Journal: bioRxiv

Article Title: Fitness Landscapes of APOBEC3G Antagonism by HIV-1 Vif proteins

doi: 10.1101/2025.10.20.683452

Figure Lengend Snippet: A. Schematic of the HIV-1 genome with the vif gene highlighted in red. The 5’ end of vif overlaps with the Integrase coding region in pol . B. Codon-specific DMS libraries were generated across amino acid positions 12-115 of Vif from two HIV-1 clade B strains (LAI and 1203). A replication-competent HIV-1 proviral backbone was used in which the endogenous vif start codon was repositioned downstream of the pol reading frame to eliminate overlap, as described previously . DMS Libraries were cloned into this backbone and transfected into HEK293T cells to produce infectious DMS virus pools. C. DMS virus pools were used to infect SupT1 cells stably expressing physiological levels of A3G (SupT1-A3G) at a low multiplicity of infection (MOI < 0.1). Viruses encoding functional Vif (gray) replicate efficiently, whereas those encoding non-functional Vif (red) incorporate A3G (blue ovals) into virions and are depleted upon subsequent infection. Viral supernatant was harvested 72 hours after the first infection (“pre-selection”) and used to infect fresh SupT1-A3G cells. During the second passage, packaged A3G (blue ovals) deaminates cytidines on minus-strand cDNA, introducing mutations. Seventy-two hours later, “post-selection” supernatant was harvested for deep sequencing of viral RNA to quantify enrichment or depletion of individual Vif variants. D. Validation of the selection assay using control virus mixtures. Top: Wild-type (WT) Vif virus was mixed with a premature stop codon mutant at Y40 (Y40*). Bottom: WT virus was mixed with a mini-library of substitutions at position 40 (Y40X). In both cases, the proportion of WT increased only after the second passage (“post-selection”), confirming that A3G-mediated pressure selects for functional Vif variants. Statistical significance was assessed using two-sample t-tests on biological replicates for each pairwise comparison: input vs. pre-selection, input vs. post-selection, and pre- vs. post-selection. Error bars indicate standard deviation across biological replicates. (* = p < 0.05, ** = p < 0.01, ns = not significant)

Article Snippet: Unmodified SUPT1 cells (ATCC CRL-1942) used in validation assays were cultured in RPMI-1640 supplemented with 10% FBS, 1% Antibiotic–Antimycotic (Gibco, #15240062), 1% sodium pyruvate (Gibco, #11360-070), 1% glucose (Gibco, #A24940-01), and 1% GlutaMAX (Gibco, #35050-061).

Techniques: Generated, Clone Assay, Transfection, Virus, Stable Transfection, Expressing, Infection, Functional Assay, Selection, Sequencing, Biomarker Discovery, Control, Mutagenesis, Comparison, Standard Deviation

Journal: bioRxiv

Article Title: Fitness Landscapes of APOBEC3G Antagonism by HIV-1 Vif proteins

doi: 10.1101/2025.10.20.683452

Figure Lengend Snippet: A. Violin plots showing the distribution of log₂ enrichment ratios for all single missense substitutions across residues 12–115 of LAI Vif following two rounds of selection in SupT1 cells expressing A3G. Log enrichment reflects the relative fitness of each variant; higher values indicate retention or enhancement of A3G antagonism. Filled circles mark residues with significant deviation from the library-wide median: red for significantly higher and blue for significantly lower mutational tolerance. Significance was assessed using Wilcoxon signed-rank tests comparing variant enrichment values at each site to the library-wide median. Sites passing a Benjamini–Hochberg false discovery rate threshold (q ≤ 0.10) in either a two-sided test or a directionally consistent one-sided test are highlighted. Sites with insufficient variant coverage (<5 missense variants) are labeled “n”. B. Scatter plot comparing the log₂ enrichment ratio of each variant to the percentage of post-selection reads with G-to-A mutations at predicted A3G target motifs. A strong negative correlation (Spearman ρ = –0.64, p = 5.67e–91) indicates that variants with higher fitness are more effective at suppressing A3G-induced hypermutation. C. Constrained residues (blue circles from panel A) mapped onto the cryo-EM structure of the Vif– A3G–VCBC–RNA complex (PDB: 8CX0). Bold outlines indicate sites passing the two-sided Wilcoxon test (q ≤ 0.10); thin outlines indicate sites passing the one-sided Wilcoxon test (q ≤ 0.10). Gray dashed lines denote known protein–protein interactions involving constrained residues.

Article Snippet: Unmodified SUPT1 cells (ATCC CRL-1942) used in validation assays were cultured in RPMI-1640 supplemented with 10% FBS, 1% Antibiotic–Antimycotic (Gibco, #15240062), 1% sodium pyruvate (Gibco, #11360-070), 1% glucose (Gibco, #A24940-01), and 1% GlutaMAX (Gibco, #35050-061).

