Role for CCR5Δ32 Protein in Resistance to R5, R5X4, and X4 Human Immunodeficiency Virus Type 1 in Primary CD4⁺ Cells

Cells and viruses.

HeLa cells purchased from the American Type Culture Collection (Rockville, Md.) were cultured in Dulbecco's modified Eagle's medium (Quality Biologicals, Gaithersburg, Md.) containing 10% fetal bovine serum (FBS) (HyClone, Logan, Utah), 2 mM l-glutamine, and antibiotics. Recombinant viruses vCB-21R (pT7-LacZ), vTF7-3 (T7 polymerase), vCB-16 (Unc), vCB-43 (BaL), and vCB-41 (LAV) have been previously described (3). The HIV-1 isolates IIIB, Ba-L, and 89.6 were obtained from the NIH AIDS Reagent Program, Rockville, Md.

Recombinant vaccinia virus encoding CCR5 was obtained from Chris Broder (Uniformed Services University of the Health Sciences, Bethesda, Md.). Construction of a recombinant vaccinia virus encoding the CCR5Δ32 protein was performed as previously described (12). Briefly, the cDNA fragment encoding the CCR5Δ32 protein was inserted into p1107 under the control of the vaccinia virus 7.5 promoter and was transfected into thymidine kinase-negative cells that were infected with the Western Reserve wt strain of vaccinia virus. The Western Reserve virus does not encode any foreign proteins and was used as a negative vector control. Positive plaques were identified by dot blot hybridization using ³²P-labeled CCR5Δ32 cDNA. After three cycles of plaque purification on thymidine kinase-negative cells, virus stocks were prepared by infection of HeLa cells and were frozen at −70°C.

Peripheral blood mononuclear cells (PBMCs) from healthy anonymous blood donors who were homozygous for CCR5Δ32 were collected at the Division of Transfusion Medicine, Warren Grant Magnuson Clinical Center, according to an NIH institutional review board approved protocol. A detailed description and an analysis of these samples have been previously described (50). PBMCs from all donors were either used as a total population or to purify the CD4⁺ fraction by positive selection using microbeads coated with antibodies against CD4 according to the instructions provided by the manufacturer (Miltenyi Biotec, Auburn, Calif.). Briefly, the PBMCs were magnetically labeled with CD4 microbeads and the cell suspension was loaded onto a column that had been placed in the magnetic field of a magnetic cell-sorting separator. The magnetically labeled CD4⁺ cells retained in the column were separated from the magnetic beads by removal of the column from the separator (removes the magnetic field) and placement of the column on a suitable tube. The CD4⁺ cells were eluted from the column by use of a plunger.

PBMCs or purified CD4⁺ cells were activated with phytohemagglutinin (PHA) (10 μg/ml) (Sigma Chemicals, St. Louis, Mo.) and 100 U of recombinant human interleukin-2 (rIL-2) (NIH AIDS Reagent Program)/ml for 3 days before use. PBMCs from two HIV-infected (−/−) individuals were obtained from H. Naif, Sydney, Australia, and H. Sheppard, San Francisco, Calif. The genotype of these samples was confirmed by reverse transcription (RT)-PCR. RT-PCR was performed on the total RNA (0.5 μg) isolated from uninfected and infected (−/−) PBMCs. Forward and reverse oligonucleotide primers for amplification were 5′-TGTGAAGCAAATCGCAGCCC-3′ and 5′ ATGGTGAAGATAAGAGCCTCACAGCC-3′, respectively. The primers were designed to amplify a 616-bp CCR5Δ32 fragment and a 648-bp CCR5 fragment.

Antisera against the CCR5Δ32 carboxy terminus.

