptomycin (RPMI-complete). The Japanese quail fibrosarcoma cell line QT6, the human embryonic kidney cell line 293T, and the human HeLa cell line TZM-bl engineered to express CD4 and CCR5 (obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH, from John C. Kappes) were cultured in Dulbecco's modified Eagle medium supplemented with 10% FBS, 2 mM glutamine, and 2 mM penicillin-streptomycin (DMEM-complete).

Env cloning and mutagenesis.

Adapted env clones from SIV-infected BC7/RhCCR5 cultures were isolated as described previously (71). Mutant env genes were created using the QuikChange site-directed mutagenesis kit (Stratagene) by following the manufacturer's protocol. To repair a premature stop codon in the cytoplasmic tail of env at position 734, this codon was reverted to the wild-type (WT) Q by QuikChange. To generate recombinant molecular clones of the SIVmac239 genome containing adapted and mutant env genes, env clones were cloned into the previously described pHVP-2 construct (also known as p239SpSp3′) containing an open nef reading frame with a corrected HindIII site at position 602, and containing the 3′ half of the viral genome, following HindIII/SacI digestion (72). Full-length genome constructs then were generated by cloning pVP-2 with p239SpSp5′ as previously described (71) (p239SpSp5′ was obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH, from Ronald Desrosiers). The identities of the recombinant clones were confirmed using restriction analysis and DNA sequencing. For the generation of Env-pseudotyped viruses, SIV env genes with a premature stop codon in the cytoplasmic tail (CT) coding region (Q734Stop) in pCR2.1 were digested with KpnI and XbaI and cloned into the similarly digested pCDNA3.1(−) expression construct.

Fusion assays.

Env fusion was assessed quantitatively on quail QT6 cells using a cell-cell fusion assay and expression constructs for CD4 and various coreceptors, and the use of a reporter plasmid encoding luciferase under the control of a T7 promoter has been previously described (71, 73, 74). Rhesus CD4 and coreceptors CCR5, CXCR6, APJ, GPR1, GPR15, CCR2, and CCR8 were kindly provided by Ronald Collman and Robert Doms.

Pseudotype viruses.

Pseudotyped viruses were generated as previously described (75) by cotransfecting 293T cells with 8 μg of plasmid encoding the SG3 Δenv-based virus backbone and 4 μg of the appropriate env expression vector for 3 to 8 h with the FuGENE6 transfection agent (Promega). Cell supernatants were collected at 48 h posttransfection and stored at −80°C.

Virus production.

To generate molecularly cloned viruses, 293T cells were transfected with full-length viral genome constructs for 5 h using calcium phosphate. Cell supernatants were collected 48 or 72 h posttransfection and stored at −80°C. The uncloned iMac239 swarm was generated from supernatants of acutely infected BC7RhR5 cells and stored at −80°C.

Viral replication assays.

Virus concentrations were determined by enzyme-linked immunosorbent assay for viral p27 Gag antigen (Advanced Bioscience Laboratories). SupT1/RhCCR5 and BC7/RhCCR5 were inoculated with equivalent amounts of p27-containing virus. Following overnight incubation at 37°C, infected cells were washed in RPMI supplemented with 5% FBS to remove excess virus and then maintained in RPMI-complete medium. Viral replication was monitored by viral reverse transcriptase (RT) activity in the culture supernatants (29, 76).

Neutralization assays.

The sensitivity of pseudotyped viruses bearing Envs of interest to neutralization by sera or plasma samples from SIVmac251-infected rhesus macaques, soluble CD4, or monoclonal antibodies to SIVmac Envs, including murine antibodies 7D3, 11F2, 17A11, 5B11, 4E11, 171C2, and 36D5 (77), human antibodies 1.4H, 6.10F, 6.10B, 9.1A, and 1.7A, or rhesus antibodies 1.10A, 3.11H (78–81), and 4.10F (unpublished data; produced as described in reference 79), was assessed in a TZM-bl pseudotype assay as previously described (75). Briefly, pseudotyped viruses were incubated for 1 h at 37°C with various dilutions of serum, plasma, soluble CD4 (sCD4), or monoclonal antibodies and then used to infect TZM-bl cells pretreated with DEAE-dextran. Cells were incubated at 37°C for 48 h and then lysed with the BriteLite plus luminescence reporter assay system (PerkinElmer). A long terminal repeat (LTR)-driven luciferase reporter gene in TZM-bl cells generated the luciferase signal. Infection was quantified by measuring luciferase activity with a Thermo LabSystems Luminoskan Ascent luminometer. Neutralization was measured as the reduction in luciferase activity compared to that of untreated controls.

Infection of PBMCs.

Purified peripheral blood mononuclear cells (PBMCs) from rhesus macaques stored at −140°C were thawed and stimulated for 3 days with 5 μg/ml concanavalin A (ConA) at a concentration of 10⁶ cells/ml in RPMI-complete medium. Cells (5 × 10⁶) then were inoculated with viruses (125 ng of p27 Gag) and medium supplemented with interleukin-2 (IL-2) (100 U/ml). After 24 h, cells were washed to remove the viral inoculum and cultured in fresh RPMI-complete medium supplemented with IL-2 (100 U/ml).

