Analysis of A3G/A3F splice variant expression in primary peripheral blood subsets. A, quantitative real-time RT-PCR of A3G and A3F analysis of splice variant expression. Primers were designed to detect total A3G/A3F mRNA or specific splice variants as indicated in Fig. 1. PBMCs, CD4 T cells, or monocytes were isolated from the peripheral blood of healthy anonymous donor. RNA was isolated from unstimulated PBMCs or PBMCs stimulated with phytohemagglutin/IL-2 for 72 h. Monocyte populations were purified by positive CD14 selection; CD4 T cells were purified by positive CD4 selection from the CD14- cell population and stimulated with phytohemagglutin/IL-2 for 72 h. Macrophages differentiated in the presence of human AB serum were left untreated or treated overnight with interferon-α (IFN-α) before RNA isolation. Total RNA was isolated from each cellular population, normalized for RNA content, and reverse transcribed with random hexamer primers. qRT-PCR was performed with the indicated primers and probes. Bars represent ratios of A3G spliceoform mRNAs (blue) to total A3G mRNA or A3F spliceoform mRNA (green) to total A3F mRNA. Values are mean ± S.E. (n = 4 independent donors). B, Western blot analysis of A3G/A3F splice variant expression. C-terminal antibodies for A3G or A3F were used to detect splice variant protein expression. Lysates from 293T cells transfected with indicated splice variants were used as a control and for comparison of the relative size of detected products. PBMCs, CD4 T cells, or monocytes were isolated from the peripheral blood of healthy anonymous donors. Total protein was isolated from unstimulated PBMCs/CD4 T cells or CD4 T cells stimulated with phytohemagglutin/IL-2 for 72 h. Monocyte populations were purified by positive CD14 selection. Macrophages differentiated in the presence of GM-CSF were left untreated or treated overnight with IFN-α before protein isolation. Total protein was isolated from each cellular population, normalized for protein content, and run on a denaturing SDS-PAGE gel. Macrophages and IFN-α-treated macrophages were run on the same blot to indicate induction of A3G, A3F, and A3A. Lanes represent two independent donors, which are representative of at least four donors analyzed for each cellular subset.

Vif sensitivity and antiviral activity of A3FΔ2 and A3FΔ2–4. A, full-length A3F or A3F spliceoform expression plasmids were cotransfected into 293T with vector alone, wild-type WT Vif, or the mutant VifSLQ at a plasmid mass ratio of 1:5 or 1:2. Lysates of transfected 293T cells were analyzed by Western blotting for transfected protein expression with antibodies specific for A3F and Vif. Anti-Hsp90 was used as control for total protein loading. B, A3FΔ2 and A3FΔ2–4 exhibit antiviral activity. NL4–3 or NL4–3ΔVif viruses were generated from 293T cells cotransfected with empty vector or A3F expression plasmids as indicated. A fixed amount of total DNA was transfected; the mass ratio of A3F to pNL4–3 was varied from 1:30 to 1:1. Viral supernatants were normalized for Gag-p24 content and used to infect the indicator cell line TZM-bl. Infection was measured as relative luciferase units after 30–48 h. Infection is indicated as the proportion of luciferase units relative to infection in the absence of A3F. Values are mean ± S.E. (n = 3 independent transfections).

