Divide et impera: An In Silico Screening Targeting HCMV ppUL44 Processivity Factor Homodimerization Identifies Small Molecules Inhibiting Viral Replication

Abstract

Human cytomegalovirus (HCMV) is a leading cause of severe diseases in immunocompromised individuals, including AIDS patients and transplant recipients, and in congenitally infected newborns. The utility of available drugs is limited by poor bioavailability, toxicity, and emergence of resistant strains. Therefore, it is crucial to identify new targets for therapeutic intervention. Among the latter, viral protein-protein interactions are becoming increasingly attractive. Since dimerization of HCMV DNA polymerase processivity factor ppUL44 plays an essential role in the viral life cycle, being required for oriLyt-dependent DNA replication, it can be considered a potential therapeutic target. We therefore performed an in silico screening and selected 18 small molecules (SMs) potentially interfering with ppUL44 homodimerization. Antiviral assays using recombinant HCMV TB4-UL83-YFP in the presence of the selected SMs led to the identification of four active compounds. The most active one, B3, also efficiently inhibited HCMV AD169 strain in plaque reduction assays and impaired replication of an AD169-GFP reporter virus and its ganciclovir-resistant counterpart to a similar extent. As assessed by Western blotting experiments, B3 specifically reduced viral gene expression starting from 48 h post infection, consistent with the inhibition of viral DNA synthesis measured by qPCR starting from 72 h post infection. Therefore, our data suggest that inhibition of ppUL44 dimerization could represent a new class of HCMV inhibitors, complementary to those targeting the DNA polymerase catalytic subunit or the viral terminase complex.

Keywords: HCMV; PAP; antivirals; pUL54; ppUL44; protein-protein interactions; screening; small molecules.

Figure 1

An in silico screening to identify SMs inhibiting ppUL44 dimerization. (A) Graphic representation of UL44(1-290) homodimers. One monomer is represented as surface (Monomer A), the other one as ribbons, with residues involved in dimerization being shown as sticks (Monomer B). (B) Inset of the homodimerization interface, with one monomer shown in surface and residues of the other monomer involved in the dimerization being shown as sticks. (C) Schematic overview of the virtual screening to identify SMs potentially disrupting UL44 homodimerization. The Glide software was used to dock molecules to the interface of the two monomers (PDB code: 1T6L). Three rounds of screening were performed using the High throughput virtual screening (HTVS), Standard Precision (SP), and Extra Precision (XP) docking settings. After each docking round, the top-ranked molecules in term of docking score were selected for the following round. The resulting 500 molecules were further filtered by visual inspection, cluster analysis and on the basis of their commercial availability, and 18 compounds were selected for further studies.

Figure 2

Identification of compounds interfering with HCMV replication. MRC5 cells were infected with TB4-UL83-EYFP at an MOI of 0.03 IU/cell and treated with each SM either at concentration of 10 μM (A,C) or 100 μM (B,D). In parallel, uninfected MRC5 cells were also treated for assessment of SMs cytotoxicity. Seven days post treatment, cells were processed for data acquisition and analysis as described in the Materials and Methods section. Mean YFP values relative to infected cells treated with the indicated SMs are expressed as a percentage of DMSO-treated cells (red bars). Cell viability was assessed by MTS (A,B) or Cell Titer Glo^® assays (C,D), and data expressed as a percentage of DMSO-treated cells (black columns). The mean + standard error of the mean (SEM) relative to 3 independent experiments is shown. * indicates the presence of precipitates. (E) The chemical structure of active molecules is shown.

Figure 3

Determination of ED₅₀ and CC₅₀ values of SMs by FRA and MTT assay. MRC5 cells were infected with TB4-UL83-EYFP at an MOI of 0.03 IU/cell and treated with increasing concentrations of the indicated compounds. In parallel, uninfected MRC5 cells were also treated for assessment of SMs cytotoxicity. Seven days post treatment, cells were processed for data acquisition and analysis as described in the Materials and Methods section. Mean values YFP values relative to infected cells treated with the indicated compounds are expressed as a percentage of DMSO-treated cells (red squares). Cell viability was assessed by MTT assays, and data expressed as a percentage of DMSO treated cells (blue circles). For each compound, representative plots are shown, along with the cytotoxic concentration 50 (CC₅₀) and effective dose 50 (ED₅₀) mean values + standard deviation of the mean relative to at least 4 independent experiments.