Antibody reagents.

Antibodies used for surface staining included anti-CD14 BV650, anti-CD20 BV605, anti-CD8 BV570 (BioLegend), anti-CD16 APC Cy7, anti-CD95 PECy5, anti-CCR5 PE (BD Biosciences), anti-CD4 PECy5.5 (Invitrogen Life Technologies), and anti-CD28 ECD (Beckman Coulter). Antibodies used for intracellular staining included anti-CD3 V450 (BD Biosciences) and anti-p27 fluorescein isothiocyanate (FITC; kindly provided by E. Rakasz, WNPRC).

Flow cytometry staining assay.

At peaks of viral replication, infected rhesus PBMCs were identified by p27 Gag positivity and flow cytometry. Aliquots of cells (1 × 10⁶ per sample) were washed once with PBS and stained for viability with Aqua amine-reactive dye (Invitrogen) for 10 min in the dark at room temperature. A mixture of surface marker antibodies was added and kept at room temperature for 30 min in the dark. Cells then were washed with PBS containing 1% bovine serum albumin (BSA) and 0.1% sodium azide and permeabilized for 17 min at room temperature using the Cytofix/Cytoperm kit (BD Biosciences). Immediately following permeabilization, cells were washed in Perm/Wash buffer (BD Biosciences), a cocktail of antibodies for intracellular markers was added, and the cells were incubated in the dark for 1 h at room temperature. Cells then were washed with Perm/Wash buffer, fixed with PBS containing 2% paraformaldehyde, and stored at 4°C until flow cytometric analysis. For flow cytometric analysis, 3 × 10⁵ events were acquired on an LSRII flow cytometer (BD Immunocytometry Systems) modified to detect up to 18 fluorophores. Antibody capture beads (BD Biosciences) were used to prepare compensation tubes for each individual antibody used in the experiment. Data analysis was performed using FlowJo software, version 9.0.1 (TreeStar).

Nucleotide sequence accession number.

The iMac239 env sequence has been deposited in GenBank under accession number KT959233.

RESULTS

Adaptation of SIVmac239 to CD4-negative BC7/RhR5 cells.

We derived a CD4-independent strain of SIV by serially passaging SIVmac239 on a 1:10 mixture of CD4-positive SupT1 cells and a CD4-negative variant of this line, BC7 (62), each of which stably expressed rhesus CCR5 (designated SupT1/RhR5 and BC7/RhR5, respectively). After 13 passages the virus could infect a pure culture of BC7/RhR5 cells and then was passaged an additional eight times in this cell line. The resulting viral swarm was able to replicate with high efficiency both in SupT1/RhR5 and in BC7/RhR5, while parental SIVmac239 could replicate only in the CD4⁺ SupT1/RhR5 cells (Fig. 1A). Env clones PCR amplified from genomic DNA of infected BC7/RhR5 cells were derived, and their ability to mediate CD4-independent fusion using rhesus CCR5 was evaluated on quail QT6 cells (73) (Fig. 1B). One clone (p8cl18) mediated comparable levels of fusion in the presence or absence of rhesus CD4 and was selected for further characterization. When this Env was inserted into a SIVmac239 backbone, the resulting virus (iMac239 p8cl18 in Fig. 1C) was able to replicate with rapid growth kinetics in both SupT1/RhR5 (data not shown) and BC7/RhR5 cells (Fig. 1C), while parental SIVmac239 was only able to replicate in SupT1/RhR5 (Fig. 1A). CD4-independent replication in BC7/RhR5 cells was retained when the expected premature stop codon acquired during passaging of this SIVmac in human cell lines (82) was corrected (data not shown), and this Env was used to construct a virus containing a full-length cytoplasmic tail, designated iMac239, that was employed in all subsequent experiments, except for the production of Env-pseudotyped viruses, for which Envs with a short cytoplasmic tail were used.

FIG 1.

Replication and fusion of CD4-independent variants of SIVmac239. (A) Replication of parental SIVmac239 and an uncloned CD4-independent viral swarm is shown in CD4⁺ SupT1/RhR5 cells (left) and CD4⁻ BC7/RhR5 cells (right), each of which stably expressed rhesus CCR5. RT, reverse transcriptase activity. (B) Fusion activity of SIVmac239 and four iMac239 env clones on QT6 cells using a cell-cell fusion assay. For each env gene, the level of CD4-independent fusion on rhesus CCR5 is shown as a percentage of fusion (luciferase activity) in the presence of rhesus CD4. Background fusion levels on cells expressing GFP only were subtracted. The data shown are the means of three experiments plus the standard errors of the means (SEM). (C) Growth curves in CD4-negative, BC7/RhR5 cells are shown for wild-type (WT) SIVmac239, the iMac239 viral swarm, and four recombinant SIVmac239-based viruses bearing the indicated iMac239 env clones. RT activity in culture supernatants was measured at the indicated time points. Results from a representative experiment are shown.

Mutations required for iMac239 CD4 independence.