Techniques: Selection, Expressing, Variant Assay, Labeling, Cryo-EM Sample Prep, Protein-Protein interactions

Journal: bioRxiv

Article Title: Fitness Landscapes of APOBEC3G Antagonism by HIV-1 Vif proteins

doi: 10.1101/2025.10.20.683452

Figure Lengend Snippet: A. Amino acid sequence alignment of HIV-1 LAI and 1203 Vif proteins. Divergent residues are bolded in black. Positions not included in the DMS libraries are shown in gray. B. Violin plots showing the distribution of log₂ enrichment ratios across residues 12–115 in the 1203 Vif DMS library following selection in SupT1-A3G cells. Filled red (higher) and blue (lower) circles indicate sites with significant deviation from the library-wide median (Benjamini–Hochberg FDR q ≤ 0.10 by either a two-sided Wilcoxon test or a direction-matched one-sided test). Sites with <5 missense variants are labeled “n”. C. Scatter plot comparing each variant’s log₂ enrichment ratio with the frequency of A3G-mediated G-to-A mutations at predicted target motifs. A significant negative correlation (Spearman ρ = –0.43, p = 3.56e–20) supports the use of enrichment scores as a proxy for A3G antagonism. D. Comparison of site-level median log₂ enrichment scores between LAI and 1203 Vif DMS datasets (Spearman ρ = 0.243, p = 0.0138). Sites showing significant differences in mutational tolerance between strains (two-sided Mann– Whitney U test, uncorrected p < 0.05) are highlighted in green (less constrained in LAI) or purple (less constrained in 1203). Triangles mark positions where the LAI and 1203 sequences differ. E. Structural mapping of the 15 divergent sites onto the Vif–A3G–VCBC–RNA cryo-EM structure (PDB: 8CX0). Green and purple circles indicate sites with reduced constraint in LAI and 1203, respectively. Dashed lines denote known interaction interfaces. F. Table summarizing divergent sites, including wild-type residues in LAI and 1203, median enrichment scores, and associated interaction interfaces.

Article Snippet: Unmodified SUPT1 cells (ATCC CRL-1942) used in validation assays were cultured in RPMI-1640 supplemented with 10% FBS, 1% Antibiotic–Antimycotic (Gibco, #15240062), 1% sodium pyruvate (Gibco, #11360-070), 1% glucose (Gibco, #A24940-01), and 1% GlutaMAX (Gibco, #35050-061).

Techniques: Sequencing, Selection, Labeling, Comparison, MANN-WHITNEY, Cryo-EM Sample Prep

Journal: bioRxiv

Article Title: Fitness Landscapes of APOBEC3G Antagonism by HIV-1 Vif proteins

doi: 10.1101/2025.10.20.683452

Figure Lengend Snippet: A. Scatter plot comparing average log₂ enrichment scores of all single amino acid Vif variants between the LAI and 1203 DMS libraries. Variants with strain-specific fitness effects (≥2 standard deviations from the mean of the log₂[LAI/1203] distribution) are colored according to their mapped interaction interface: A3G (blue), CBFβ (orange), or A3G and RNA (red). Q83C and Q83S are highlighted. B. Structural depiction of Q83 (red) in the LAI Vif–A3G–VCBC–RNA cryo-EM complex (PDB: 8CX0), showing its location at the A3G-binding interface. C. Average log₂ enrichment scores for all Q83 substitutions in LAI (green squares) and 1203 (purple circles) Vif DMS libraries. The dashed line at 0 indicates the enrichment score of the wild-type residue. Note: enrichment scores are derived from pooled deep sequencing following two rounds of replication in SupT1-A3G cells and are directly comparable to those shown in . D–E. Infectivity assays for replication-incompetent, VSV-G pseudotyped HIV-1 viruses encoding Q83 variants in the LAI (D) or 1203 (E) Vif background. Infectivity was measured in SupT1 cells expressing physiological levels of A3G and normalized within each experiment to a no-A3G vector control. Bars represent mean fold change in infectivity relative to wild-type Vif in the presence of A3G. Wild-type and ΔVif viruses were included as controls. Error bars represent the standard deviation across biological replicates. Statistical significance was assessed using one-way ANOVA followed by Dunnett’s multiple comparisons test, comparing each variant to wild-type Vif: ns = not significant (P > 0.05); *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001.

Article Snippet: Unmodified SUPT1 cells (ATCC CRL-1942) used in validation assays were cultured in RPMI-1640 supplemented with 10% FBS, 1% Antibiotic–Antimycotic (Gibco, #15240062), 1% sodium pyruvate (Gibco, #11360-070), 1% glucose (Gibco, #A24940-01), and 1% GlutaMAX (Gibco, #35050-061).

Techniques: Cryo-EM Sample Prep, Binding Assay, Residue, Derivative Assay, Sequencing, Infection, Expressing, Plasmid Preparation, Control, Standard Deviation, Variant Assay