Figure 4

Effect on cell viability and proliferation of SMs inhibiting HCMV replication. MRC5 cells were treated with increasing concentrations of indicated compounds or with solvent only, as described in the Materials and Methods section. At the indicated time, post treatments cells were processed for intracellular ATP quantification. Data shown are the mean + standard error of the mean relative to 3 independent experiments.

Figure 5

Inhibition of AD169 replication by B3 in PRA. Dose-response curves for B3 (blue) or GCV (red) were obtained by infecting HFF cells with HCMV AD169 and then treating them with different concentrations of the indicated compounds. Data shown are the means ± standard deviation of the mean relative to four independent experiments performed in duplicate.

Figure 6

B3 efficiently inhibits replication of the GCV-resistant AD169-GFP26 virus. MRC5 were infected with either AD169-GFP virus or its GCV-resistant counterpart AD169-GFP26 at a MOI of 0.05 IU/cells and treated with increasing concentrations of the indicated compounds. At 7 days p.i., cells were lysed and plates processed for FRAs (A,C), while supernatants were collected and used for VYRAs (B,D). For the latter experiments, MRC5 cells were infected with serial dilutions of supernatants derived from infected cells. At 7 days p.i., viral titers were calculated using the TCID₅₀ method. Data from both assays were used to calculate the ED₅₀ relative to the two viruses (A,B) as well as the ratio between the ED₅₀ calculated for AD169-GFP26 and AD169-GFP (C,D), for both GCV and B3. Data shown are single measurements, means, and standard deviation of the mean relative to at least three independent experiments (see Table 1), along with the p-value relative to the Student’s t-test the indicated groups; *: p ≤ 0.05.

Figure 7

B3 specifically impairs early and late HCMV AD169 gene expression. MRC5 were infected with HCMV AD169 and treated as described in the Materials and Methods section. At the indicated time points p.i., cells were lysed and processed for Western blotting to detect the expression of the immediate early IE1 antigen (A,B; at 6 and 12 h p.i.), the early-late antigens ppUL44 and pp65 (C,D; at 24 and 48 h p.i.), and the late antigen pp28 (E,F; at 72 and 96 h p.i.). GAPDH or β-actin were also detected as loading controls. (G,H): Loading controls were used to normalize signal intensity relative to each antigen after treatment with GCV (G) or B3 (H). Lysates of mock infected cells were analyzed to verify antibody specificity. Data shown are the mean + standard deviation of the mean relative to three independent experiments.

Figure 8

B3 impairs HCMV genome replication. MRC5 cells were infected with HCMV TB4-UL83-EYFP and treated with the indicated compounds or vehicle alone (DMSO) as described in the Materials and Methods section. At the indicated time points p.i., cells were lysed and processed for qPCR to detect the presence of viral and host cell genomic DNAs as described in the Materials and Methods section. Data are shown as HCMV DNA fold changes with respect to DMSO-treated cells at 24 h p.i. Means ± standard deviation of the mean from two independent experiments performed in quadruplicate are reported.

Figure 9

Analysis of the theoretical binding mode of B3 to ppUL44. A graphic representation of ppUL44(1-290) dimerization pocket is shown in the absence (A) or in the presence (B) of B3. One monomer is represented as surface, with residues involved in dimerization represented as green (Monomer A), or gray sticks (Monomer B), whereas B3 is shown as yellow sticks. (C) The 2D ligand interaction diagram of the theoretical binding mode for B3 is shown. Hydrogen bond interactions are shown as violet arrows. Positive and negative charged amino acids are represented in blue and red, respectively. Residues involved in hydrophobic or polar interactions are shown in green and light blue, respectively. Ligand-exposed fractions are indicated as a gray, circular shadow.