Sequencing of the iMac239 p18cl8 env gene revealed seven coding changes from SIVmac239, four in gp120 and three in gp41 (Table 1; also see Fig. S2 in the supplemental material). None of the gp120 mutations occurred in regions analogous to those of HIV-1 gp120 that contribute to the CD4 binding site in HIV-1, indicating that although the iMac239 Env and virus were CD4 independent, they likely retained the ability to interact with CD4. For gp120, three mutations (D178G, D337Y, and R427K) occurred within variable loop regions V1/V2, V3, and V4, respectively, while a single mutation (H224Q) occurred in a region analogous to the HIV-1 C2 region flanking the V1/V2 stem. In gp41, two mutations in the ectodomain (K573T and N673I) were located within regions comparable to HIV-1 heptad-repeat regions 1 and 2 (HR1 and HR2), respectively, while one (L820M) occurred in the cytoplasmic tail.

TABLE 1.

Amino acid differences in envelope glycoproteins between SIVmac239 and CD4-independent iMac239^a

Virus	Env region and amino acid position
	gp120				gp41
	V1/V2	C2	V3	V4	HR1	HR2	CT
	178	224	337	427	573	673	820
SIVmac239	D	H	D	R	K	N	L
iMac239	G	Q	Y	K	T	I	M

^aAlso see Fig. S2 in the supplemental material.

The contributions of these mutations to CD4 independence first were evaluated in a cell-cell fusion assay (Fig. 2A). In the absence of CD4 the iMac239 Env generated fusion levels on CCR5 that were comparable to or slightly greater than those in the presence of CD4, while parental SIVmac239 exhibited <10% fusion. When iMac239 mutations were introduced singly into the SIVmac239 Env, D178G in the V1/V2 loop was sufficient to confer CD4-independent fusion at levels approximately 50% greater than those in the presence of CD4, although gp41 mutations K573T and N673I each produced modest increases in fusion to levels 40 to 50% of the fusion in the presence of CD4.

FIG 2.

Determinants for iMac239 Env CD4 independence in cell-cell fusion and viral replication assays. (A) Fusion activity on rhesus CCR5 in the presence or absence of rhesus CD4 is shown for SIVmac239 Envs containing the indicated single mutations from iMac239. Data from 3 experiments plus SEM are shown as described for Fig. 1B. Two-way analysis of variance (ANOVA) with multiple comparisons was used to determine significance. (B) Replication of SIVmac239-based viruses bearing the indicated Envs is shown in CD4⁺ SupT1/RhR5 (left) and CD4⁻ BC7/RhR5 (right) cells. Four changes in gp120 are sufficient to confer CD4-independent replication. A virus with D178G alone was unable to replicate in either cell type but was rescued for replication in SupT1 by H224Q. RT activity was measured at the indicated time points. Results from a representative experiment are shown.

We next evaluated CD4 independence in an infection assay on SupT1/RhR5 and BC7/RhR5 cells using viruses containing Envs with various combinations of iMac239 mutations (Fig. 2B). When all 4 gp120 mutations were introduced into the SIVmac239 Env (SIVmac239 D178G H224Q D337Y R427K in Fig. 2B), robust CD4-independent replication was observed in BC7/RhR5 cells with kinetics and levels that were comparable to those of a virus with the full iMac239 Env. However, a virus containing only the gp41 K573T and N673I mutations replicated poorly in both cell types (not shown). Interestingly, although the D178G mutation alone was sufficient to confer CD4 independence in the cell-cell fusion assay, a virus containing only this mutation replicated poorly in BC7/RhR5 cells and was noninfectious on CD4⁺ SupT1/RhR5 cells (Fig. 2B). Virions from this virus exhibited levels of Env similar to those of both SIVmac239 WT and iMac239 virus, as measured by Western blotting, indicating that this defect was not the result of a failure of Env incorporation into virions (data not shown). However, when viruses contained the D178G mutation in combination with the gp120 H224Q mutation, replication was restored in both SupT1/RhR5 and BC7/RhR5 cells, although CD4-independent replication occurred more slowly in the latter (Fig. 2B). This apparent rescue of infectivity for virus containing the D178G mutation alone was not seen when the other gp120 mutations were inserted individually (not shown). Significantly, the removal of the D178G mutation from the iMac239 Env with all four gp120 changes resulted in a virus (SIVmac239 H224Q D337Y R427K) that was replication competent on SupT1/RhR5 cells but no longer CD4 independent and unable to infect BC7/RhR5 cells (Fig. 2B). Thus, among the gp120 mutations that conferred CD4 independence to SIVmac239, while D178G in V1/V2 was critical, this mutation alone resulted in a virus that was noninfectious in both CD4-positive and -negative cell types but could be rescued by the H224Q change in gp120.

Neutralization sensitivity of iMac239.

