Complex of HIV-1 Integrase with Cellular Ku Protein: Interaction Interface and Search for Inhibitors

Abstract

The interaction of HIV-1 integrase and the cellular Ku70 protein is necessary for HIV replication due to its positive effect on post-integration DNA repair. We have previously described in detail the Ku70 binding site within integrase. However, the integrase binding site in Ku70 remained poorly characterized. Here, using a peptide fishing assay and site-directed mutagenesis, we have identified residues I72, S73, and I76 of Ku70 as key for integrase binding. The molecular dynamics studies have revealed a possible way for IN to bind to Ku70, which is consistent with experimental data. According to this model, residues I72 and I76 of Ku70 form a "leucine zipper" with integrase residues, and, therefore, their concealment by low-molecular-weight compounds should impede the Ku70 interaction with integrase. We have identified such compounds by molecular docking and have confirmed their capacity to inhibit the formation of the integrase complex with Ku70. Our data demonstrate that the site of IN binding within Ku70 identified in the present work may be used for further search for inhibitors of the integrase binding to Ku70.

Keywords: HIV-1; Ku70; integrase; protein–protein interactions.

Figure 1

Ku70 residues 47–74 are involved in the HIV-1 integrase binding. (A) Ku70 amino acid sequence. Ku70 residues covered by MS-identified peptides after GluC treatment are colored in blue, and after trypsin digestion are underlined. Colored boxes indicate peptides precipitated with full-length HIV-1 IN. (B,C) MS identification of Ku70 peptides precipitated with HIV-1 IN in peptide fishing experiments. The His₆-Ku70 was digested by trypsin (B) or GluC (C) proteases and then incubated with glutathione-agarose beads with (lower graphs, signed +IN) or without (upper graphs, signed −IN) immobilized GST-IN. The following m/z 1354.7 and 1397.7 correspond to Ku70^81–92 and its carbamoylated form; 3226.5, 3269.5, and 3302.5 – to Ku70^47–74, its carbamoylated form and β-mercaptoethanol adduct, respectively; 1695.9 and 1939.0 – to Ku70^94–107 and Ku70^92–107, respectively. (D) Solvent accessible surface area (SASA) values of the identified peptides, and the peptide positions in the Ku70/Ku80 heterodimer structure (PDB ID: 1JEQ). (E) The structure of truncated Ku70 mutants (left), and analysis of their binding to IN by GST pull-down assay (right).

Figure 2

Identification of Ku70 residues involved in the HIV-1 IN binding. (A) Schematic representation of the Ku70_1–78 mutant secondary structure (top), and GST-pull down analysis of the interaction between truncated Ku70 mutants and HIV-1 IN (bottom). (B) GST-pull down analysis of the interaction between HIV-1 IN and Ku70 mutants: Ku70_1–250_Δ51–57/insAG, Ku70_1–250-Q65A/Q68A, Ku70_1–250_S69A/I72A, and Ku70_1–250_S73A/I76A. Top left insert depicts the position of Ku70 residues selected for mutagenesis in the structure of Ku70 (dark gray)/Ku80 (light gray) heterodimer (PDB ID: 1 JEQ). (C) His₆ pull-down analysis of the interaction between full-length Ku70 or its double mutants and HIV-1 IN. Quantitative analysis of WB presented at Figure S1. (D) Immunoprecipitation of the HA-tagged IN with 3xFLAG-tagged Ku70_wt, Ku70_S69A/I72A, or Ku70_S73A/I76A in HEK293T cells. Eluates were analyzed by Western blot (bottom panel) with anti-FLAG antibodies to detect Ku70_3FLAG and anti-HA antibodies to detect HIV-1 IN. Ten percent of the cell lysates were used for input analysis by Western blot (top panel) with anti-Ku70 antibodies for the detection of Ku70 (Ku70_3xFLAG migrates slightly slower than endogenous Ku70), and anti-HA to detect HA-IN; anti-tubulin detection was used as loading control. (E) GST pull-down analysis of the interaction between Ku70_1–250 point mutants and HIV-1 IN by Western blot (left) and quantitative analysis (right). Mean values ± SD of three independent experiments are presented. Significance was determined by two-way ANOVA, * = adjusted p-value < 0.05, **** = adjusted p-value < 0.0001.

Figure 3

Three-dimensional model of the Ku70/IN complex. (A) Gibbs energy profile of the N-terminal domain of Ku70 and IN binding. The error bars are calculated using umbrella integration approach. The inset demonstrates the hydrophobic binding motif of the so-called “leucine zipper”. Ku70 and IN carbon atoms are colored pink and cyan, respectively, and hydrogen atoms are white. (B) Calculated structure of the Ku70/Ku80 heterodimer and IN dimer complex. IN dimer is shown in light blue: the monomer that interacts with the N-terminal domain (up to 250th residue) of Ku70 is shown in cartoon representation. The dark isosurface is the Ku70/Ku80 heterodimer, except the N-terminal domain of the Ku70, which is shown in pink cartoon representation. (C) Alignment of the crystal structure of Ku70/Ku80 heterodimer (PDB ID: 1JEQ [33]) and the calculated structure. The N-terminal domain of Ku70 from the PDB ID: 1JEQ is shown in green. The arrow shows the rotation direction of the N-terminal Ku70 domain.

Figure 4

Y021-2376 blocks HIV-1 IN binding to Ku70. (A) Distribution of relative SASA S69 + I72 + S73 + I76 calculated for complexes of the Ku heterodimer with top 31 investigated compounds. White dots correspond to the median of the distributions. (B) The effect of 100 μM compounds on the Ku70/IN complex formation. (C) Fluorescent pull-down assay analysis of the interaction of His6-Ku70-tRFP (200 nM) and GST-mCer-IN (200 nM) in the presence of an increasing concentration of Y021-2376. (D) Chemical structure of the most active compound Y021-2376 (top panel) and probable inhibitor position in the Ku70/Ku80 heterodimer (bottom panel). Ku70 residues involved in HIV-1 IN binding are pink. Ku70 surface is colored in dark gray and Ku80 in light gray.