To obtain antisera specific for the CCR5Δ32 protein, rabbits were immunized with a peptide corresponding to the carboxy terminus of the protein (IKDSHLGAGPAAACHGHLLLGNPKNSASVSK) conjugated to keyhole limpet hemocyanin. The carboxy-terminal region was chosen because this region has no shared amino acid homology with CCR5, and antibodies against this domain are specific for the CCR5Δ32 protein. The immunoglobulin G fraction was purified from crude antisera by using protein G-Sepharose, and further purification was performed by using a CCR5Δ32-specific peptide affinity column. The peak fraction containing affinity-purified antibodies was stored at −70°C until use.

FACS analysis.

Cells were washed twice in fluorescence-activated cell sorting (FACS) buffer (supplemented with 0.5% FBS and 0.02% sodium azide), resuspended in 100 μl of FACS buffer at 10⁷/ml, and incubated with a 1:200 dilution of monoclonal antibodies (MAbs) raised against the different coreceptors at 4°C for 30 min. All MAbs and matched isotype controls were purchased from Pharmingen Inc., San Diego, Calif. Cells were then washed twice, resuspended in 100 μl of ice-cold FACS buffer in the presence of phycoerythrin (PE)-conjugated anti-mouse immunoglobulin G, and incubated at 4°C for 30 min. Finally, cells were washed twice, resuspended in 500-μl of ice-cold FACS buffer, and analyzed in a FACScan cytometer (Becton Dickinson).

Fresh PBMCs from 10 different donors were isolated by Ficoll centrifugation. Three-color flow cytometry was performed on unstimulated PBMCs or on cells stimulated with either PHA plus IL-2 or αCD3 antibody plus IL-2 for 3 days. A PE-conjugated MAb against CXCR4 (12G5), a fluorescein isothiocyanate-conjugated MAb against CCR5 (2D7), a cychrome-conjugated MAb against CD4 (RPA-T4), and a cychrome-conjugated CD3 MAb (HIT3Aa) were obtained from Pharmingen. Staining was performed by incubating 10⁶ PBMCs in phosphate-buffered saline (PBS) containing 1% bovine serum albumin with the respective MAb(s) alone or in combination depending on the gate used (i.e., either CD3 or CD4 gated). Incubations were done for 30 min at 4°C and reaction mixtures were washed three times before analysis in a FACScan apparatus (Becton Dickinson).

Construction of recombinant vectors encoding CCR5 or CCR5Δ32.

DNA fragments encoding either CCR5 or CCR5Δ32 were subcloned into pMCV-SV24 under control of the major late promoter of adenovirus type 2. The plasmid construct and method used to generate recombinant adenoviruses encoding measles virus hemagglutinin (MVHA) have been previously described (2). We have used this method to generate recombinant adenoviruses carrying several genes of measles virus (2, 5, 6). Briefly, CCR5 or CCR5Δ32 cDNA was inserted by homologous recombination into the early region of the adenovirus type 5 (Ad5) genome, replacing the E1 region (reviewed in references 10 and 18). The resulting E1-deficient viruses are defective for replication and are propagated to high titers in human 293 complementation cells by providing the missing E1 gene products in trans (19). Efficient expression of CCR5Δ32 or CCR5 by Ad5/Δ32 or Ad5/CCR5 is accomplished by infecting primary cells at a high multiplicity of infection, since these recombinant Ad5 vectors are replication defective in cells other than 293 cells.

HIV-1 Env-mediated fusion using Ad5/CCR5-infected target cells.

The basic features of the fusion assay were developed by using the HIV-1 Env-CD4 interaction of two different populations of cells, with one expressing CD4 and the other expressing the HIV-1 envelope glycoprotein (Env) (29). Target cells (purified CD4⁺ cells) were coinfected with Ad5/Δ32, Ad5/CCR5, Ad5 vector, or Ad5/MVHA and Ad5Pol3. Ad5Pol3 was obtained from Frank Graham, McMaster University, Hamilton, Ontario, Canada. Ad5Pol3 is a recombinant adenovirus that encodes T7 RNA polymerase. HeLa cells coinfected with vCB-21R (LacZ under control of the T7 promoter) and one of the HIV-1 Env proteins served as effector cells. After mixing of the effector and target cell populations and incubation at 37°C for 2.5 h, the fusion specificity was measured by β-galactosidase (β-Gal) production in a colorimetric lysate assay. In experiments shown in Fig. 6, 293-CD4 cells were used to analyze CCR5Δ32 protein interactions and the protein's effect on Env-mediated cell fusion. 293-CD4 cells were coinfected with either Ad5/CCR5 plus Ad5/Δ32 or Ad5CCR5 plus Ad5 and Ad5Pol3 and were challenged with HeLa cells expressing HIV-1 Env and LacZ under control of the T7 promoter.