For HIV-1 and SIV, CD4-independent Envs typically are highly neutralization sensitive, owing to their more open conformation of the Env trimer on virions (41, 42) and formation of highly immunogenic epitopes that typically are induced only in the presence of CD4 (11, 32, 39, 40). Given the well-described neutralization resistance of the SIVmac239 Env, we were interested in determining the neutralization sensitivity of the iMac239 Env to a panel of sera and plasma samples from SIVmac-infected rhesus macaques and to a panel of anti-SIVmac gp120 monoclonal antibodies previously shown to potently neutralize laboratory-adapted SIVmac251 but not SIVmac239 (77). Neutralization assays were performed on TZM-bl cells using viral particles pseudotyped with Envs. In addition to iMac239, we also evaluated SIVmac239 Envs containing the four gp120 changes that were sufficient to confer CD4 independence (Fig. 2B) in viral replication assays and an Env containing only the D178G mutation, which was CD4 independent in the cell-cell fusion assay (Fig. 2A).

As expected, whereas SIVmac239 was resistant to neutralization by anti-SIV sera or plasma, with reciprocal 50% inhibitory dilutions (ID₅₀) of ≤270, iMac239 was highly sensitive, with ID₅₀ values of >2 million for plasma and 8,703 for sera (Fig. 3A). We sought to determine if iMac239's neutralization sensitivity could be mapped to particular epitopes; thus, we utilized a panel of monoclonal antibodies to SIVmac variable loops (V2, V3, and V4) and to the CD4 and CCR5 binding sites (77–81) in TZM-bl cell neutralization assays. As with the plasma and serum samples, SIVmac239 was highly resistant to all of the antibodies in the panel, while iMac239 was highly sensitive to 7 of 13 antibodies and resistant only to an anti-CD4 binding site (5B11), anti-CD4/CCR5 binding site (17A11), anti-V2 (171C2), and anti-V4 loop (1.7A) antibody (Fig. 3A). Surprisingly, iMac239 Envs containing the minimum number of mutations in gp120 that conferred CD4 independence in either cell-cell fusion or viral infection assays (Fig. 2) remained neutralization resistant at levels comparable to those of parental SIVmac239. These findings indicate that although typically associated, CD4 independence and increased neutralization sensitivity can be dissociated. Moreover, these findings also suggest that changes in the iMac239 gp41 that were selected for in vitro and were not present in the SIVmac239 D178G H224Q D337Y R427K Env used in this assay were key determinants for its marked neutralization sensitivity.

FIG 3.

Neutralization sensitivity of Envs with iMac239 mutations. Viral pseudotypes containing the indicated Envs were preincubated with various dilutions of plasma, serum, or monoclonal antibodies prior to infection of TZM-bl cells. Reciprocal 50% inhibitory dilutions (ID₅₀) for plasma and serum are color coded (<100, green; 100 to 1,000, yellow; 1,000 to 100,000, orange; >100,000, red). Fifty percent inhibitory concentrations (IC₅₀) of monoclonal antibodies (mAb) are color coded (>2 μg/ml, green; 0.2 to 2 μg/ml, yellow; <0.01 to 0.2 μg/ml, red). (A) Neutralization of SIVmac239 and iMac239 relative to two viruses (SIVmac239 D178G and SIVmac239 D179G H224Q D337Y R427K), each containing the indicated mutations in gp120. (B) Neutralization of iMac239 and iMac239-ΔD385, containing a deletion of Asp-385 within the CD4 binding site. Gray-shaded areas indicate assays that were not tested.

CD4 independence of iMac239 is retained following ablation of the CD4 binding site.

Although CD4-independent Envs are structurally altered and expose or form neutralization epitopes (32, 39–42, 47), as noted above, these Envs retain the ability to bind and use CD4. In order to determine if the iMac239 Env would remain competent for fusion and infection even after its CD4 binding site had been ablated, we introduced a 3-nucleotide deletion removing a codon for an aspartic acid at amino acid position 385 that is highly conserved throughout HIV and SIV phylogeny (10) (see Fig. S1 in the supplemental material). For HIV-1, the analogous aspartic acid at position 368 (HXB numbering) forms a salt bridge with arginine 59 of CD4 (1, 2), and a D368R mutation in HIV-1 gp120 ablates CD4 binding and most CD4 binding site antibody epitopes (10, 66).

The effects of the D385 deletion (ΔD385) on SIVmac239 and iMac239 Envs were assessed in cell-cell fusion assays and on viral replication on CD4-positive and CD4-negative cell lines. Remarkably, whereas ΔD385 largely ablated the fusion of SIVmac239 Env on target cells bearing CD4 and CCR5 to levels of <10% of the wild type, iMac239 fusion was unaffected and actually was enhanced in the presence of this mutation (Fig. 4A). When viral replication was assessed in SupT1/RhR5 and BC7/RhR5 cells, SIVmac239 containing the ΔD385 mutation was unable to replicate in either cell type, whereas iMac239 with or without the ΔD385 mutation replicated in both cell types with similar kinetics (Fig. 4B, left and right). We confirmed that viruses used in these infection assays expressed comparable amounts of gp120 relative to p27 Gag (not shown).

FIG 4.