Schematic of A3G and A3F spliceoform mRNAs and predicted protein products. A, A3G pre-mRNA contains eight exons (E1–E8) and seven introns. Primers used for PCR amplification of splice variants are shown in black. One forward primer was placed in the 5′UTR of exon 1, and seven reverse primers were designed in each of the remaining exons (black arrows). Total RNA was isolated from PBMCs of uninfected, healthy donors and reverse transcribed with random hexamers. Seven individual PCR reactions were performed to detect inclusion and exclusion of exons in splice variants. Red arrows indicate positions of primers used for subsequent real-time qRT-PCR of specific splice variants. B, A3F pre-mRNA contains seven exons (E1–E7) and six introns. Primers used for PCR detection of A3F splice variants are shown in black. PCR reactions were performed on cDNA generated as described in A. Primers used for subsequent real-time qRT-PCR of specific splice variants are shown in red. C, schematic of full-length A3G protein structure and predicted structures of A3G splice variants. If translation of A3GΔ2 and A3GΔ2–5 starts at the first ATG, a frameshift will occur, resulting in a stop codon immediately downstream. However, if A3GΔ2 and A3GΔ2–5 utilize an alternate downstream ATG, an in-frame protein will be produced. D, structure of full-length A3F protein and predicted structures of A3F splice variants. As described above, translation initiation at the first ATG will produce a frameshift and subsequent stop codon for A3FΔ2.

A3G spliceoforms are not sensitive to Vif and lack antiviral activity. A, full-length A3G expression plasmids or A3G spliceoform expression plasmids were cotransfected into 293T with vector alone, wild-type (WT) Vif, or the mutant Vif^SLQ at a plasmid mass ratio of 1:5 or 1:2. Transfected 293T lysates were analyzed by Western blotting for transfected protein expression with antibodies against A3G and Vif. Anti-Hsp90 was used as a control for total protein loading. B, NL4–3 or NL4–3ΔVif viruses were generated from 293T cells cotransfected with empty vector or A3G expression plasmid DNA. A fixed amount of total DNA was transfected; the mass ratio of A3G to pNL4–3 was varied from 1:30 to 1:1. Viral supernatants were normalized for p24-Gag content and used to infect the indicator cell line, TZM-bl. Levels of infection were measured by detection of luciferase activity 36–48 h later. Infection level is indicated as the proportion of luciferase units generated in the presence and absence of A3G. Values are mean ± S.D. (n = 6 independent infections). C, viral preparations were generated as described above using a 1:1 mass ratio of A3G spliceoforms to pNL4–3. Relative infection levels are shown. Values are mean ± S.E. (n = 3 independent transfections).

A3F spliceoforms exhibit reduced deoxycytidine deaminase activity and lower virion incorporation. A, FLAG-tagged A3F, A3FΔ2, or A3FΔ2–4 were immunoprecipitated with anti-FLAG beads, eluted with FLAG peptide, serially diluted 5-fold (1:1 to 1:125), and analyzed in a deoxycytidine deaminase assay that utilizes a Cy5.5-labeled DNA probe substrate containing a consensus dinucleotide site for A3F-deamination. Upon the addition of uracil N-glycosidase, an abasic site is produced that is subsequently cleaved by addition of NaOH. Enzymatic activity was measured by the appearance of a smaller Cy5.5- labeled cleavage product. Normalized levels of immunoprecipitated proteins were blotted with anti-FLAG antibody. B, viral supernatants were generated by cotransfecting 293T cells with various mass ratios of A3F/A3FΔ2/A3FΔ2–4 and pNL4–3 WT/pNL4–3ΔVif expression plasmids. A3 incorporation was measured by Western blot analysis of concentrated, iodixanol-purified virions. Anti-p24-Gag was used as a control for virion protein loading. A3F and A3FΔ2 virions or A3F and A3FΔ2–4 virions were electrophoresed on the same gel and blotted simultaneously. A3FΔ2–4 was incorporated at such low levels that a longer exposure is given for the anti-FLAG antibody. Results presented for A3F represent a 30-s exposure, whereas A3FΔ2–4 results were obtained following a 5-min exposure. No signal for A3FΔ2–4 was detected after a 30-s exposure; the A3F signal was saturated at 5 min. A3F and A3FΔ2 are shown at the same exposure. NL4–3ΔVif-producing 293T cells were blotted as a control for transfection efficiency. Levels of FLAG-tagged A3F/A3FΔ2 and A3F/A3FΔ2–4 are shown at the same exposure time. Hsp90 was employed as a protein loading control for the cellular lysates.