Western blot analysis.

For Western blotting, cell lines or primary CD4⁺ cells purified from CCR5^+/+ PBMCs were infected with Ad5/Δ32, and the expression of CCR5Δ32 was monitored by immunoblotting. For CCR5Δ32 protein detection in CCR5Δ32^−/− individuals, primary CD4⁺ cells were purified from CCR5Δ32^−/− PBMCs. Cell lysates were prepared, fractionated by sodium dodecyl sulfate-12.5% polyacrylamide gel electrophoresis (PAGE), and immunoblotted onto polyvinylidene difluoride membranes (Millipore). After blocking, membranes were reacted with polyclonal antibodies directed against the N terminus of CCR5 (5) or with affinity-purified CCR5Δ32-specific antibodies at a 1/100 dilution, washed, and incubated with horseradish peroxidase-conjugated secondary antibodies. Protein bands were detected by the addition of a substrate.

Analysis of CCR5Δ32 protein interactions.

The yeast Matchmaker two-hybrid system 3 (GAL4-based) was utilized to study protein-protein interactions as suggested by the manufacturer (Clontech). The system provides a transcriptional assay for the detection of protein interactions in vivo in the yeast Saccharomyces cerevisiae. The CCR5Δ32 ORF was expressed as a fusion to the GAL4 DNA-binding domain (DNA-BD), while CCR5 or CXCR4 was expressed as a fusion to the GAL4 activation domain (AD). The DNA-BD is amino acids 1 to 147 of the yeast GAL4 protein, which binds to the GAL upstream activation sequence upstream of the reporter genes. If CCR5 or CXCR4 interacted with CCR5Δ32, AD amino acids 768 to 881 act as a transcriptional activator to drive expression of the reporter genes. Strain AH109, which uses the reporters Ade2, HIS3, and LacZ under the control of distinct GAL4 upstream activity sequences, was used to study coreceptor-CCR5Δ32 interactions. At medium stringency, HIS3 was used as a reporter gene, with 10 mM 3-amino-1,2,4-triazole (3-AT) as a competitive inhibitor.

Complex formation, or heterodimerization, between CCR5Δ32 and either CCR5 or CXCR4 was analyzed in 293-CD4 cells that had been coinfected with Ad5Pol3 and either Ad5/CCR5 plus Ad5/Δ32 or Ad5/CCR5 plus Ad5. At the appropriate time after infection, a portion of infected cells was used in a cell fusion assay as described above, and the other portion was lysed in RIPA buffer (0.5% Triton X-100, 250 mM NaCl, 50 mM Tris-HCl [pH 7.4], 0.2 mM phenylmethylsulfonyl fluoride) containing 0.1% bovine serum albumin, rabbit polyclonal anti-CCR5Δ32 serum, and protein A-Sepharose. For resolution of the protein complexes, the precipitated proteins (complexed to protein A-Sepharose) were washed three times in RIPA buffer and resuspended in perfluoro-octanoic acid (PFO) loading buffer (50 mM Tris base [pH 8.0], 4% [wt/vol] NaPFO, 10% glycerol, and 0.005% bromophenol blue). Samples were fractionated in a freshly poured 10% Tris-glycine-acrylamide gel without sodium dodecyl sulfate. The running buffer contained 25 mM Tris, 192 mM glycine, and 0.5% PFO. It was previously demonstrated that the detergent PFO protects interactions within protein oligomers (34). The fractionated proteins were transferred onto a polyvinylidene difluoride membrane by blotting (Millipore). After blotting, the membrane was treated with 5% skim milk powder in PBS for 2 h, incubated with anti-CCR5Δ32 antibody for 16 h at 4°C, and washed with PBS containing 0.1% Triton X-100. Bound antibodies were detected by incubation with horseradish peroxidase-conjugated secondary antibodies and the addition of a substrate after washing. The membrane was stripped and used to react with an anti-CCR5 MAb (CTC-6; Protein Design Labs). After detection, the membrane was stripped again and reacted with antibodies specific to CXCR4 (17).