Effect of the ΔD385 mutation in cell-cell fusion, viral replication, and neutralization assays. (A) Fusion activities of SIVmac239, SIVmac239-ΔD385, iMac239, and iMac239-ΔD385 on rhesus CCR5 only are shown for each Env as the percentage of fusion in the presence of rhesus CD4. Background was subtracted as described for Fig. 1B. The data shown are the means from four experiments plus SEM. (B) Replication of SIVmac239, SIVmac239-ΔD385, iMac239, and iMac239-ΔD385 viruses in CD4⁺ SupT1/RhR5 cells (left) and CD4⁻ BC7/RhR5 cells (right). RT activity in culture supernatants was measured at the indicated time points. Results from a representative experiment are shown. (C) Soluble CD4 (sCD4) neutralization of viral pseudotypes containing SIVmac239, SIVmac251.6, iMac239, and iMac239-ΔD385 Envs is shown on TZM-bl cells. Percent neutralization was calculated using luciferase activity normalized to infection in the absence of sCD4. Results from a representative experiment are shown.

The sensitivity of viral pseudotypes bearing these Envs to neutralization by sCD4 also was assessed as an indicator of CD4 binding to Env trimers on virions. The infectivity of pseudotypes containing SIVmac239, SIVmac251.6 (a laboratory-adapted SIVmac), iMac239, or iMac239-ΔD385 Envs was evaluated on TZM-bl cells in the presence of various concentrations of sCD4. While sCD4 sensitivity was observed for SIVmac239 (IC₅₀, 7.8 μg/ml) and was markedly enhanced for SIVmac251.6 (IC₅₀, 0.1 μg/ml) and iMac239 (IC₅₀ <0.01 μg/ml), iMac239 containing the ΔD385 mutation was highly resistant (IC₅₀, >20 μg/ml) (Fig. 4C). Collectively these findings indicate that although the iMac239 Env retained a CD4 binding site and was highly sensitive to sCD4 neutralization, this Env could mediate entry while lacking a CD4 binding site.

Identifying neutralization epitopes on iMac239 and iMac239-ΔD385.

Given the exquisite sensitivity of iMac239 Env to sera from SIVmac-infected macaques and monoclonal antibodies to gp120 epitopes (Fig. 3A) and the ability of iMac239 virus to replicate without a functional CD4 binding site (Fig. 4B), we sought to determine if conformational changes associated with its altered antigenicity were affected by the loss of the CD4 binding site. Neutralization of viral particles pseudotyped with Envs from iMac239 or iMac239-ΔD385 was assessed on TZM-bl cells using the same panel of monoclonal antibodies as that shown in Fig. 3A. As seen previously (Fig. 3A), the iMac239-ΔD385 Env also was highly neutralization sensitive at levels that were comparable to or greater than those of iMac239 (Fig. 3B). Thus, relative to SIVmac239, CD4-independent iMac239 was globally neutralization sensitive to multiple antibodies, and this sensitivity was further enhanced by the ΔD385 mutation. These findings also indicate that while the ΔD385 mutation largely ablated CD4 binding function, it did not disrupt the antigenicity of the CD4 binding site, as determined by the antibodies in our panel.

Replication of iMac239 and iMac239-ΔD385 in primary rhesus macaque PBMC.

Given the ability of CD4-independent iMac239 to replicate in T cell lines with or without a CD4 binding site, we assessed the infectivity of iMac239 and iMac239-ΔD385 on ConA/IL-2-stimulated primary rhesus PBMCs. Similar to T cell lines, parental SIVmac239 containing the ΔD385 mutation was completely noninfectious. However, both iMac239 and iMac239-ΔD385 replicated to levels identical to those of SIVmac239. Peak intracellular viral p27 Gag expression occurred at different days postinoculation, with SIVmac239 infection peaking at 4 days postinfection and iMac239 and iMac239-ΔD385 peaking at 10 days (Fig. 5A). These data indicate that while CD4 binding could facilitate iMac239 replication, a CD4 binding site was not required for this virus to infect primary cells.

FIG 5.

Replication of SIVmac239 and iMac239 with and without the ΔD385 mutation in rhesus PBMCs. Replication of SIVmac239, SIVmac239-ΔD385, iMac239, and iMac239-ΔD385 in ConA/IL-2-stimulated rhesus PBMCs is shown. p27 Gag in culture supernatants was quantified by ELISA at the indicated time points. (A) Results from a representative experiment are shown. (B, left) For each virus, the percentage of total CD3⁺ T cells that are positive for p27 Gag are indicated. (Right) Flow cytometry cytograms from a representative experiment show the percentage of p27 Gag⁺ CD3⁺ T cells. (C, left) The percentage of p27 Gag⁺ CD3⁺ T cells that express CD4 and/or CD8 at peak infection is shown. (Right) Cytograms from a representative experiment show that for iMac239 and iMac239-ΔD385, a marked increase in p27 Gag is detectable in CD8⁺ T cells. One-way analysis of variance (ANOVA) was used to determine significance.