HIV-1 infection of wt primary CD4⁺ cells transduced with Ad5/Δ32.

PHA-plus-IL-2-activated CD4⁺ cells isolated from an HIV-seronegative (+/+) donor were infected with either Ad5/Δ32 or Ad5/CCR5 at a multiplicity of infection of 50 PFU/cell. Infected cells were incubated for 2 days to allow the expression of recombinant proteins (CCR5 or CCR5Δ32 protein) and then were infected with either IIIB (X4), Ba-L (R5), or 89.6 (R5X4) HIV-1. The virus was adsorbed for 2 h, and cells were then washed three times with PBS and maintained in RPMI 1640 supplemented with 10% FBS, PHA, and IL-2. The culture fluid (50 μl) was harvested after cell resuspension every 3 days and was replaced with fresh medium. The amount of p24 antigen in the cell-containing supernatants was measured by using an enzyme-linked immunosorbent assay kit purchased from DuPont (Wilmington, Del.).

Reduced susceptibility to X4 HIV-1 infection of primary CD4⁺ cells isolated from CCR5Δ32 homozygous individuals.

A biological activity of CCR5Δ32 protein has been described as downmodulating CCR5 at the cell surface in a TDN manner (8, 13). The recent isolation of X4 (27) and R5X4 (28) from infected (−/−) individuals suggested that viral transmission occurred through CXCR4. However, why CXCR4 is rarely used in uninfected (−/−) individuals remains an unresolved question. To investigate this issue, we first compared the susceptibility to X4 HIV-1 infection of CD4⁺ cells purified from four HIV-uninfected CCR5Δ32 homozygotes (−/−) and from individuals carrying the homozygous wt CCR5 allele (+/+). The genotypes of these individuals were confirmed by RT-PCR amplification of the fragment spanning the 32-bp deletion (Fig. 1A). CD4⁺ cells purified from −/− or +/+ individuals will be referred to as −/− or +/+ CD4⁺ cells. The +/+ CD4⁺ cells supported high levels of both X4 and R5 HIV infection. The −/− CD4⁺ cells did not support R5 HIV infection, as expected, but unexpectedly were also markedly deficient compared to +/+ cells in the ability to support X4 HIV infection (Fig. 1B). This correlated with a marked reduction of CXCR4 expression on the surfaces of −/− versus +/+ cells by FACS, whereas CD4 levels were similar (Fig. 1C and D). We hypothesized that CCR5Δ32 exerts a TDN effect on CXCR4 expression through expression of the CCR5Δ32 protein.

Stimulation of PBMCs reduces the percentage of CCR5⁺ CXCR4⁺ double-positive cells.