Given the potential for CD4-independent iMac239 and iMac239-ΔD385 to have an expanded cellular tropism, we next assessed their infectivity for different cell types in the stimulated rhesus macaque PBMC population using flow cytometry with a panel of antibodies to T cells, B cells, and monocyte subsets and to CCR5. Gating strategies were performed with uninfected cells (see Fig. S3 in the supplemental material). Among CD3⁺ T cells assayed at the viral peak, iMac239 infection produced a significant increase in p27 Gag-positive cells compared to both SIVmac239 and iMac239-ΔD385 (i.e., 25.4% versus 5.56% and 6.46%, respectively, in the representative experiment shown in Fig. 5B). Among p27 Gag⁺ CD3⁺ T cells, the vast majority (>90%) of SIVmac239-infected cells were negative for CD4 and CD8, most likely reflecting CD4⁺ T cells from which CD4 was downregulated by the effects of Nef and Env expression (83–86), and only rare cells (<1%) expressed CD8. In marked contrast, for iMac239 and iMac239-ΔD385 infections, on average 60% and 40% of p27 Gag⁺ cells, respectively, expressed CD8 (Fig. 5C). The infection of monocytes (CD16⁺ and CD14⁺) or B cells (CD20⁺) was not observed (data not shown), although we note that culture conditions did not support the expansion of these cell types. Further analysis will be required to determine if any additional CD4-negative cell types in rhesus PBMC populations are infected by iMac239 and iMac239-ΔD385. Collectively, these data indicate that both CD4-independent iMac239 and iMac239-ΔD385 have expanded tropism on primary cells, which includes CD8⁺ T cells.

No use of alternative coreceptors by iMac239 and iMac239-ΔD385.

SIVmac239 and other SIVs have been shown to use coreceptors in addition to CCR5, including CXCR6, APJ, GPR1, GPR15, CCR2, and CCR8 (87–91). To determine if CD4-independent use of CCR5 by iMac239 and iMac239-ΔD385 affected CD4-dependent or -independent use of alternative coreceptors, a cell-cell fusion assay was used to assess fusion on target cells expressing rhesus CXCR6, APJ, GPR1, GPR15, CCR2, and CCR8 with or without rhesus CD4. In the presence of CD4, SIVmac239, iMac239, and iMac239-ΔD385 exhibited some capacity to use CXCR6, GPR1, and GPR15, although the levels of fusion were less than those for CCR5 (Fig. 6A, left). However, in the absence of CD4, only iMac239 and iMac-ΔD385 exhibited CD4-independent fusion, and only on CCR5 (Fig. 6A, right). We also assessed the infection of two CD4⁺, CCR5-negative cell lines, CEMx174 and HUT-78, previously shown to be permissive for SIVmac infection, most likely through their expression of GPR15 (91–96). In contrast to SIVmac239, both iMac239 and iMac239-ΔD385 were unable to replicate in these cell lines (Fig. 6B). Thus, while adapted for the CD4-independent use of rhesus CCR5, these findings suggest that iMac239 and iMac239-ΔD385 are strictly CCR5 tropic and unable to use alternative coreceptors for infection in the presence or absence of CD4.

FIG 6.

Use of alternative coreceptors by SIVmac239, iMac239, and iMac239-ΔD385. (A) Fusion activity of the indicated Envs on rhesus coreceptors in the presence (left) or absence (right) of rhesus CD4 was assessed in a cell-cell fusion assay. In each panel, luciferase activity was normalized to values for rhesus CCR5. Background fusion levels were subtracted prior to normalization. Data shown are the means from three experiments plus SEM. (B) Replication of SIVmac239, iMac239, and iMac239-ΔD385 viruses in CD4⁺ CCR5⁻ CEMx174 (left) and HUT-78 (right) cells. RT activity in culture supernatants was measured at the indicated time points. Results from a representative experiment are shown.

DISCUSSION

CD4 tropism is conserved among all primate lentiviruses and has been proposed to play a key role in protecting viruses from neutralizing antibodies that are sterically restricted from accessing the Env trimer once virions have bound to the cell surface (11, 12). However, by focusing infection onto cells that are critical to host adaptive immune responses, CD4 tropism likely also exerts potent immunomodulatory effects that contribute to disease progression and/or viral persistence. Although CD4-independent viruses have been observed in vivo, particularly in nonhuman primate models of pathogenic SIV infection, they typically appear in the setting of highly immunocompromised hosts with advanced neurological or pulmonary complications (44, 58, 69) at sites where nonlymphoid cells (primarily macrophages) with little or no CD4 are infected. In rhesus macaques depleted of CD4 T cells with anti-CD4 antibodies prior to SIVmac infection, CD4-independent viruses rapidly appeared, in association with SIV encephalitis and macrophage infection (46, 53), indicating that this phenotype can readily emerge in vivo. Because CD4-independent viruses are characteristically neutralization sensitive, it is likely that they are strongly selected against during typical pathogenic infection (32, 48, 49, 97). Recent publications by Yen et al. have suggested that CD4-independent entry serves as a mechanism of cell-cell viral spread in tissue macrophages, which is more efficient and can shield virus from neutralizing antibodies than cell-free transmission, thereby allowing CD4-independent variants to circulate in compartments such as the brain (59, 60). Interestingly, primary isolates of HIV-2, which is less pathogenic than HIV-1 (reviewed in reference 98), have been reported to exhibit CD4 independence in vitro (61, 99). However, for HIV-1 (and as we show for SIVmac239), extensive passaging is required to derive CD4-independent viruses in vitro, indicating that for these viruses there likely are additional barriers to their emergence (29, 30, 34, 36, 37, 39, 47). Of note, CD4-independent Envs typically retain their CD4 binding site, and their infectivity generally is enhanced in the presence of CD4 (30, 32, 46, 61). Thus, while CD4-independent viruses arising in vivo or in vitro provide potentially useful tools to understand conformational changes associated with coreceptor engagement and viral entry (34, 35, 39, 40, 63), their applicability to questions of what role CD4 interactions play in pathogenesis and on host immune responses has been limited.