To exert a TDN effect on CXCR4, the mutant CCR5Δ32 protein must be coexpressed in CXCR4⁺ cells. Previous studies reported a low percentage of CCR5⁺ CXCR4⁺ cells in CD3-gated, PHA-plus-IL-2-stimulated T lymphocytes (11). To readdress this issue, we performed three-color immunofluorescence flow cytometry on PBMCs from 10 healthy donors and found that the percentage of cells coexpressing CCR5 and CXCR4 is four- to fivefold higher in unstimulated PBMCs than in PHA and IL-2 blasts (Fig. 2 and Table 1). The percentage of double-positive cells depended on the cell population gated and the method of stimulation. For all donors tested, the percentage of double positives in CD3-gated cells was lower than in the CD4-gated cell population (Fig. 2 and Table 1). The reduction was donor dependent and also depended on the cell activation method. For seven donors, PHA plus IL-2 stimulation resulted in the redistribution of coreceptor expression, causing a dramatic reduction in the percentage of double-positive cells. All 10 donors tested showed a dramatic reduction in double-positive cells (CD3 gated) upon anti-CD3 plus IL-2 stimulation. The results demonstrate that the percentage of CCR5⁺ CXCR4⁺ cells was consistently higher in unstimulated CD4-gated cells (34.33 to 76.5%) than in CD3-gated cells (16 to 43%). Table 1 also shows the percentages of CD4⁺ CXCR4⁺ cells that coexpress CCR5. These values may represent the percentages of CD4⁺ CXCR4⁺ cells that could potentially be affected by coexpression of the CCR5Δ32 protein in −/− individuals. The results of this analysis demonstrate that the percentage of double-positive cells (CXCR4⁺ CCR5⁺) is higher than was previously thought and may explain the three- to fourfold drop in sensitivity to X4 virus infection shown in Fig. 1B.

Endogenous expression of the truncated CCR5Δ32 protein.

The genotypes of PBMCs isolated from −/− and +/+ individuals were confirmed by RT-PCR analysis using primers that span the 32-bp deletion (Fig. 3A). To analyze the potential protective effects of the CCR5Δ32 protein, we examined whether the truncated protein is expressed in CD4⁺ cells purified from −/− or +/+ PBMCs by using a polyclonal antiserum raised against the frame-shifted 31 amino acid residues found in CCR5Δ32 but not in CCR5. The analysis revealed a high level of expression of the CCR5Δ32 protein, which appeared as a 30-kDa band on a Western blot of CD4⁺ cells purified from PBMCs isolated from five different −/− individuals (Fig. 3B and C). The mutant CCR5Δ32 protein was also detected by using polyclonal antibodies against the N-terminal region of CCR5 (common to CCR5 and CCR5Δ32) (4) (Fig. 3C). Interestingly, the CCR5Δ32 protein was not significantly detected in CD4⁺ cells purified from two different unrelated infected −/− individuals (Fig. 3B). As expected, we were unable to detect the protein in cells from any of the +/+ subjects tested (Fig. 3B and C).

To examine the effect of cell activation on protein expression, we compared the levels of the CCR5Δ32 protein made in PHA-plus-IL-2-stimulated cells to those in unstimulated CD4⁺ cells. The results indicated down-regulation of CCR5Δ32 expression and up-regulation of CXCR4 expression upon stimulation (Fig. 3D).

Construction of recombinant adenoviruses encoding either CCR5 or CCR5Δ32.

For analysis of the apparent TDN effect of the CCR5Δ32 protein in more detail, the cDNA fragment encoding either CCR5 or CCR5Δ32 protein was inserted into the genome of Ad5 (Fig. 4A). We used this expression system because Ad5 viruses are replication defective in PBMCs and do not cause the cytopathic effect associated with the wt virus and because primary cells infected with these vectors can be infected with HIV-1 and monitored for p24 production. To demonstrate that these vectors can deliver the CCR5 or CCR5Δ32 gene into most of the infected cells, we used a similar Ad5/GFP vector encoding green fluorescent protein (GFP). We used 293 cells as a target for either infection with Ad5/GFP (10 PFU/cell) or transfection with a plasmid carrying GFP under cytomegalovirus promoter. The results demonstrated that infection resulted in efficient GFP expression in most infected cells, whereas only 10% of the transfected cells expressed GFP (data not shown).