In this report, we describe the derivation and characterization of a CD4-independent and truly non-CD4-tropic variant of SIVmac239 that lacks the ability to interact with CD4. A CD4-independent virus, iMac239, was first derived in vitro and shown to be highly competent in mediating fusion and infection of cells bearing rhesus CCR5 in the absence of CD4. Unlike SIVmac316, a macrophage tropic variant of SIVmac239 that was CD4 independent in cell-cell fusion assays (32, 44), the determinants for iMac239's altered tropism resided solely within gp120, and this virus did not require a truncated cytoplasmic tail to exhibit this phenotype. Notably, after deletion of the codon for a highly conserved aspartic acid in the CD4 binding loop on gp120, shown for HIV-1 to be critical for CD4 binding, iMac239 remained fully infectious on CD4-negative cell lines expressing rhesus CCR5 and on primary peripheral blood lymphocytes. This mutation in parental SIVmac239 completely ablated its function in cell-cell fusion and infection assays. Moreover, whereas SIVmac239 and especially iMac239 were sensitive to neutralization by soluble CD4, iMac239 containing the ΔD385 deletion was completely resistant, consistent with the view that CD4 binding for this virus was ablated or at least markedly reduced (Fig. 4).

Among the seven mutations in the iMac239 Env, four changes in gp120 were sufficient to confer CD4-independent infection of CCR5-expressing cells. A D178G mutation in the iMac239 V1/V2 loop was critical in that this change alone conferred CD4-independent fusion to the SIVmac239 Env, and correction of this change alone ablated CD4-independent infection by a virus bearing the minimum set of gp120 mutations required for CD4 independence. For HIV-1 and SIV, changes in V1/V2 frequently are associated with CD4 independence (30, 33–35, 43, 100–102) and/or an enhanced ability to infect cells that express low levels of CD4 (60, 99). Structural studies of HIV-1 soluble SOSIP timers (7, 9, 103–105) and cryoelectron microscopic analyses of virion-associated trimers have shown, in the absence of CD4, V1/V2 loops to be oriented toward the apex of the trimer, in contrast to their more lateral positioning upon CD4 activation (8, 41). Given that soluble gp120 even in the absence of CD4 is thermodynamically favored to assume a CD4-bound conformation as an apparent default structure (106) but that its conformation is likely restrained by the V1/V2 and V3 variable loops (104–106), changes in V1/V2 that perturb its quaternary interactions within or between adjacent protomers could favor the spontaneous opening of the trimer to a CD4-bound conformation and promote CD4-independent function. Interestingly, although D178G alone enabled the SIVmac239 Env to fuse independently of CD4 and a viral pseudotype bearing this Env could permit infection in a single-round entry assay, SIVmac239 virus containing only this change was not replication competent on either CD4-positive or -negative cells. However, this Env could be rescued by iMac239's H224Q mutation distal from the V1/V2 stem. Because the Env trimer has been modeled as a metastable structure with the potential to assume conformations that are either favorable or nonpermissive for fusion (107, 108), we interpret these results to indicate that D178G, while necessary for CD4-independent fusion and entry, requires H224Q to guide conformational changes and/or promote Env stability during fusion and to prevent fusion-incompetent conformations from being formed, as described for HIV-1 Envs after cold treatment (107) and/or small-molecule CD4 binding site agonists (108–110).

As noted, it is likely that the enhanced neutralization sensitivity of CD4-independent viruses results from a more open structure, as their Env trimers assume conformations that typically occur only in the presence of CD4, exposing epitopes that are shielded on resting virions (41). In addition, CD4-induced epitopes that contribute to the coreceptor binding site and are highly immunogenic (11) are poorly formed in the absence of CD4 binding and inaccessible to antibodies on cell-bound virions but are able to be targeted on CD4-independent viruses on which they are formed and exposed (39) or sampled more frequently in the absence of CD4 (111). As we demonstrated, iMac239 as well as its non-CD4 binding derivative, iMac239-ΔD385, were globally neutralization sensitive to sera from SIVmac-infected animals and to monoclonal antibodies to CD4-induced and noninduced epitopes (Fig. 3). However, an Env containing only the iMac239 gp120 changes, while CD4 independent, remained highly neutralization resistant, similar to parental SIVmac239. Similar findings were reported by Yen and coworkers, in which the loss of a glycosylation site in the SIVmac239 V2 loop (Asn-173) conferred the ability to infect macrophages in the context of cell-to-cell transmission while retaining the neutralization-resistant phenotype of SIVmac239 (59, 60). In addition to indicating that CD4 independence and enhanced neutralization sensitivity can be dissociated, our findings also suggest that changes in the gp41 ectodomain that arose with iMac239's CD4 independence contribute to its neutralization sensitivity. The effect of gp41 mutations on gp120 neutralization sensitivity has been more extensively characterized in HIV-1 (112–115). Our finding also is consistent with the model of intrinsic reactivity of the Env trimer proposed by Haim et al. (107), in which changes in gp41 enhanced the spontaneous formation/exposure of the HR1 coiled coil, decreasing the threshold for Env to transition upon activation from a high-energy to a lower-energy state.

As described, CD4-independent iMac239 virus, following deletion of Asp-385, remained fully infectious on CCR5-expressing cell lines and on primary lymphocytes. Although the structure of the SIVmac gp120 in complex with CD4 has not been resolved at the crystallographic level, for HIV-1 this residue forms a covalent bond with Arg-59 on human CD4 (corresponding to Lys-59 on rhesus CD4) and is highly conserved across nearly all HIV and SIV isolates (see Fig. S1 in the supplemental material) (10). While we cannot rule out the possibility that iMac239 containing this mutation maintained some low-level interactions with CD4, the finding that it became completely resistant to soluble CD4 while iMac239 was exquisitely sensitive strongly supports the view that CD4 binding was markedly impaired (Fig. 4). We chose to introduce a deletion rather than a point mutation at this position to create a CD4-binding site mutant that would be less likely to revert in vivo in macaques. In vitro, when iMac239-ΔD385 was serially passaged up to 20 times in CD4⁺ SupT1/RhR5 cells, this mutation remained stable (not shown), indicating that the loss of CD4 binding function, at least in cell lines, did not confer a major fitness cost during long-term propagation in vitro.

In rhesus PBMCs cultured with T cell mitogens, SIVmac has been shown to infect CD4 effector and central memory T cells, consistent with the expression of CCR5 on these cells and SIVmac's highly efficient use of this coreceptor for entry (49, 116, 117). Although alternative coreceptors can be used by SIVs in vivo (87, 118), it is likely that levels of CCR5 expression are a key determinant of tropism and pathogenicity. Consistent with this hypothesis, sooty mangabeys, a natural host for nonpathogenic SIVsm infection, exhibit low CCR5 expression on central memory CD4 T cells, likely accounting for the sparing of this subset in the context of SIVsmm infection (117, 119). Among peripheral blood cells stimulated with T cell mitogens, iMac239 with and without the ΔD385 deletion exhibited an expanded host range that included CD8⁺ T cells, most likely through their expression of CCR5 (Fig. 5). We observed that 30 to 65% of CD8⁺ T cells in these cultures expressed CCR5 (not shown), which was associated with the infection of approximately 20% and 4% of CD3⁺ CD8⁺ T cells by iMac239 and non-CD4-tropic iMac239-ΔD385, respectively, in contrast to <0.15% for SIVmac239. The adaptation of SIVmac239 for CD4-independent use of CCR5 led to a reduced capacity to utilize alternative coreceptors (Fig. 6), suggesting that its expanded tropism in vitro was driven largely by CCR5 expression. Collectively, these findings clearly show that the tropism of SIVmac239 on primary cells can be altered and redirected from its exclusive infection of CD4⁺ target cells. Whether these viruses can infect additional cell types, such as NK, B cells, or monocytes, remains to be determined.

The ability to remove CD4 tropism from SIVmac creates new opportunities to assess the role of CD4 in pathogenesis. Nonhuman primate models of AIDS have clearly shown that during early SIV infection, CD4⁺ T cells that express CCR5 and reside in mucosal tissues are selectively and rapidly depleted (116, 120, 121), associated with a disruption in the epithelial barrier that contributes to microbial translocation and systemic immune activation (20, 27, 28, 122, 123). In addition, by focusing infection onto T cell subsets that provide help for adaptive immune responses, including Th1, Th17, and Tfh cells, it is likely that CD4 tropism has profound effects on antiviral immune responses, which ultimately are inadequate to contain viral replication and disease progression. Binding of gp120 to CD4 also has the potential to disrupt CD4's physiologic interaction with HLA class II on antigen-presenting cells, which underlies T-cell immunologic helper functions. Although iMac239-ΔD385 exhibited expanded cell tropism in vitro, its inability to selectively target CD4⁺ T cell subsets raises the possibility that T cell help for cytotoxic CD8 and CD4 cellular responses will be qualitatively or quantitatively altered and that B cell maturation and memory responses, which are dependent on interactions with Tfh cells, will lead to improved antibody responses. Future studies that assess the quality of anti-SIV responses in the context of a CD4-sparing infection will provide new insights into pathogenesis and possibly inform interventions that can be directed to improve host immune responses to infection and vaccines.

Funding Statement

The Nonhuman Primate Humoral Immunology Core Laboratory also received funds through National Institute of Allergy and Infectious Diseases contract HHSN27201100016C. Extensive assistance was provided by the Viral/Molecular and Nonhuman Primate Cores of the Penn Center for AIDS Research (P30-AI045008).