We found that the CCR5 protein encoded by Ad5/CCR5 is a functional receptor in terms of its coreceptor activity, as demonstrated by cell fusion assays. Immunoblot analysis of cellular lysates prepared from Ad5/Δ32-infected 293 cells revealed expression of the 30-kDa CCR5Δ32 protein, as detected with either CCR5Δ32-specific antibodies (Fig. 4B) or CCR5 antibodies directed against the common N terminus (Fig. 4C). Protein bands above the 34-kDa marker band were sometimes detected in CCR5Δ32-expressing 293 cells (Fig. 4B). These protein species could represent complexes between CCR5Δ32 molecules and cellular proteins or oligomers of the CCR5Δ32 protein itself. In order to determine that our adenovirus vector does not result in the overexpression of CCR5Δ32 protein, we demonstrated that the levels of Ad5-encoded CCR5Δ32 are comparable to those endogenously made in CD4⁺ cells purified from −/− PBMCs (Fig. 4D). As expected, uninfected CD4⁺ cells or cells infected with Ad5/CCR5 or Ad5 did not show any protein band corresponding to the CCR5Δ32 protein.

trans-downmodulation and impairment of HIV coreceptor activity.

For examination of the TDN effect of Ad5-encoded CCR5Δ32 protein, CD4⁺ cells from healthy donors (wt CCR5) were infected with wt Ad5, Ad5/Δ32, or Ad5/CCR5 and stained with MAbs against CCR5, CXCR4, CXCR2, or CD4. FACS analysis demonstrated that primary CD4⁺ cells expressing recombinant CCR5Δ32 protein, but not any other protein, showed specific downmodulation of CCR5 and CXCR4 (75 to 80% reduction) (Fig. 5A). Interestingly, expression of the CCR5Δ32 protein had no significant effect on cell surface CD4 or CXCR2, indicating the specificity of the downmodulation effect (Fig. 5A). In other experiments, we found that expression of the CCR5Δ32 protein did not affect CCR4 or CCR2 expression (data not shown). These downmodulation results were confirmed by using CD4⁺ cells purified from PBMCs from 10 different donors.

For analysis of the effect of CCR5Δ32-induced downmodulation on CCR5 coreceptor activity, the same CD4⁺ cells used for Fig. 5A were mixed with HeLa cells expressing either R5 or X4 HIV-1 Env. The extent of Env-mediated fusion was measured by a reporter gene activation assay (β-Gal production). The results indicated that CCR5Δ32 expression impaired the ability of R5 and X4 Env proteins to fuse with CD4⁺ cells (Fig. 5B). As expected, R5 and X4 Env-mediated fusion was not reduced by the expression of Ad5 proteins (Ad5), CCR5, or MVHA. To exclude the possibility that the TDN effect was due to the Ad5 vector system, we repeated the analysis with a recombinant vaccinia virus vector encoding the CCR5Δ32 protein. The results confirmed those obtained with the Ad5 system (Fig. 5C and D). In contrast, the vaccinia virus-encoded CCR5Δ32 protein had no effect on fusion mediated by human T-cell leukemia virus type 1 Env, from a closely related retrovirus that recognizes a wide range of mammalian cell targets (41) (Fig. 5E). Thus, we were able to demonstrate the specificity of the TDN CCR5Δ32 activity against CCR5 and CXCR4 in two different vector systems.

Analysis of CCR5Δ32 interaction with CCR5 and CXCR4.

The yeast two-hybrid system has been used as a powerful genetic tool to rapidly select previously uncharacterized proteins and to identify novel components of signaling pathways. The expression of functional CXCR4 in yeast has recently been described (35). The yeast two-hybrid system has been previously used to demonstrate CCR5Δ32 protein interaction with wt CCR5 (8). To control for nonspecific interactions, we used either empty vectors (Table 2, rows 1, 3, 4, and 5) or a vector that expressed a nonrelevant protein (pGADT-7-MVHA). A positive interaction was detected upon cotransformation with pGBKT7-53 plus pGBKT7-TAg (positive control), pGBKT-7-CCR5Δ32 plus pGADT-7-CCR5, and pGBKT7-CCR5Δ32 plus pGADT7-CXCR4, but not with pGBKT7-CCR5Δ32 plus pGADT7-CXCR2 (under high-stringency conditions), indicating the specificity of the CCR5Δ32 protein interaction with the HIV coreceptors. A strong positive interaction under conditions of medium stringency had a colony count of >1,000. The colonies obtained with medium stringency were further analyzed, and the interaction was confirmed by using a β-Gal colony-lift filter assay (high-stringency conditions). Strongly positive colonies appeared dark blue within 30 min. These results imply that the CCR5Δ32 protein binds directly with CXCR4 and CCR5.

To confirm the yeast two-hybrid system results, we analyzed CCR5Δ32 protein-coreceptor interactions by coimmunoprecipitation. In order to detect CCR5Δ32-coreceptor heterodimers, we used 293-CD4 cells, which constitutively express CXCR4, coinfected with Ad5/Δ32 plus Ad5/CCR5 or Ad5/CCR5 plus Ad5. The cells were first examined for the CCR5Δ32 effect on HIV-1 Env fusion, and as expected, R5 and X4 fusion was dramatically inhibited by CCR5Δ32 protein expression but not by wt CCR5 expression (Fig. 6A). After confirming the CCR5Δ32 effect in our fusion assay, cell lysates were prepared and immunoprecipitated with anti-CCR5Δ32 antibodies and subjected to PFO-PAGE and blotting. Sequential probing with anti-CCR5Δ32, stripping and reprobing with anti-CCR5 (CTC-6; Protein Design Labs), and stripping and reprobing with anti-CXCR4 antibodies revealed the specific detection of high-molecular-weight bands that do not correspond to the sizes of CCR5Δ32, CCR5, and CXCR4 monomer bands (Fig. 6B). Detection of the same high-molecular-weight protein bands with different antibodies against CCR5Δ32, CCR5, and CXCR4 proteins suggested that complex formation occurred between CCR5Δ32 and either CCR5 or CXCR4. The results provide biochemical evidence for the interaction of the CCR5Δ32 protein with CCR5 and CXCR4.

Expression of CCR5Δ32 protein confers resistance to diverse HIV-1 strains.

To show efficient gene delivery by our adenovirus vector system, we used a similar recombinant adenovirus, Ad5/GFP, encoding GFP under the control of a cytomegalovirus promoter. The infection of primary CD4⁺ cells at 25 to 50 PFU/cell resulted in efficient GFP expression (green fluorescence) in most of the infected CD4⁺ cells (data not shown). This high efficiency of gene delivery allowed a detailed analysis of the biological activities of the CCR5Δ32 protein in primary cells. In order to examine Δ32 effects on HIV-1 infection, healthy seronegative (+/+) CD4⁺ cells were transduced to express Ad5-encoded CCR5 or Ad5-encoded CCR5Δ32 protein. The expression of CCR5Δ32 protein in this experiment was verified by Western blotting to show that the Ad5-encoded CCR5Δ32 protein was expressed at levels comparable to those endogenously made in −/− CD4⁺ cells (Fig. 7A). Purified CD4⁺ cells expressing wt CCR5 protein were sensitive to HIV-1 infection and showed the expected levels of p24 protein. In contrast, CD4⁺ cells expressing Ad5-encoded CCR5Δ32 protein showed a dramatic inhibition of productive infection by X4 (Fig. 7B), R5 (Fig. 7C), and R5X4 isolates (Fig. 7D). Since primary +/+ cells express endogenous CCR5, the expression of Ad5-encoded CCR5 was verified by FACS analysis of infected cells. The results indicated a twofold augmentation of CCR5 surface expression in Ad5/CCR5-infected cells (data not shown). These results are consistent with the CCR5Δ32 effects on Env fusion and provide evidence for a broad protective CCR5Δ32 effect in primary cells.